Functional Neuroscience: Evoked Potentials and Related Techniques: Presentations from the 8th International Evoked Potentials Symposium (Fukuoka, ... 59 (Supplements to Clinical Neurophysiology)

Functional Neuroscience: Evoked Potentials and Related Techniques Selected Presentations from the 8th International Evok...

Author: Colin Barber | Sadatoshi Tsuji MD

85 downloads 1061 Views 14MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Functional Neuroscience: Evoked Potentials and Related Techniques Selected Presentations from the 8th International Evoked Potentials Symposium (Fukuoka, Japan) EDITED BY

C. BARBER Medical Physics Department, Queen’s Medical Centre Nottingham, University Hospital NHS Trust, Nottingham NG7 2UH, UK

S. TSUJI, S. TOBIMATSU, T. UOZUMI, N. AKAMATSU Departments of Neurology and Clinical Neurophysiology, University of Occupational and Environmental Health, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555, Japan

A. EISEN Department of Neurology, University of British Columbia, 2826 Highbury Street, Vancouver, BC V6R 3T6, Canada

SUPPLEMENTS TO CLINICAL NEUROPHYSIOLOGY VOLUME 59 2006

Supplements to Clinical Neurophysiology, 2006, Vol. 59

ELSEVIER B.V. Sara Burgerhartstraat 25 P.O. Box 211, 1000 AE Amsterdam The Netherlands

ELSEVIER Inc. 525 B Street, Suite 1900 San Diego, CA 92101-4495 USA

ELSEVIER Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB UK

ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK

© 2006 Elsevier B.V. All rights reserved. This work is protected under copyright by Elsevier B.V., and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier Health Science Rights Department, Elsevier Inc., 625 Walnut Street, Philadelphia, PA 19106, USA; phone: (+1) 215 238 7869, fax: (+1) 215 238 2239, e-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://www.elsevier.com), by selecting ‘Customer Support’ and then ‘Obtaining Permissions’. In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+1) (978) 7508400, fax: (+1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier Health Science Rights Department, at the phone, fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2006 Library of Congress Cataloging in Publication Data A catalog record is available from the Library of Congress. British Library Cataloguing in Publication Data A catalogue record is available from the British Library. ISBN-10: 0-444-51697-2 (this volume) ISBN-13: 978-0-444-51697-8 (this volume) ISSN (Series): 1567-424X The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.

Preface

In the period – a little over a quarter of a century – since this series of Evoked Potentials (EPs) symposia began, the field has evolved and matured to the extent that, strictly, EPs can no longer be considered a field of study per se. They have become instead an indispensable part of the panoply of investigative techniques available to functional neuroscience studies. It is a measure of their success that they have become so well accepted, so well used, so well integrated with other techniques, that they rarely appear in an individual starring role, but as players in a team. Thus, this volume, which comprises invited presentations from the eighth (and probably last) International Evoked Potentials Symposium, is somewhat difficult to sectionalise. It is no longer possible to categorise chapters by modality and we have chosen, instead, to group them under broader headings reflecting the use of EPs in the search for sources and origins of phenomena; the insights they offer into cognitive processing; and their role in clinical advances. They do not fit neatly, and some span even these broad boundaries, but we hope this approach will guide you in finding the work that interests you particularly. If we have failed in this regard please persevere, for the work reported herein is of the highest order. It reflects the state of the art across the globe, and across the wide variety of endeavours in which EPs play a part. We should like to thank all the contributors to this volume for sharing with us their scientific advances and for showing that, for all their maturity, EPs can still generate intellectual excitement. Colin Barber Sadatoshi Tsuji Shozo Tobimatsu Takenori Uozumi Naoki Akamatsu Andrew Eisen

List of Contributors

Akamatsu, N.

Department of Neurology, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555, Japan.

Andersen, O.K.

Center for Sensory-Motor Interaction, Laboratory for Experimental Pain Research, Aalborg University, Fredrik Bajersvej 7, D3, DK-9220 Aalborg, Denmark.

Arai, M.

Department of Neurology, Dokkyo University School of Medicine, Tochigi 321-0293, Japan.

Arendt-Nielsen, L.

Center for Sensory-Motor Interaction, Laboratory for Experimental Pain Research, Aalborg University, Fredrik Bajersvej 7, D3, DK-9220 Aalborg, Denmark.

Arimura, K.

Department of Neurology and Geriatrics, Kagoshima University Graduate School of Medical and Dental Sciences, 8-35-1 Sakuragaoka, Kagoshima 890-8520, Japan.

Ashby, P.

Division of Neurology, Department of Medicine, University of Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan.

Avitable, M.

Center for Scientific Computing, SUNY Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203-2098, USA.

Azuma, Y.

Division of Clinical Electrophysiology, Department of Neurology, University of Iowa, College of Medicine, 200 Hawkins Drive (0150 RCP), Iowa City, IA 52242, USA.

Baba, M.

Department of Neurological Sciences, Hirosaki University School of Medicine, Hirosaki, Japan.

Battista, V.

Department of Neurology, The Neurological Institute, Eleanor and Lou Gehrig MDA/ALS Research Center, Columbia University, 710 W 168th Street, New York, NY 10032, USA.

viii Bodis-Wollner, I.

Department of Neurology, SUNY Downstate Medical Center, 450 Clarkson Avenue, Box 1213, Brooklyn, NY 11203-2098, USA.

Celesia, G.G.

Loyola University of Chicago, 3016 Heritage Oak Lane, Oak Brook, IL 60523, USA.

Chen, R.


Chu, N.-S.

Department of Neurology, Chang Gung Memorial Hospital, and College of Medicine, Chang Gung University, 199 Tung-Hwa North Road, Taipei 10591, Taiwan.

Davis, K.D.

Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, ON M5T 2S8, Canada.

Deletis, V.

Institute for Neurology and Neurosurgery, St. Luke’s-Roosevelt Hospital, 1000 10th Avenue, New York, NY 10019-1147, USA.

Dostrovsky, J.O.

Department of Physiology, University of Toronto, Toronto, ON M5T 2S8, Canada.

Floyd, A.


Forgacs, P.B.

Department of Neurology, SUNY Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203-2098, USA.

Goto, Y.

Department of Clinical Neurophysiology, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582, Japan.

Han, T.R.

Department of Rehabilitation Medicine, Seoul National University College of Medicine, 28 Yeongundong, Jongrogu, 110-744 Seoul, Republic of Korea.

Hanajima, R.


Harhula, M.

Department of Neurology, SUNY Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203-2098, USA.

Hashimoto, I.

Human Information Systems Laboratory, Kanazawa Institute of Technology, Tokyo, Japan.

ix Hashimoto, T.


Hayashi, T.

Department of Investigative Radiology, Research Institute, National Cardiovascular Centre, 5-7-1 Fujishirodai, Suita, Osaka 565-8565, Japan.

Hirata, K.


Hoshiyama, M.

Department of Health Sciences, Faculty of Medicine, Nagoya University, Nagoya 461-8673, Japan.

Hozumi, A.


Hristova, A.


Husain, A.M.

Department of Medicine (Neurology), Duke University Medical Center, and Neurodiagnostic Center, Veterans Affairs Medical Center, Box 3678, 202 Bell Building, Durham, NC 27710, USA.

Hutchison, W.D.

Department of Physiology, University of Toronto, Toronto, ON M5T 2S8, Canada.

Igasaki, T.

Department of Electrical and Computer Engineering, Kumamoto University, 2-39-1 Kurokami, Kumamoto City, Kumamoto 860-8555, Japan.

Ikemoto, T.

Department of Orthopedics, Kochi Medical School, Kohasu Oko-cho, Nankoku City, Kochi 783-8505, Japan.

Ilić, T.V.

Motor Cortex Laboratory, Department of Neurology, Johann Wolfgang Goethe-University of Frankfurt, Schleusenweg 2–16, D-60528 Frankfurt am Main, Germany.

Inagaki, M.

Division of Diagnostic Research, Department of Developmental Disorders, National Institute of Mental Health, National Center of Neurology and Psychiatry (NCNP), 1-7-3 Kohnodai, Ichikawa 272-0827, Japan.

Inui, K.

Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585, Japan.

x Ishida, K.


Ishiguchi, H.


Izumi, S.-I.

Department of Physical Medicine and Rehabilitation, Tohoku University Graduate School of Medicine, 2-1 Seiryo-cho, Aoba-ku, Sendai 980-8575, Japan.

Jones, S.J.

Department of Clinical Neurophysiology, The National Hospital for Neurology and Neurosurgery, Queen Square, London WC1N 3BG, UK.

Jung, P.

Motor Cortex Laboratory, Department of Neurology, Johann Wolfgang GoetheUniversity of Frankfurt, Schleusenweg 2–16, D-60528 Frankfurt am Main, Germany.

Jung, S.H.

Department of Rehabilitation Medicine, Seoul National University College of Medicine, 28 Yeongundong, Jongrogu, 110-744 Seoul, Republic of Korea.

Kaga, M.


Kaga, Y.


Kaji, R.

Department of Neurology, Tokushima University Faculty of Medicine, 8-15-3 Chome, Kuramoto-Cho, Tokushima City, Tokushima 770-8503, Japan.

Kakigi, R.


Kanda, M.

Department of Neurology, Takeda General Hospital, 28-1 Moriminami-machi, Fushimi-ku, Kyoto 601-1495, Japan.

Kaufmann, P.


Kim, D.-S.

Department of Neurology, Pusan National University Hospital, Pusan National University College of Medicine, 1-10 Ami-Dong, Seo-Gu, Busan, Republic of Korea.

xi Kimura, J.

Department of Neurology, University of Iowa Health Care, 200 Hawkins Drive, Iowa City, IA 52242, USA.

Kojima, Y.


Kurokawa, H.


Le Pera, D.

Department of Neurological Rehabilitation, IRCCS, San Raffaele Pisana, Rome, Italy.

Lesser, R.P.

Department of Neurology and Neurosurgery, 2-147 Meyer Building, Johns Hopkins University, 600 N Wolfe Street, Baltimore, MD 21287-7247, USA.

Lozano, A.M.

Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, ON M5T 2S8, Canada.

Maeda, R.

Department of Neurosurgery, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555, Japan.

Mäkelä, J.P.

BioMag Laboratory, Helsinki University Central Hospital, P.O. Box 340, FIN-00029 HUS, Helsinki, Finland.

Matsunaga, K.


Miki, K.


Mitsumoto, H.


Mouraux, A.

Laboratoire de Neurophysiologie (NEFY), Université Catholique de Louvain, B-1200 Brussels, Belgium.

Murase, N.


Murayama, N.


xii Nakata, H.


Neshige, R.

Neshige Neurological Clinic, 38-17 Tyuou-machi, Kurume, Fukuoka 830-0023, Japan.

Ng, A.R.


Nihei, K.

National Center for Child Health and Development, 2-10-1 Okura, Tokyo 157-8535, Japan.

Nonaka, Y.

Nihon Kohden Corporation, Tokyo, Japan.

Ohnishi, T.

Department of Radiology, National Centre Hospital for Mental, Nervous, and Muscular Disorders, National Centre of Neurology and Psychiatry, 4-1-1 Ogawahigashi, Kodaira, Tokyo 187-8551, Japan.

Okabe, S.

Department of Neurology, Division of Neuroscience, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan.

Okazaki, S.

Institute of Disability Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan.

Ozaki, H.

Laboratory of Physiology, Ibaraki University, Ibaraki, Japan.

Ozaki, I.

Faculty of Health Sciences, Aomori University of Health and Welfare, 58-1 Mase, Hamadate, Aomori 030-8505, Japan.

Park, B.K.

Department of Rehabilitation Medicine, Pusan National University Hospital, Pusan National University College of Medicine, 1-10 Ami-Dong, Seo-Gu, Busan, Republic of Korea.

Park, K.-H.

Department of Neurology, Pusan National University Hospital, Pusan National University College of Medicine, 1-10 Ami-Dong, Seo-Gu, Busan, Republic of Korea.

Plaghki, L.

Unité de Réadaptation (READ), Université Catholique de Louvain, B-1200 Brussels, Belgium.

Puce, A.

Department of Radiology, Center for Advanced Imaging, West Virginia University, Morgantown, VA 26506, USA.

Pullman, S.L.


xiii Qiu, Y.


Sadato, N.

Department of Cerebral Research, National Institute for Physiological Sciences, 38 Nishigonaka, Myodaiji, Okazaki 444-8585, Aichi, Japan.

Selesnick, I.

Department of Electrical and Computer Engineering, Polytechnic University, Brooklyn, NY 11201, USA.

Shimadu, H.


Skrandies, W.

Institute of Physiology, Justus-Liebig University, Aulweg 129, D-35392 Giessen, Germany.

Sonoo, M.

Department of Neurology, Teikyo University School of Medicine, Kaga 2-11-1, Itabashi-ku, Tokyo 173-8605, Japan.

Stewart, M.

Department of Physiology and Pharmacology, State University of New York, Downstate Medical Center, Box 31, Brooklyn, NY 11203, USA.

Tamagawa, A.


Tanaka, H.


Tang, M.-X.

Biostatistics and Sergievsky Center, Columbia University, 710 W 168th Street, New York, NY 10032, USA.

Tani, T.


Taniguchi, S.


Taniwaki, T.

Departments of Clinical Neurophysiology and Neurology, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582, Japan.

Tanoue, K.


xiv Terao, Y.


Tobimatsu, S.


Tsuboya, H.


Tsuji, S.


Tsuji, T.


Tsurusawa, R.

Departments of Clinical Neurophysiology and Pediatrics, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582, Japan.

Ugawa, Y.


Uozumi, T.


Urasaki, E.


Urushihara, R.


Ushida, T.


Valeriani, M.

Department of Neurology, Ospedale Pediatrico Bambino Gesù, IRCCS, Piazza Sant’Onofrio 4, 00165 Rome, Italy.

Von Gizycki, H.

Center for Scientific Computing, SUNY Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203-2098, USA.

xv Wang, X.


Watanabe, O.


Watanabe, S.


Wei, F.-C.

Department of Plastic and Reconstructive Surgery, Chang Gung Memorial Hospital, and College of Medicine, Chang Gung University, 199 Tung-Hwa North Road, Taipei 10591, Taiwan.

Yaegashi, Y.

Department of Social Science, Akita Keijoh College, Oodate, Japan.

Yamada, T.


Yamaguchi, S.

Departments of Neurology, Hematology and Immunology, Shimane University School of Medicine, Izumo, Shimane 693-8501, Japan.

Yamasaki, T.


Yanagisawa, T.


Yokota, A.


Zeng, X.-H.

Department of Neurology, Dokkyo University School of Medicine, Tochigi 321-0293, Japan, and Neurological Sciences Institute, Oregon Health and Sciences University, 505 NW 185th Avenue, Beaverton, OR 97006, USA.

Ziemann, U.

Motor Cortex Laboratory, Department of Neurology, Johann Wolfgang GoetheUniversity of Frankfurt, Schleusenweg 2–16, D-60528 Frankfurt am Main, Germany.

Functional Neuroscience: Evoked Potentials and Related Techniques (Supplements to Clinical Neurophysiology, Vol. 59) Editors: C. Barber, S. Tsuji, S. Tobimatsu, T. Uozumi, N. Akamatsu, A. Eisen © 2006 Elsevier B.V. All rights reserved

3

Chapter 1

Neurocognitive development of visuocognitive motor behavior revealed by event-related potentials in cued continuous performance test Shinji Okazakia,* and Hisaki Ozakib a

Institute of Disability Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8572 (Japan) b Laboratory of Physiology, Ibaraki University, Ibaraki (Japan)

1. Introduction Many prior studies have reported that motor response is controlled by various underlying cerebral structures (Ballard, 2001). For example, Mesulam (1981) suggested a cortical network for attention. In this model, the posterior parietal cortex contributes to spatial attention, and a prefrontal contribution to anticipated motor control is also described. Brainstem structures might also play an important role in maintaining the arousal level. The continuous performance test (CPT) (Rosvold et al., 1956) has been used to examine motor response control (Fallgatter et al., 1997). The CPT is an attention task, in which a series of stimuli are presented, and is divided into CPT-X, CPT-AX, and CPT-double by target demand (Corkum and Siegel, 1993). Above all, CPT-AX is the paradigm where subjects are asked to respond to a target within a cue-target sequence such as A-X. As expectation by the warning stimulus might affect motor

*Correspondence to: Shinji Okazaki, Institute of Disability Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8572, Japan. Tel/Fax: +81-29-853-6804; E-mail: [email protected]

behavior, CPT-AX is the optimal paradigm for investigating anticipated motor control. As well as the development of sustained attention, the developmental change in motor control in children has been studied using reaction time tasks (Williams et al., 1999) and CPT (Greenberg and Waldman, 1993). Recent electrophysiological data have complemented and extended the neurocognitive findings on motor control (Overtoom et al., 1998; Van Leeuwen et al., 1998). However, motor control difficulty is a crucial symptom in attention-deficit hyperactivity disorder (ADHD), which is the most prevalent childhood psychiatric disorder (Barkley, 1997). Motor inhibition deficits in children with ADHD might involve the functional states of their prefrontal brain cortices (Strik et al., 1998; Sergeant et al., 2002). In this study, we focused our interest on the neurocognitive process of motor execution/inhibition in several age groups of children with ADHD, as well as normal controls, under CPT-AX with different intervals between stimulus signals. 2. Subject and methods Ten normal adults (7 males, 3 females, mean age 25.2 years, age range 22.7–32.2 years), and 28 normal

4 children participated. The children were divided into a 9-year-old group (4 males, 4 females, mean age 9.0 years, age range 8.7–9.5 years), an 11-year-old group (7 males, 2 females, mean age 11.3 years, age range 10.10–11.10 years), and a 13-year-old group (10 males, 1 female, mean age 13.0 years, age range 12.1– 13.10 years). Twenty-seven children with ADHD also participated. The subjects with ADHD were divided into three age groups, a 9-year-old group: N = 12, mean age 9.6 years, age range 8.4–9.10 years, an 11-year-old group: N = 12, mean age 10.11 years, age range 10.6– 11.7 years, and a 13-year-old group: N = 12, mean age 13.5 years, age range 11.10–15.0 years, all males. All children were diagnosed with ADHD combined type by DSM-IV. All children with ADHD were medication free for at least 24 h prior to testing. In children with ADHD, 11 children in the 11-year-old group participated in experiments under pre- and postmethylphenidate medication. Event-related potentials (ERPs) were initially recorded medication free (nonmedicated ADHD). After the initial recording of ERPs, methylphenidate was medicated. One hour after the administration, ERP recording was repeated (medicated ADHD). In CPT-AX, subjects were asked to press a button when “9” was presented immediately after “1.” To maintain temporal uncertainty in the stimulus series, we implemented three different inter-stimulus intervals (ISIs) between the warning stimulus and the subsequent target and the non-target (no-go) with preceding warning. The probability of the target and non-target was 5% under each ISI condition. A series of 400 digits was presented per block and two blocks were tested. EEGs were recorded from 17 locations on the scalp against linked earlobes as a reference. Vertical EOG was also monitored. EEG was filtered with a bandpass of 0.05–30 Hz and digitized at 500 Hz. Trials with artifacts of 80 μV or more were excluded from further analysis. Trials with response failure or with false alarms were also excluded. ERPs were averaged for target and no-go in each ISI condition. To examine the time course of the brain electric field, global field power (GFP), global dissimilarity and centroid locations (positive and negative) were computed using an

average reference (Lehmann and Skrandies, 1980; Wakkermann et al., 1993). 3. Results 3.1. ERPs during CPT-AX with different ISIs in adults Regardless of target or no-go conditions, N1 and P3 components were clearly observed in all ISI conditions. P3 in the target condition was observed dominantly in the parietal area. On the other hand, P3 in the no-go condition was observed dominantly in the precentral area. ERP waveforms, GPF curves, and positive and negative centroid locations of N1 and P3 components are shown in Fig. 1. The negative centroid location of the N1 component due to the target was similar to that due to no-go stimuli. On the other hand, the positive centroid of the P3 component due to the target was located in the central area. However, the positive centroid of the P3 component due to no-go was located more anteriorly. 3.2. Developmental change of motor behavior in CPT-AX with different ISIs Adaptive segmentation of grand mean ERP yielded four different successive segments under each ISI condition, i.e. P1, N1, P2/N2 with posterior positivity and anterior negativity, and P3. ERP peak maps of the P2/N2 component are shown in Fig. 2A. The amplitude of P2/N2 fell according to development. However, the distribution of no-go P3 (Fig. 2B) significantly shifted to a more anterior area, due to prolonged ISI conditions (Okazaki et al., 2004). 3.3. Neurocognitive process of motor control in children with ADHD during CPT-AX under various ISI The ERP map series in ADHD is shown in Fig. 3. Regardless of the ISI condition, P2/N2 and P3 intensities due to the target in ADHD were lower than those in the control, especially in the younger group (Fig. 3A). However, the intensity of these ERP components in

5

Fig. 1. ERP waveforms, GFP curves, and centroid locations of N1 and P3 components due to target and no-go stimuli.

elder ADHD was enhanced comparable to the control. In the no-go condition (Fig. 3B), P2/N2 and P3 intensities in ADHD were also lower than those in the control. However, an anterior shift of no-go P3 was also observed in the elder ADHD group, comparable to the elder control group. 3.4. ERP during CPT-AX with and without medication in children with ADHD P2/N2 and P3 components were also observed in ERPs with medication. However, P2/N2 intensity increased due to medication and became comparable

to that in the controls. Due to medication, P3 intensity improved. 4. Discussion The fate of ERP spatio-temporal features in adult subjects depends on the mode of motor behavior, i.e. P3 due to target is dominant in the centro-parietal area and P3 due to no-go is dominant in the centro-frontal area (no-go P3). Prolonged ISI induces more anterior distribution of no-go P3, which indicates motor inhibition. Therefore, different cerebral structures might be activated according to whether the motor set prepared

6

A: P2/N2

B: P3

No-go Short

Medium

Long (ISI)

9 years

Short

No-go Medium Long (ISI)

9 years

11 years

11 years

13 years

13 years −10.0

+15.0 (μV)

Fig. 2. ERP peak maps of P2/N2 (A) and P3 (B) at the GFP peak time due to no-go stimulus for each ISI condition. (Reprinted with permission from Okazaki et al., 2004.)

was released or inhibited. However, ERP spatiotemporal features in children are rather different from those in adults, i.e. P3 due to no-go in younger children tends to distribute dominantly in the posterior brain area, and according to maturation, the dominant distribution of no-go P3 shifted toward a more anterior

area. Therefore, the developmental course of motor control might be concerned with the automation of orientation and evaluation of stimulus relevance. Maturational change of the connective network between anterior and posterior cortices is needed for efficient motor control.

Fig. 3. ERP map series from 300 to 500 ms (per 20 ms) due to target (A) and no-go (B) stimuli (medium ISI).

7 ERP in ADHD was characterized by a lower intensity of P2/N2 and P3, suggesting under-activation at a relatively early stage of visual information processing in ADHD. Such under-activation of early processing might disturb the acquisition of appropriate motor behavior. These mal-acquisitions might be concerned with insufficient interaction between anterior and posterior cerebral structures. Significant improvement of ERP intensity was brought about by methylphenidate medication for ADHD. Such ERP enhancement implies that the medication contributed to the allocation of resources to orientate their attention to the relevant stimulus, and also to inhibit later processing for irrelevant stimulus. 5. Acknowledgements We wish to thank Satoshi Futakami (DevelopmentalBehavioral Pediatrics, Izu Iryou Fukushi Center) for giving us valuable advice and comments in the preparation of this study. References Ballard, J.C. (2001) Assessing attention: comparison of responseinhibition and traditional continuous performance test. J. Clin. Exp. Neuropsychol., 23: 331–350. Barkley, R.A. (1997) ADHD and the Nature of Self-Control. Guilford Press, New York. Corkum, P.V. and Siegel, L.S. (1993) Is the continuous performance task a valuable research tool for use with children with attentiondeficit hyperactivity disorder? J. Child Psychol. Psychiatry, 34: 1217–1239.

Fallgatter, A.J., Brandeis, D. and Strik, W.K. (1997) Robust assessment of the no-go anteriorisation of P300 microstates in a cued continuous performance test. Brain Topogr., 9: 295–302. Greenberg, L.M. and Waldman, I.D. (1993) Developmental normative data on the test of variables of attention (TOVA). J. Child Psychol. Psychiatry, 34: 1019–1030. Lehmann, D. and Skrandies, W. (1980) Reference-free identification of components of checkerboard-evoked multichannel potential fields. Electroencephalogr. Clin. Neurophysiol., 48: 609–621. Mesulam, M.-M. (1981) A cortical network for directed attention and unilateral neglect. Ann. Neurol., 10: 309–325. Okazaki, S., Hosokawa, M., Kawakubo, Y., Ozaki, H., Maekawa, H. and Futakami, S. (2004) Developmental change of neurocognitive motor behavior in a continuous performance test with different interstimulus intervals. Clin. Neurophyiol., 115: 1104–1113. Overtoom, C.C.E., Verbaten, M.N., Kemner, C., Kenemans, J.L., Van Engeland, H., Buitelaar, J.K. et al. (1998) Associations between event-related potentials and measures of attention and inhibition in the continuous performance task in children with ADHD and normal controls. J. Am. Acad. Child Adolesc. Psychiatry, 37: 977–985. Rosvold, H.E., Mirsky, A.F., Sarason, I., Bransome Jr., E.D. and Beck, L.H. (1956) A continuous performance test of brain damage. J. Consult. Psychol., 20: 343–350. Sergeant, J.A., Geurts, H. and Oosterlaan, J. (2002) How specific is a deficit of executive functioning for attention-deficit/ hyperactivity disorder? Behav. Brain. Res., 130: 3–28. Strik, W.K., Fallgatter, A.J., Brandeis, D. and Pascual-Marqui, R.D. (1998) Three-dimensional tomography of event-related potentials during response inhibition: evidence for phasic frontal lobe activation. Electroencephalogr. Clin. Neurophysiol., 108: 406–413. Van Leeuwen, T.H., Steinhausen, H.-Ch., Overtoom, C.C.E., Pascual-Marqui, R.D., Van’t Klooster, B., Rothenberger, A. et al. (1998) The continuous performance test revisited with neuroelectric mapping: impaired orienting in children with attention deficits. Behav. Brain Res., 94: 97–110. Wakkermann, J., Lehmann, D., Michel, C.M. and Strik, W.K. (1993) Adaptive segmentation of spontaneous EEG map series into spatially defined microstates. Int. J. Psychophysiol., 14: 269–283. Williams, B.R., Ponesse, J.S., Schachar, R.J., Logan, G.D. and Tannock, R. (1999) Development of inhibitory control across the life span. Dev. Psychol., 35: 205–213.


9

Chapter 2

Studying higher cerebral functions by transcranial magnetic stimulation Yasuo Terao* and Yoshikazu Ugawa Department of Neurology, Division of Neuroscience, Graduate School of Medicine, University of Tokyo, Tokyo 113-8655 (Japan)

1. What is the virtual lesion paradigm? Transcranial magnetic stimulation (TMS) is known to exert both excitatory and inhibitory effects on the cerebral cortex. While one major use of TMS is to elicit motor evoked potentials from the motor cortex (i.e. example of an excitatory effect), the inhibitory effect may be used to investigate the functions of cortical areas other than the motor cortex, from which no overt response can be elicited. Penfield and colleagues (Penfield and Rasmussen, 1949; Penfield and Roberts, 1959) explored the cortical surface by intraoperative electrical stimulation and found vast cortical regions whose functions cannot be revealed by electrical stimulation unless they are involved in a task that is currently being performed, which they named the elaborative cortex. Unfortunately, most cortical areas are “elaborative cortices” also in terms of TMS. In this chapter, we describe the use of TMS to study the function of nonmotor and non-visual cortical areas from which no overt stimulation signatures can be evoked. *Correspondence to: Yasuo Terao, M.D., Ph.D., Department of Neurology, Division of Neuroscience, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Tel: +81-3-3815-5411, ext. 33784; Fax: +81-3-5800-6548; E-mail: [email protected]

Before the advent of TMS, most investigations on the functions of elaborative cortices relied on lesion studies in neuropsychology. The purpose of lesion studies is to establish a correlation between a circumscribed region of damaged brain and changes in some aspects of an experimentally controlled behavioral performance. In a similar manner, TMS can produce a “virtual lesion” by adding noise to cortical processing. Using a figure-ofeight coil, such “lesions” can be made focal. In the virtual lesion paradigm, we applied TMS to focal regions of the brain to temporarily interfere with local information processing and observe the resultant behavioral changes, e.g. a change in reaction time (RT) in a RT task. Functional mapping of the brain is performed by assuming that if the performance of a task is delayed or facilitated by TMS at a certain time, the focal cortical area just underneath the coil is then active and necessary. If we plot the delay of RT induced by TMS as a function of the stimulus location, we would obtain a “functional map” which shows the location where TMS affects RT. By comparing the effect of TMS at various time periods, we can see how the physiological activities of those cortical regions evolve with time. Aspects of cognitive and higher functions addressed by the virtual lesion paradigm are not necessarily different from those that can be addressed by neuropsychology or other neuroimaging techniques. However, by

10 using the virtual lesions paradigm, a clear chain of cause and effect between the activity of the brain and behavior can be established; if you stimulate a certain cortical region and observe a resultant behavioral change, you can be certain that the “stimulated” cortical region is necessary to induce that behavior. In addition, another advantage of the virtual lesion paradigm is that you can produce lesions “wherever you want” without the confound of reorganization. Furthermore, the virtual lesion paradigm has a time resolution that lesion studies in neuropsychology lack. In other words, it is capable of realizing “real time neuropsychology” (Walsh and Pascual-Leone, 2003).

A

Central spot Target spots Video camera

Switch button

Controller AD convertor

DC amplifier

TV monitor

2. Visualization of the information flow through human oculomotor cortical regions by TMS Microcomputer

B

Here, we describe an attempt at functional mapping of the brain by TMS to visualize the information flow through oculomotor cortical regions during the performance of an antisaccade task (Terao et al., 1998). The subjects are seated in front of a dome-shaped screen of 90-cm diameter, with light-emitting diodes (LEDs) embedded in horizontal, vertical, and oblique arrays (Fig. 1A). The targets are presented as LEDs, and the magnetic stimulator can be triggered at a programmed time relative to their presentation. A 1–2 cm grid reference system covering the skull was constructed over a plastic cap worn by each subject, and TMS is delivered over each grid point while the subjects perform an oculomotor task (Fig. 1B). The most commonly used oculomotor task is the visually guided saccade task (Fig. 2A, top). In this paradigm, a fixation spot is presented at the center of the dome, on which the subjects have to fixate. After a random interval, a target is presented to the left or right of it, at the same time as the fixation spot goes off, and the subject is required to make a saccade toward it. We used the antisaccade paradigm in our experiments. In this paradigm, the target presentation is identical, but the subjects are required to make a saccade of the same size but in the opposite direction (Fig. 2A, bottom). Therefore, to produce an antisaccade, the subjects have to suppress the reflexive prosaccade made toward the presented target, and at the same time generate a saccade in a symmetrical position.

Fig. 1. (A) Experimental setup for oculomotor tasks. (B) The subjects wore a plastic cap over which a grid system was constructed. TMS was applied over each of the grid points while the subjects performed a RT task.

Without TMS, antisaccades begin 200–250 ms after the target presentation (Fig. 2B). Delivered just before the expected saccade onset, TMS delays the latency by up to 50 ms over some scalp locations. As mentioned, the delay is assumed to occur because TMS interferes with cortical processing occurring at that time. Having the subjects perform the task while stimulating various scalp locations, we obtain a map showing the locations where TMS effectively delays

11 A

F

Target: Lt & Rt ; 20deg

Prosaccade

Target delay: 0 ms

T

Antisaccade F T

B

Without TMS

Angle

Target on Expected onset of saccade

TMS With TMS

Delay induced by TMS

Target on

Time Fig. 2. (A) The visually guided saccade (prosaccade) and antisaccade paradigm. In both paradigms, the targets were randomly presented 20° to the left and right of the fixation spot. (B) Schema of the TMS experiments. TMS applied just before the expected onset of saccade delays the saccade reaction time (lower trace) as compared to when no TMS is applied (top trace). The horizontal axis of the traces gives the time, and the vertical axis gives the angle.

saccade onset. Figure 3A is a map, color-coded according to the delay of saccade onset induced by TMS at each location. With TMS delivered at 80 ms after the target presentation, the maximal delay was induced over posterior scalp regions of bilateral hemispheres, 6–8 cm posterior to hand motor areas, and somewhat lateral, presumably covering the posterior parietal cortices (PPC). At 100 ms, the maximal delay was induced over the frontal regions, 2–4 cm anterior and 2–4 cm lateral to the hand motor areas bilaterally. These were considered to represent the frontal eye field (FEF) and perhaps also some part of the dorsolateral prefrontal cortex. Although not shown here, TMS at 120 ms did not result in any significant change of saccade latency over any of the regions. Therefore, there was information flow through human oculomotor cortical regions during the presaccadic period, i.e. from posterior (e.g. PPC) at an earlier interval (80 ms) to anterior cortical regions (e.g. FEF, dorsolateral prefrontal cortex) at a later interval (100 ms). TMS was also effective in inducing reflexive prosaccades in the direction of the target that should not be made (Fig. 3B). Prosaccades were induced significantly more frequently than the baseline when TMS was applied to the contralateral hemisphere. For example, as shown in this figure, rightward prosaccades were induced over the left PPC and FEF at 80 ms, and over the left FEF at 100 ms. The mechanism of antisaccade generation can be explained by a scheme shown in Fig. 4A. Let us consider a case in which the target is presented to the right. The visual input in the right hemifield reaches the left primary visual cortex by 40–60 ms, and is then transmitted to the parietal cortex by 80 ms. In the case of antisaccade, parietal information is sent to the corresponding region in the opposite (in this case, the right) hemisphere. At 100 ms, the bilateral information is sent to the frontal cortex including the FEF via cortico-cortical connections. The final saccadic motor output is sent out from the right FEF to produce a saccade to the left, opposite to the visual stimulus. For the bilateral effect of TMS to occur, there must be an interhemispheric transfer of information. This transfer may be mediated by callosal fibers, which are known to connect the FEFs and PPCs of both hemispheres.

A

Lt-sided view

Delay of RT (ms)

35

Rt-sided view

-20

Lt-sided view

10

Rt-sided view

TMS at 80ms

TMS at 100ms

For legend see opposite page

Increase in error rate (%)

B

0

13 Fig. 3. Spatiotemporal shift of regions affecting saccade latency (A) and eliciting erroneous prosaccades. The delay in antisaccade latency (A) or increase in the error rate (B) induced by TMS was plotted (z axis) against the site (x–y axis), which was then transformed into a contour map. The map is shown as viewed from two sides. White dots mark the positions of the bilateral hand motor areas, and the white cross indicates the position of the vertex. The maps are color coded to help identify the effective regions; the red areas indicate regions where saccade delay or elicitation of prosaccades was maximal, and the green and blue areas indicate regions where these were minimal. For cues, see the panel linked to the right-hand side of each figure.

A similar but somewhat different cortical information flow can be proposed for the mechanism of reflexive prosaccades (Fig. 4B). Antisaccades require the subject to inhibit prosaccade to the presented target, and instead to generate a saccade in the opposite direction. TMS presumably interferes with such a dual process, leading to an increased incidence of reflexive prosaccades. This time, the distribution of effective regions was unilateral, i.e. contralateral to the side of the visual stimulus, and no interhemispheric transfer of information was required. 3. Cortical motor preparation for human vocalization The same mapping procedure can be used to investigate the cortical preparation for vocalization. Penfield and Rasmussen (1949) could induce speech arrest by intraoperatively stimulating the lip–face motor representation of both hemispheres. To see the effect of TMS on vocalization, we asked the subject to produce a Japanese vowel sound “a” quickly in response to a visual signal (Fig. 5A). TMS was applied to various locations over the motor strip of bilateral hemispheres just before the expected onset of voice, which, without TMS, occurred at about 300–350 ms after the visual signal (Fig. 5B). A significant delay in voice onset was noted when TMS was delivered 0–150 ms before the expected onset, and the hotspot of the TMS effect was over locations 6 cm to the left and right of Cz irrespective of the time of stimulation (Fig. 5C and 5D). These locations correspond to the face or lip representation of the motor cortex of both hemispheres, consistent with the findings of Penfield et al. (1949, 1959).

We then investigated the time courses of TMS effect over the left and right face motor areas (Fig. 6A). There were two distinct phases in the prevocalization period; 50–150 ms before the expected onset, and a later phase, i.e. 0–50 ms before. Plotting the delay of RT against the time of TMS, the effect of TMS was larger over the left than over the right hemisphere 50–150 ms before the expected onset of voice, whereas this lateralization was reversed when TMS was applied just before vocalization onset, i.e. 0–50 ms before the expected onset of voice. This suggests that, during the cortical preparation for human vocalization, hemispheric lateralization alternates between the two motor cortices, with mild left hemispheric predominance at the early phase switching over to robust right hemispheric predominance during the late phase. A magnetoencephalographic study by Gunji et al. (2000) also demonstrated the involvement of bilateral motor areas during cortical motor preparation for vocalization. Although the authors did not mention it explicitly, activation of the left motor area precedes that of the right motor area, which is taken over by the predominant activity of the right motor area just before voice onset. In primates, the periaqueductal gray serves as a bottleneck region for vocalization that receives all the descending inputs from supraspinal centers, including the cerebral cortex, and relays these to the phonatory motoneuron pools located in the medulla and spinal cord (Jürgens, 1998). If we postulate a similar pathway for humans, the cortical preparation for vocalization may be considered a process through which motor buffer is formed within the bilateral motor cortices (motor programming phase) and released to relevant brainstem centers (motor output phase). We considered the early

14

A

Target

80ms

Saccade direction

100ms FEF

Callosal transmission

FEF Callosal transmission

Cortico-cortical connections

PPC

PPC V1

Antisaccades Target

B

80ms

Saccade direction

100ms FEF FEF

PPC

Cortico-cortical connections

PPC

V1

Elicitation of prosaccades Fig. 4. Schematic diagram showing the cortical processing presumed to occur during antisaccades (A) and elicited prosaccades (B). Black ellipses represent the active areas at each time shown in the top left-hand corner, and gray areas indicate regions whose activities are low or have subsided.

and late phases of the prevocalization period each representing the motor programming and motor output phases of vocalization. Given that bilateral motor areas are active during the cortical motor preparation for vocalization, what happens when both motor areas are stimulated simultaneously? We compared the effect of unilateral vs. bilateral TMS delivered over the left and right motor areas (Fig. 6B). During the period preceding the expected onset by 50–150 ms (early phase), the delay induced by unilateral TMS and bilateral TMS was almost identical. In comparison, during the period 0–50 ms before the

expected onset of voice, bilateral TMS induced a significantly larger delay than that induced by unilateral TMS (late phase). If TMS mainly interfered with motor buffer formation, the cortical process for this formation may be slowed, but the buffer itself would not disappear. Once formed and ready in both hemispheres, the motor buffer is released to relevant brainstem centers, so that the effect of TMS is not significant even if delivered bilaterally. If bilateral TMS delays buffer formation in both hemispheres by the same amount of time Δt, RT would also be delayed by Δt, because in both

15 A

B

Visual signal

Sites of stimulation

-130ms

A B C D E F Cz

C

G

3cm

%

30

20

-80ms

10 0

D

%

20 15

-30ms

10 5 0

-10

Left

-5

5

0

Cz

10 cm

Right

Site

control

-50

100 ms/div

0ms

-100

50ms

-150

100ms

-200

150ms

Fig. 5. An example of voice recordings (A) and sites of TMS over the scalp (B). (A) Each trace shows the superimposition of voice recordings for 10 trials. The bottom trace is a recording when the magnetic coil was delivered off the scalp, but the subject heard the clicking sound accompanying the magnetic pulse. The time of visual cue presentation is indicated by a vertical solid line, and the control reaction time is marked by a vertical dashed line. The time of TMS delivery is indicated by white triangles. In the top three traces, TMS was applied ~130, 80, and 30 ms before the expected onset of voice. The onsets of voice (marked by black triangles) were progressively more delayed in comparison with the control reaction time when TMS was applied at a later interval. (B) The figure-of-eight coil was placed over points spanning the motor strip, namely Cz, (point D), and points 3 cm to the left and right (points C and E), points 6 cm to the left and right (points B and F), and points 9 cm to the left and right (points A and G). The effect of TMS over the motor strip in one subject (C) and all subjects (D). The RT delay was plotted as a function of the site where focal TMS was delivered over the motor strip. The four curves in the top figure depict the delay when TMS was applied 0–50, 50–100, 100–150, and 150–200 ms before the expected onset of voice. In this and the following figures, the delay is expressed as a percentage of the control reaction time in the same session.

16 A

B

%

% 30

25 Rt motor area

unilateral TMS

Lt motor area

20

bilateral TMS

20

15 10

10

5 0

0 -5 -200

-150

-100

-50

0

-10 -200

50 ms

-150

-100 -50 time of TMS

0 ms

Without TMS

C

TMS over rt hemisphere

TMS over both hemispheres

TMS over It hemisphere

Fig. 6. (A) Comparison of the effects when TMS was applied over the left (gray curve) and right motor areas (black curve). (B) Comparison of the effects of unilateral (filled dots) and bilateral TMS (white circles). Dots represent unilateral TMS, whereas circles stand for bilateral TMS. In both these figures, the abscissa gives the time of TMS relative to the expected onset of voice, and the ordinate gives the delay induced by TMS. (C) Schematic illustration of the possible TMS effect when it was applied over the right or left hemisphere or bilaterally over both motor areas at the same time.

hemispheres, buffer formation is completed with delayed Δt. Unilateral TMS may delay buffer formation in the stimulated hemisphere by time Δt, but not in the unstimulated hemisphere. Here again, RT may be delayed by the same amount of time Δt, if we postulate that the motor buffers in both hemispheres should be completed for them to be released for motor output. TMS during the early phase may have interfered

mainly with buffer formation, because the induced delay was nearly identical to unilateral or bilateral TMS. On the other hand, if TMS blocked the release of motor buffer into relevant brainstem centers, bilateral TMS would induce RT delay well in excess of that induced by unilateral TMS. This is because bilateral TMS would abolish the descending commands from

17 both hemispheres, greatly reducing the motor output, whereas unilateral TMS would spare at least the motor output from the unstimulated hemisphere (Fig. 6C). Thus, we reasoned that the late phase was mainly dedicated to the release of motor commands into relevant brainstem centers. 4. Summary TMS can be used to study higher cerebral functions by the virtual lesion paradigm. The major advantages of this method are that it could be used to produce a lesion anywhere the researcher wants without confusing cortical reorganization, and that it helps to establish a chain of cause and effect between the activity of the brain and behavior. With elucidation of the mechanism underlying the cortical function blocking, this technique will open up new possibilities for studying higher cerebral functions. In contrast to the online method in which TMS is delivered while subjects perform a certain task, the off-line method uses repetitive TMS to achieve lasting effects even after stimulation has ceased. The application of the offline method will extend from improving cognitive

functions by TMS to the treatment of neurological and psychiatric patients. References Gunji, A., Kakigi, R. and Hoshiyama, M. (2000) Spatiotemporal source analysis of vocalization-associated magnetic fields. Cogn. Brain Res., 9: 157–163. Jürgens, U. (1998) Neuronal control of mammalian vocalization, with special reference to the squirrel monkey. Naturwissenschaft, 85: 376–388. Penfield, W. and Rasmussen, T. (1949) Vocalization and arrest of speech. Arch. Neurol. Psychiatry, 61: 21–27. Penfield, W. and Roberts, L. (1959) Speech and Brain Mechanism. Princeton University Press, Princeton, NY. Terao, Y., Fukuda, H., Ugawa, Y., Hikosaka, O., Hanajima, R., Furubayashi, T., Sakai, K., Miyauchi, S., Sasaki, Y. and Kanazawa, I. (1998) Visualization of the information flow through human oculomotor cortical regions by transcranial magnetic stimulation. J. Neurophysiol., 80: 936–946. Terao, Y., Ugawa, Y., Enomoto, H., Furubayashi, T., Shiio, Y., Machii, K., Hanajima, R., Nishikawa, M., Iwata, N.K., Saito, Y. and Kanazawa, I. (2001) Hemispheric lateralization in the cortical motor preparation for human vocalization. J. Neurosci., 21: 1600–1609. Walsh, V. and Pascual-Leone, A. (2003) Transcranial Magnetic Stimulation – A Neurochronometrics of the Mind. MIT Press, Cambridge, MA.


19

Chapter 3

Long-term potentiation (LTP)-like plasticity and learning in human motor cortex – investigations with transcranial magnetic stimulation (TMS) Ulf Ziemann*, Tihomir V. Ilić and Patrick Jung Motor Cortex Laboratory, Department of Neurology, Johann Wolfgang Goethe-University of Frankfurt, D-60528 Frankfurt am Main (Germany)

1. Introduction Motor learning occurs in humans throughout life, for instance during acquisition of a new motor skill, or during adaptation to a changing environmental situation, or during re-learning a lost motor function after brain lesion. The mechanisms of motor learning are not entirely clear though. It is currently thought that synaptic plasticity in the form of long-term potentiation (LTP) and long-term depression (LTD) is an important candidate mechanism of motor learning (Donoghue et al., 1996; Sanes and Donoghue, 2000). LTP and LTD are typically induced by particular stimulation protocols (Hess and Donoghue, 1996; Hess et al., 1996). If motor learning is an LTP-dependent process, then it should interact with stimulation-induced LTP. Several theories predict the ways of this interaction. The most influential

*Correspondence to: Prof. Ulf Ziemann, Department of Neurology, Johann Wolfgang Goethe-University of Frankfurt, Schleusenweg 2–16, D-60528 Frankfurt am Main, Germany. Tel: +49 69 6301 5739; Fax: +49 69 6301 6842; E-mail: [email protected]

theory is the Bienenstock–Cooper–Munro (BCM) theory (Bienenstock et al., 1982). The key assumptions of this theory are that (i) Active synapses are bi-directionally modifiable, i.e. they can express LTP and LTD; (ii) The sign of this modification (LTP or LTD) depends on the level of the postsynaptic activity relative to a modification threshold; and (iii) The value of this modification threshold is not set but varies with the history of postsynaptic activity, i.e. the threshold for LTP induction increases, and at the same time the threshold for LTD induction decreases, with the mean level of previous postsynaptic activity (Bienenstock et al., 1982). The BCM theory provides an important conceptual framework for stabilizing activity in neuronal networks (Abbott and Nelson, 2000). If motor learning is an LTP-dependent process, then, according to the BCM theory, it should result in a decrease of subsequent stimulation-induced LTP but an increase in subsequent stimulation-induced LTD. This prediction was recently proven correct in rat motor cortex (RioultPedotti et al., 1998, 2000). Rats learned a new motor skill by practising the retrieval of food pellets from a

20 box by their preferred forelimb. Success rate improved significantly through the first three training days. Thereafter, rats were sacrificed and field potentials were elicited across horizontal layer II–III connections in slices of the trained and non-trained primary motor cortex (M1). Field potentials were larger in the trained compared to non-trained M1 (Rioult-Pedotti et al., 1998). In addition, induction of LTP was depressed while induction of LTD was enhanced in the trained vs. non-trained M1 (Rioult-Pedotti et al., 2000). These findings are consistent with the BCM theory, and for the first time, provided strong experimental evidence that motor learning is an LTP-dependent process. The objective of the present experiments is to demonstrate that a similar interaction between motor learning and LTP/LTD also occurs in the human M1. In experiment 1, we explored the effects of motor learning on subsequent associative LTP/LTD-like plasticity in an experimental design similar to the experiments by Rioult-Pedotti et al. (1998, 2000). These data have been published recently (Ziemann et al., 2004). Since the findings strongly suggested that motor learning is an LTP-dependent process in humans, it follows according to the BCM theory that stimulation-induced LTP should reduce subsequent motor learning whereas stimulationinduced LTD should enhance it. If true, this would be a way to purposefully modify learning. Therefore, in experiment 2, we reversed the order of events in experiment 1 by exploring the effects of LTP/LTD-like plasticity on subsequent motor learning. Experiment 2 is in progress and only preliminary data from single subjects will be shown.

2. Methods 2.1. Experiment 1 Twelve naive right-handed healthy subjects (24–42 years, 4 females) participated after giving written informed consent. Approval was obtained from the local ethics committee. Each subject participated in five experiments of fixed order, with consecutive experiments separated by at least one week: the first two sessions applied paired associative stimulation alone

(PAS alone), i.e. protocols to induce LTP/LTD-like plasticity in human M1 (Stefan et al., 2000; Wolters et al., 2003), the third session employed motor practice alone (MP alone) (Muellbacher et al., 2001), and the last two sessions assessed the interactions between MP and subsequent PAS (MP + PAS). PAS consisted of 200 electrical stimuli to the right median nerve at the wrist paired with focal transcranial magnetic stimulation (fTMS) of the hand area of the left M1 optimal for eliciting motor evoked potentials (MEPs) in the right abductor pollicis brevis (APB) muscle (rate of paired stimulation, 0.25 Hz). One of two interstimulus intervals (ISIs) between median nerve stimulation and fTMS was applied, either equalling the individual N20 cortical component of the median nerve somatosensory evoked potential (PASN20) to induce an LTP-like increase in MEP amplitude in the APB, or shorter by 5 ms (PASN20−5) to induce an LTD-like decrease (Stefan et al., 2000; Wolters et al., 2003; own unpublished observations). Each subject was assigned randomly with 50% chance and in a double-blind fashion to PASN20 (“LTP group”) or PASN20−5 (“LTD group”). LTP/LTD effects were quantified by comparing MEP amplitude in the resting APB over a period of 30 min after PAS with MEP amplitude before PAS (baseline). At baseline, fTMS intensity was adjusted to elicit MEP amplitudes of 1 mV (MEP1mV) on average. For MP, subjects were seated in a comfortable chair with their right arm adducted in the shoulder, flexed at right angle in the elbow, and the semi-pronated forearm supported by a plate. Forearm, wrist, and fingers II–V were fixed in a cast, leaving the thumb free for movements in all directions. MP consisted, in a slightly modified version of a previous motor-learning protocol (Muellbacher et al., 2001) of fastest thumb abduction movements of the right hand for 30 min, paced by a metronome at a rate of 0.5 Hz. Acceleration of the thumb movements was measured in a two-dimensional plane along the abduction– adduction and flexion–extension axes. Motor learning was quantified by the increase in the first peak acceleration of trained fastest possible thumb abduction movements immediately after MP compared to those before MP. In addition, learning-induced changes in M1 excitability were tested by measuring MEP amplitude in the APB.

21 2.2. Experiment 2 Nine naive right-handed healthy subjects (22–40 years, 2 females) participated. The experiments reversed the order of events in experiment 1, i.e. PAS preceded motor learning. Each subject was tested in three different sessions of randomized and counter-balanced order at least two weeks apart, using one of three different ISIs for PAS: N20 plus 2 ms (PASN20 + 2) to induce an LTP-like increase in MEP amplitude in the flexor pollicis brevis (FPB) muscle, N20 minus 5 ms (PASN20−5) to induce an LTD-like decrease, or 100 ms (PAS100) to induce no change (control experiment). These intervals were derived from previously published protocols (Stefan et al., 2000; Wolters et al., 2003) and own unpublished observations. Otherwise, the details of the PAS protocols followed the specifications given under experiment 1. The PAS induced change in MEP amplitude was determined by the ratio of MEPs measured immediately after PAS relative to those before PAS. MP consisted of two blocks (15 min)

A

of fastest possible thumb flexion movements externally paced by a metronome at a rate of 0.25 Hz. Motor learning was quantified by comparing the maximum first peak acceleration of the trained thumb flexion movement measured between the two blocks of MP, and over a period of 30 min after the second block of MP with the peak acceleration before MP. 3. Results 3.1. Experiment 1 PASN20 alone resulted in an LTP-like increase in MEP amplitude in the APB in both sessions (session 1: 0.88 ± 0.24 mV → 1.36 ± 0.39 mV, P = 0.005; session 2: 0.92 ± 0.24 mV → 1.47 ± 0.44 mV, P = 0.015, black bars in Fig. 1A). In contrast, PASN20−5 alone resulted in an LTD-like decrease of MEP amplitude in both sessions (session 1: 1.03 ± 0.32 mV → 0.73 ± 0.33 mV, P = 0.021; session 2: 1.01 ± 0.30 mV → 0.75 ± 0.17 mV, P = 0.011, black bars in Fig. 1B). MP alone resulted in

B

Fig. 1. (A) Long-term changes in MEP amplitude in the APB induced by PASN20 alone (two sessions, black bars), MP preceding PASN20 without (MP + PASN20, grey bar) and with adjustment of MEP amplitude to MEP1mV after MP (control, white bar). MEP amplitudes were measured for 30 min after PASN20 and compared to the baseline MEP amplitude immediately before PASN20 (amplitude ratio, y axis). All data are means (n = 6 subjects) + 1 SEM. *indicate LTP-like effects (MEP ratio > 1.0, P < 0.05), ‡ and ¶ indicate differences from the first and second sessions of PASN20 alone, respectively (P < 0.001). (B) Long-term changes in MEP amplitude in the APB induced by PASN20−5 alone (two sessions, black bars), MP preceding PASN20−5 without (MP + PASN20−5, grey bar) and with adjustment of MEP amplitude to MEP1mV after MP (control, white bar). All data are means (n = 5 subjects) + 1 SEM. *indicate LTD-like effects (MEP ratio < 1.0, P < 0.05), § and # indicate differences from the first and second sessions of PASN20−5 alone, respectively (P < 0.05). Note that MP suppressed subsequent LTP-like plasticity but enhanced subsequent LTD-like plasticity.

22 significant learning as measured by an increase in the peak acceleration of fastest possible voluntary thumb abduction movements from 8.44 ± 4.83 to 13.02 ± 8.42 m/s2 (P = 0.011) without differences between the subjects in the LTP and LTD group (P = 0.59). Learning was accompanied by an increase in MEP amplitude in the APB from 1.04 ± 0.23 to 1.32 ± 0.40 mV (P = 0.010), again without differences between the subjects in the LTP and LTD group (P = 0.60). The LTP-like increases in MEP amplitude in the APB induced in the two PASN20 alone experiments (black bars, Fig. 1A) were abolished, even with a nonsignificant trend towards MEP depression, when motor learning immediately preceded PASN20 (grey bar, Fig. 1A). In contrast, the LTD-like decreases in MEP amplitude induced in the two PASN20−5 alone experiments (black bars, Fig. 1B) were enhanced if motor learning immediately preceded PASN20−5 (grey bar, Fig. 1B). These interactions were not explained by the learning induced increase in MEP amplitude in the APB because a control experiment with lowered intensity of fTMS to correct for this increase (reinstallation of MEP1mV) showed a very similar suppression of LTP-like plasticity when learning preceded PASN20 (white bar, Fig. 1A), and a similar enhancement of LTD-like plasticity when learning preceded PASN20−5 (white bar, Fig. 1B), compared to the main experiments without adjustment of fTMS intensity (grey bars, Fig. 1). 3.2. Experiment 2 At the time of writing this manuscript, not all subjects had completed the study. Therefore, only data from two individual subjects are shown (Fig. 2). Both subjects exhibited a clear increase in MEP amplitude of the FPB after PASN20 + 2, a clear decrease after PASN20–5, and not much change in the control condition after PAS100 (Fig. 2A and 2C). The different PAS conditions strongly influenced subsequent motor learning. In both subjects, PASN20−5 enhanced motor learning when compared to PAS100 (Fig. 2B and 2D). PASN20 + 2 had more variable effects. In one subject PASN20 + 2 tended to suppress learning when compared to PAS100 (Fig. 2B), while in

the other subject, it slightly enhanced learning, but less so than PASN20−5 (Fig. 2D). 4. Discussion The main finding from experiments 1 and 2 is that associative LTP/LTD-like plasticity and motor learning in human cortex interact in specific ways, consistent with the Bienenstock–Cooper–Munro (BCM) theory of bi-directional synaptic plasticity (Bienenstock et al., 1982). Two important conclusions follow from these findings: (1) LTP-like plasticity is a mechanism of motor learning in humans, consistent with previous evidence from animal preparations (Donoghue et al., 1996; Sanes and Donoghue, 2000; Rioult-Pedotti and Donoghue, 2003); (2) Motor learning is enhanced after LTD-like plasticity. According to the BCM theory, LTP-dependent processes are more likely to occur in the context of a history of low postsynaptic activity. Similar to the present findings, this may explain why motor consolidation and motor retention processes in which the primary motor cortex is specifically involved (Muellbacher et al., 2002), are enhanced by spaced practice. The majority of studies showed that a period of rest of greater than 4 h, or a night of sleep, results in improvement in performance between the first and second training sessions (Karni and Sagi, 1993; Brashers-Krug et al., 1996; Shea et al., 2000; Walker et al., 2002). In addition, it may explain why it is not possible to learn more than one motor task at a time (Brashers-Krug et al., 1996). Evidence for the existence of bi-directional synaptic plasticity consistent with the BCM theory, and underlining its key principle of a sliding modification threshold for LTP/LTD, was first provided in rat visual cortex where it was shown that binocular deprivation shifted the LTP threshold to lower stimulation frequencies of the induction protocol, and that this was reversed by restoring normal vision (Kirkwood et al., 1996). This sliding threshold of LTP/LTD induction was also referred to as metaplasticity, i.e. the plasticity of synaptic plasticity (Abraham and Bear, 1996). One candidate mechanism of the sliding LTP/LTD threshold is an

23 Subject 1 A

B

Subject 2 C

D

Fig. 2. Change in MEP amplitude in subject 1 (A) and 2 (C) induced by three different PAS protocols (x axis) given as MEP amplitude after PAS (P) normalized to baseline (y axis). Motor learning in subject 1 (B) and 2 (D) given as a ratio of maximum peak acceleration of the trained thumb flexion movement after the first block of motor practice (P1), immediately after the second block of motor practice (P2), and 10 min (P3), 20 min (P4), and 30 min (P5) later normalized to the maximum peak acceleration before practice (y axis). The different symbols indicate motor learning after PASN20+2 (black squares), PASN20−5 (grey triangles), and PAS100 (white circles). Note that in both subjects, PASN20−5 enhanced subsequent learning when compared to learning after PAS100.

24 activity-dependent slowing of N-methyl-D-aspartate (NMDA) receptor-mediated excitatory postsynaptic potentials (EPSP) after a period of low postsynaptic activity which in turn results in an increased probability for EPSP summation and enhanced intracellular Ca2+ entry (Philpot et al., 2001). These changes in NMDA receptor kinetics are likely explained by activitydependent changes in the subunit composition of the NMDA receptor (Quinlan et al., 1999). There is now abundant evidence from several other cortical systems in experimental animal preparations that synaptic plasticity depends crucially on the recent history of synaptic and cellular activity. In rat motor cortex, the depression of LTP and enhancement of LTD by prior motor skill learning (Rioult-Pedotti et al., 2000) is an important example consistent with the BCM theory. These findings strongly supported the view that LTP is a mechanism of motor skill learning. In the present experiments, we used a paired associative stimulation protocol (PAS) (Stefan et al., 2000, 2002; Wolters et al., 2003) in order to induce LTP/ LTD-like plasticity in human motor cortex in vivo at the systems level. While it is clear and has to be kept in mind that this experimental approach is indirect compared to experiments in slices, it was shown that this form of LTP/LTD-like plasticity is associative, inputspecific, prevented by NMDA receptor blockade and has a duration > 60 min (Stefan et al., 2000, 2002; Wolters et al., 2003). These are the main characteristics of LTP/LTD as defined in slices (Bliss and Collingridge, 1993). Therefore, other researchers and we believe that LTP/LTD in the slice and PAS-induced long-term potentiation or depression of MEP amplitude share very similar or even identical mechanisms, and that it is therefore justified to refer to the long-term changes in MEP amplitude as LTP/LTD-like plasticity (Classen and Ziemann, 2003; Ziemann, 2004). The acceptance of this argument is a premise for the main conclusion of the present experiments that the specific interactions between PAS-induced plasticity and motor learning support the hypothesis that learning in human motor cortex depends on LTP. Very recently, others have also used transcranial magnetic stimulation (TMS) to provide evidence for metaplasticity in human motor cortex. In one study, an

LTD-like depression of MEP amplitude induced by low-frequency (1 Hz) repetitive transcranial magnetic stimulation (RTMS) was enhanced if preceded by priming higher frequency (6 Hz) RTMS (Iyer et al., 2003). This finding is directly consistent with earlier data obtained in rat hippocampus (Christie and Abraham, 1992) and with the BCM theory. Another study explored the interactions between preconditioning anodal or cathodal transcranial direct current stimulation (tDCS) and the effects of subsequent 1 Hz RTMS (Siebner et al., 2004). Anodal tDCS alone results in a long-term LTP-like increase in MEP amplitude, cathodal tDCS in a long-term LTD-like decrease (Nitsche and Paulus, 2000; Nitsche et al., 2003). In the interaction experiments, anodal tDCS enhanced the LTD-like effect of 1 Hz RTMS, while cathodal tDCS resulted in a switch from an LTD-like to an LTP-like effect of 1 Hz RTMS (Siebner et al., 2004), again consistent with the idea of an activity-dependent sliding threshold of LTP/LTD induction according to the BCM theory. In summary, the current data and other recent studies point out that TMS offers the opportunity to study mechanisms of learning in human motor cortex. The findings support the view that learning depends on LTP and that principles of metaplasticity consistent with the BCM theory regulate learning: it appears that learning performance is enhanced when the probability for LTP induction is high because learning is an LTP-dependent process itself. This knowledge may be utilized for strategies to enhance learning, for instance when patients after brain lesion practice during the process of neurorehabilitation to re-learn lost motor function. 5. Acknowledgements Tihomir V. Ilić was a fellow of the Alexander von Humboldt Foundation. References Abbott, L.F. and Nelson, S.B. (2000) Synaptic plasticity: taming the beast. Nature Neurosci., 3 (Suppl): 1178–1183. Abraham, W.C. and Bear, M.F. (1996) Metaplasticity: the plasticity of synaptic plasticity. Trends Neurosci., 19: 126–130.

25 Bienenstock, E.L., Cooper, L.N. and Munro, P.W. (1982) Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex. J. Neurosci., 2: 32–48. Bliss, T.V. and Collingridge, G.L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature, 361: 31–39. Brashers-Krug, T., Shadmehr R. and Bizzi, E. (1996) Consolidation in human motor memory. Nature, 382: 252–255. Christie, B.R. and Abraham, W.C. (1992) Priming of associative long-term depression in the dentate gyrus by theta frequency synaptic activity. Neuron, 9: 79–84. Classen, J. and Ziemann, U. (2003) Stimulation-induced plasticity in the human motor cortex. In: S.J. Boniface and U. Ziemann (Eds.), Plasticity in the Human Nervous System. Investigation with Transcranial Magnetic Stimulation. Cambridge University Press, Cambridge, pp. 135–165. Donoghue, J.P., Hess, G. and Sanes, J.N. (1996) Substrates and mechanisms for learning in motor cortex. In: J. Bloedel, T. Ebner and S.P. Wise (Eds.), Acquisition of Motor Behavior in Vertebrates. MIT Press, Cambridge, MA, pp. 363–386. Hess, G. and Donoghue, J.P. (1996) Long-term depression of horizontal connections in rat motor cortex. Eur. J. Neurosci., 8: 658–665. Hess, G., Aizenman, C.D. and Donoghue, J.P. (1996) Conditions for the induction of long-term potentiation in layer II/III horizontal connections of the rat motor cortex. J. Neurophysiol., 75: 1765–1778. Iyer, M.B., Schleper, N. and Wassermann, E.M. (2003) Priming stimulation enhances the depressant effect of low-frequency repetitive transcranial magnetic stimulation. J. Neurosci., 23: 10867–10872. Karni, A. and Sagi, D. (1993) The time course of learning a visual skill. Nature, 365: 250–252. Kirkwood, A., Rioult, M.C. and Bear, M.F. (1996) Experiencedependent modification of synaptic plasticity in visual cortex. Nature, 381: 526–528. Muellbacher, W., Ziemann, U., Boroojerdi, B., Cohen, L.G. and Hallett, M. (2001) Role of the human motor cortex in rapid motor learning. Exp. Brain Res., 136: 431–438. Muellbacher, W., Ziemann, U., Wissel, J., Dang, N., Kofler, M., Facchini, S., Boroojerdi, B., Poewe, W. and Hallett, M. (2002) Early consolidation in human primary motor cortex. Nature, 415: 640–644. Nitsche, M.A. and Paulus, W. (2000) Excitability changes induced in the human motor cortex by weak transcranial direct current stimulation. J. Physiol., 527: 633–639. Nitsche, M.A., Fricke, K., Henschke, U., Schlitterlau, A., Liebetanz, D., Lang, N., Henning, S., Tergau, F. and Paulus, W. (2003) Pharmacological modulation of cortical excitability shifts induced by transcranial direct current stimulation in humans. J. Physiol., 553: 293–301.

Philpot, B.D., Sekhar, A.K., Shouval, H.Z. and Bear, M.F. (2001) Visual experience and deprivation bidirectionally modify the composition and function of NMDA receptors in visual cortex. Neuron, 29: 157–169. Quinlan, E.M., Philpot, B.D., Huganir, R.L. and Bear, M.F. (1999) Rapid, experience-dependent expression of synaptic NMDA receptors in visual cortex in vivo. Nature Neurosci., 2: 352–357. Rioult-Pedotti, M.-S. and Donoghue, J. (2003) The nature and mechanisms of plasticity. In: S.J. Boniface and U. Ziemann (Eds.), Plasticity in the Human Nervous System. Investigations with Transcranial Magnetic Stimulation. Cambridge University, Cambridge, pp. 1–25. Rioult-Pedotti, M.-S., Friedman, D., Hess, G. and Donoghue, J.P. (1998) Strengthening of horizontal cortical connections following skill learning. Nature Neurosci., 1: 230–234. Rioult-Pedotti, M.-S., Friedman, D. and Donoghue, J.P. (2000) Learning-induced LTP in neocortex. Science, 290: 533–536. Sanes, J.N. and Donoghue, J.P. (2000) Plasticity and primary motor cortex. Annu. Rev. Neurosci., 23: 393–415. Shea, C.H., Lai, Q., Black, C. and Park, J.H. (2000) Spacing practice sessions across days benefits the learning of motor skills. Hum. Mov. Sci., 19: 737–760. Siebner, H.R., Lang, N., Rizzo, V., Nitsche, M.A., Paulus, W., Lemon, R.N. and Rothwell, J.C. (2004) Preconditioning of lowfrequency repetitive transcranial magnetic stimulation with transcranial direct current stimulation: evidence for homeostatic plasticity in the human motor cortex. J. Neurosci., 24: 3379–3385. Stefan, K., Kunesch, E., Cohen, L.G., Benecke, R. and Classen, J. (2000) Induction of plasticity in the human motor cortex by paired associative stimulation. Brain, 123: 572–584. Stefan, K., Kunesch, E., Benecke, R., Cohen, L.G. and Classen, J. (2002) Mechanisms of enhancement of human motor cortex excitability induced by interventional paired associative stimulation. J. Physiol., 543: 699–708. Walker, M.P., Brakefield, T., Morgan, A., Hobson, J.A. and Stickgold, R. (2002) Practice with sleep makes perfect: sleepdependent motor skill learning. Neuron, 35: 205–211. Wolters, A., Sandbrink, F., Schlottmann, A., Kunesch, E., Stefan, K., Cohen, L.G., Benecke, R. and Classen, J. (2003) A temporally asymmetric Hebbian rule governing plasticity in the human motor cortex. J. Neurophysiol., 89: 2339–2345. Ziemann, U. (2004) TMS induced plasticity in human cortex. Rev. Neurosci., 15: 252–266. Ziemann, U., Ilić, T.V., Pauli, C., Meintzschel, F. and Ruge, D. (2004) Learning modifies subsequent induction of LTP-like and LTD-like plasticity in human motor cortex. J. Neurosci., 24: 1666–1672.


27

Chapter 4

Cortical activities elicited by viewing mouth movements: a magnetoencephalographic study Kensaku Mikia,b,*, Shoko Watanabea, Ryusuke Kakigia,b,c and Aina Puced a

Department of Integrative Physiology, National Institute for Physiological Sciences, Okazaki 444-8585 (Japan) b Japan Space Forum, Ohtemachi, Chiyoda, Tokyo 100-0004 (Japan) c RISTEX, Japan Science and Technology Agency, Ohtemachi, Chiyoda, Tokyo 100-0004 (Japan) d Department of Radiology, Center for Advanced Imaging, West Virginia University, Morgantown, VA 26506 (USA)

1. Introduction Recent neuroimaging studies examining brain responses when viewing the actions of others indicate that in addition to regions such as the superior temporal sulcus (STS) and middle temporal gyrus (MTG), MT/V5 also plays a prominent role in processing these complex motion stimuli (Bonda et al., 1996; Puce et al., 1998; Kourtzi and Kanwisher, 2000; Watanabe et al., 2001). Previously, fMRI activity when observing eye movements over and above that seen in response to general motion has been reported in MT/V5 (Puce et al., 1998), leading to speculation that specialized cortical regions for the movement of facial parts might be present in MT/V5 (Watanabe et al., 2001). This activity occurs at around 200 ms post-motion onset, as indicated by both event-related potentials (ERP) (Puce et al., 2000, 2003) and magnetoencephalography (MEG) studies (Watanabe et al., 2001).

*Correspondence to: Kensaku Miki, Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585, Japan. Tel: +81 564-55-7814; Fax: +81 564-52-7913; E-mail: [email protected]

In this study, we investigated both temporal and spatial characteristics of MEG responses elicited by viewing mouth movements (opening and closing), as compared to viewing control movement types such as eye aversion movement and general motion. We used apparent motion, which is perceived by the same mechanism as real motion (Kaneoke et al., 1997; Watanabe et al., 2001). We have already reported this study (Miki et al., 2004) and summarized it in this chapter. 2. Methods 2.1. Subjects We studied 17 right-handed volunteers (4 females, 13 males) ranging in age from 24 to 43 years (mean, 32.2) with normal or corrected visual acuity. All subjects gave informed consent to participate in the experiment, which was approved by the Ethics Committee of the National Institute for Physiological Sciences. 2.2. Visual stimulation We used a series of apparent motion conditions, where the first stimulus, S1, was replaced by a second

28 stimulus, S2, with no inter-stimulus interval (Fig. 1): M-OP – the mouth suddenly opened; M-CL: the mouth suddenly closed; EYES: the eyes stared at the viewer and suddenly deviated to the right; RADIAL: the background rings moved inward; CONTROL: this condition differentiated MEG responses elicited by viewing movement vs. stimulus onset (Fig. 1: CONTROL). S1 and S2 were identical. Subjects noted no change from S1 onset to S2 offset. A sixth stimulus type (Fig. 1: Filler) was presented between stimulus conditions, so as to avoid large luminance and contrast changes during the experiment. This interval stimulus consisted of a scrambled image of the face on the radial background. Hence, there were no overall luminance and contrast differences between S1, S2, and the Filler. In all five conditions, S1 was shown for 800 ms, as was S2. The five stimulus conditions (consisting of S1 + S2 combinations) were presented randomly during the course of the experiment. The filler stimulus was presented for a random interval of 1000–1200 ms between each stimulus condition, producing a period of 2600–2800 ms between successive trials. Stimuli were presented using a personal computer (PC, IBM) and video projector (LP-9200, Sanyo, Japan)

housed outside a magnetically shielded room (Vacuumschmerze GmbH, Germany). Mean luminance in the room was 0.2 cd/m2. The distance between the subject’s eyes and the display was 148 cm. Stimuli were projected centrally, and subtended a visual angle of 11.6° × 11.6°. Subjects were asked to maintain their gaze at a point at the top of the nose and between the eyes of the stimulus face. The mean luminance of the center (fixation) point of the face was 14.0 cd/m2. 2.3. MEG recording MEG was measured using a 37-channel biomagnetometer (Magnes, Biomagnetic Technologies Inc., San Diego, CA). Subjects lay supine on a bed and the MEG probe was positioned at the back of the head overlying the occipito-temporal area of one hemisphere in all subjects with separate recording from the other hemisphere also made in the same recording session as our previous studies (Watanabe et al., 1999a, b, 2001, 2003). The probe sampled activity from the primary visual cortex and the occipito-temporal junction. The left and right hemispheres were studied independently, with a counterbalanced order across subjects.

Fig. 1. Examples of five stimulus conditions and their timing during the experiment.

29 MEG and vertical and horizontal electrooculograms (EOGs) were simultaneously recorded with a bandpass of 0.1–50 Hz and digitized at a sampling rate of 520.8 Hz. Epochs in which signal variations were larger than 3 pT in MEG and 80 μV in EOG were excluded from the average. The analysis window of 1500 ms was divided into two sections: 800 ms after the S1 onset and 700 ms after the S2 onset. A 300 ms pre-stimulus baseline was used for responses to S1 and S2. The amplitude of recognizable components was measured as a root mean square (RMS) value across the 37 channels of averaged response data in the order of fT. Peak latency was measured at the point with the largest RMS at visible peaks of each component. 2.4. Data analysis We used a multi-dipole model, brain electric source analysis (BESA) (Scherg and Bucher, 1993) (Neuroscan, McLean, VA), computation of theoretical source generators in a three-layer spherical head model. We then accepted the dipole model based on two criteria: (1) Goodness of fit (GoF) values larger than 85% were defined as indicating adequate multiple dipole models (see Watanabe et al., 1999a, b, 2003). An increase in the dipole number mathematically increases the GoF, since a greater number of dipoles will account for more variance (Watanabe et al., 1999a, b, 2003); however, the generated solutions may not always be physiologically plausible. (2) Sources estimated in gray matter after overlaying on MRI. For the four stimulus conditions (M-OP, M-CL, EYES, and RADIAL), taking the results of our previous studies (Watanabe et al., 1999a, b, 2001, 2003) and other neuroimaging studies into account (e.g. Puce et al., 1998, 2003) and based on a principal component analysis (PCA), we made a 4-source model as follows: source 1: the occipito-temporal junction, MT/V5 homologue in humans, source 2: left primary visual cortex (V1), source 3: right V1; source 4: the fusiform gyrus. Our MEG sensor locations in this study covered V1 bilaterally, as well as the lateral occipito-temporal cortex in one hemisphere. Initially, we placed each of

the 4 sources around each corresponding region. The BESA calculation allows some change in the initial location and freedom in the orientation of each source on PCA, so it is possible for each source to move to a nearby location, if it better fits the data results. We computed the best location and orientation of each source separately in each condition. We used the anatomical landmark criteria reported by Dumoulin et al. (2000) to confirm the location of source 1 in MT/V5. Their MT/V5 fMRI activation by moving random checkerboard patterns was located along the inferior temporal sulcus (ITS), which they separated into 3 parts: (1) the posterior ITS; (2) the ascending limb of the ITS (ALITS); and (3) the posterior continuation of the ITS (PCITS). We used analysis of variance (ANOVA) with post hoc tests of Fisher’s protected least significant difference (PLSD), or paired t-tests (p < 0.05) to assess significant differences between conditions. 2.5. MRI overlay Locations of sources calculated by BESA were converted into pixels using the MRI transformation matrix and overlaid onto corresponding MR images. 3. Results The number of trials averaged per subject was 95.4 ± 2.9, 94.4 ± 3.0, 94.7 ± 3.3, 94.6 ± 3.6, and 94.7 ± 4.0 for M-OP, M-CL, EYE, RADIAL, and CONTROL, respectively. We present detailed results for S2, or motion onset, as this was the focus of our study. 3.1. The right hemisphere 3.1.1. Waveform characteristics All subjects reported experiencing motion perception from four apparent motion conditions, but not from the CONTROL. The most prominent component, 1M, was observed in all conditions with apparent motion (M-OP, M-CL, EYES, and RADIAL) in 12 out of 17 subjects (Fig. 2). 1M peak latency, which was

30 measured at the point with the largest RMS at the visible peak of each component, was similar under facial motion conditions (Table 1), peaking around 160 ms, and was around 20 ms shorter for RADIAL. The RADIAL latency was significantly shorter than that under facial motion conditions (p < 0.05). There were no significant differences in 1M latency between the two mouth movement types (M-OP and M-CL), nor did they differ relatively to EYES (see Table 1). The signal strength of 1M, measured as maximum RMS, and the maximum RMS for the mouth movement conditions (M-OP and M-CL) did not differ significantly; however, M-OP was significantly smaller than RADIAL (p < 0.05), but was not different to EYES. M-CL was significantly smaller than EYES (p < 0.05) and RADIAL (p < 0.01). There were no significant differences between EYES and RADIAL.

Fig. 2. Right hemisphere S2 MEG activity shown in a 37-channel superimposed display for all conditions. In subject 1, 1M peak latency was 154.8, 156.7, 162.5, and 148.1 ms for M-OP, M-CL, EYES, and RADIAL, respectively. Associated maximum RMS values were 62.6, 66.7, 122.0, and 119.4 fT.

3.1.2. Source analysis using BESA BESA results for M-OP and M-CL fulfilled our strict criteria in 9 out of 12 subjects whose 1M was observed in all conditions with apparent motion (M-OP, M-CL, EYES, and RADIAL) (Fig. 3). We analyzed dipole moment in the nAm of source 1. The maximum dipole moment for the mouth movement conditions (M-OP, M-CL, and EYES) did not

TABLE 1 1M PEAK LATENCY, MEASURED AT THE POINT WITH THE MAXIMUM RMS AT VISIBLE PEAKS OF EACH COMPONENT, AND MAXIMUM RMS VALUES FOR S2 TO FOUR STIMULUS TYPES IN THE RIGHT AND LEFT HEMISPHERES (MEANS AND STANDARD DEVIATIONS)

M-OP M-CL EYES RADIAL

Latency (ms) RMS (fT) Latency (ms) RMS (fT) Latency (ms) RMS (fT) Latency (ms) RMS (fT)

*p < 0.05, **p < 0.01: comparison with results of RADIAL. # p < 0.05, ##p < 0.01: comparison with results of EYES.

Right (n = 12)

Left (n = 11)

159.8 ± 17.3* 62.5 ± 23.5** 161.9 ± 15.0* 59.1 ± 21.8**,# 161.2 ± 18.9** 82.5 ± 32.7 140.1 ± 18.0 98.7 ± 29.7

162.4 ± 11.6** 56.0 ± 28.1**,# 160.9 ± 9.8** 50.1 ± 17.5**,## 164.6 ± 14.2** 87.3 ± 42.7 138.4 ± 9.0 99.4 ± 33.0

31

Fig. 3. Right hemisphere 4-source BESA model for S2 for the time interval highlighted in green. The same subjects as in Fig. 2. Source 1 (red) was located in human MT/V5, sources 2 (pink) and 3 (green) in the left and right primary visual fields (V1), and source 4 (blue) in the right fusiform gyrus. Overall GoF values are also displayed. In subject 1, the activity of source 1 was very large relative to the other sources, and this pattern was observed in the majority of subjects. TABLE 2 DIPOLE MOMENT IN THE nAm OF SOURCE 1 TO FOUR STIMULUS TYPES IN THE RIGHT AND LEFT HEMISPHERES (MEANS AND STANDARD DEVIATIONS)


Right (n = 9)

Left (n = 9)

7.9 ± 1.9* 7.8 ± 3.2** 10.0 ± 6.8 13.8 ± 4.9

7.4 ± 2.8** 6.7 ± 3.0** 9.3 ± 4.3** 13.6 ± 1.8

*p < 0.05, **p < 0.01: comparison with RADIAL results.

differ significantly (Table 2). However, M-OP and MCL were significantly smaller than RADIAL (p < 0.05). We used the anatomical landmark criteria reported by Dumoulin et al. (2000) to confirm the location of source 1 in MT/V5. In this study, the source 1 for each movement condition in all subjects was located in one of three defined regions: in 2 subjects in or adjacent to (1) the posterior ITS, in 4 subjects in or adjacent to (2) the ascending limb of the ITS (ALITS), and in the other 3 subjects in or adjacent to (3) the posterior continuation of the ITS (PCITS). The locations of source 1 for M-OP and M-CL, as indicated by their x, y, and z coordinates, were similar (Table 3, Fig. 4). Source 1 for both M-OP and M-CL appeared to be located more anterior and superior

32 TABLE 3 SOURCE 1 LOCATIONS FOR S2. MEANS AND STANDARD DEVIATIONS OF x, y, AND z COORDINATES (mm) FOR M-OP, M-CL, EYES, AND RADIAL IN THE RIGHT AND LEFT HEMISPHERES. x IS POSITIVE TO RIGHT, y TO ANTERIOR, AND z TO SUPERIOR Right (n = 9)


Left (n = 9)

x (mm)

y (mm)

z (mm)

x (mm)

y (mm)

z (mm)

40.8 ± 5.8 41.8 ± 6.1 40.0 ± 7.9 43.0 ± 7.8

−27.6 ± 10.7 −26.2 ± 10.1 −30.1 ± 9.4 −27.2 ± 8.7

60.9 ± 8.1 60.8 ± 7.5 59.9 ± 7.2 61.2 ± 6.4

−35.1 ± 5.6 −35.2 ± 7.1 −35.8 ± 7.6 −31.6 ± 6.5

−31.2 ± 9.1 −30.6 ± 8.8 −34.9 ± 3.8 −31.2 ± 3.9

63.2 ± 6.0 60.6 ± 6.1 59.1 ± 5.3 63.3 ± 6.4

Fig. 4. Right hemisphere locations for S2 source 1 ECDs for apparent motion conditions overlaid on axial, coronal, and sagittal MRI slices, and the volume-rendered brain of subject 1.

33 relative to EYES, and more medial relative to RADIAL. However, these differences were not significant.

3.2. The left hemisphere 3.2.1. Waveform characteristics In the left hemisphere, as in the right, the most prominent MEG component was 1M, which was observed for all apparent motion conditions in 11 out of 17 subjects. The latency of the 1M was significantly shorter for RADIAL than all other conditions (p < 0.01). There were no significant latency differences between the mouth movement conditions (M-OP and M-CL), or between these conditions and EYES. The maximum RMS values for M-OP and M-CL were significantly smaller than those for EYES (p < 0.01) or RADIAL (p < 0.01), but showed no significant differences among themselves. 1M RMS values for EYES and RADIAL did not differ significantly. 3.2.2. Source analysis using BESA Using our GoF criteria, BESA results for M-OP and M-CL fulfilled the strict criteria in 9 out of 11 subjects who showed clear elicited MEG activity. In all 9 subjects who met the GoF criteria for apparent motion stimulus conditions, source 1 located in the lateral temporal region around MT/V5, was very large in amplitude, but the other 3 sources were very small in amplitude or showed no significant activity. We analyzed dipole moment in the nAm of source 1. The maximum dipole moment for the mouth and eye movement conditions (M-OP, M-CL, and EYES) did not differ significantly; however, M-OP, M-CL, and EYES were significantly smaller than RADIAL (p < 0.05). In both the left and right hemispheres, the source 1 for each movement condition in all subjects was located in or adjacent to the three anatomical regions defined earlier: region (1) in 2 subjects, region (2) in 4 subjects, and region (3) in 3 subjects. Therefore, our findings were in agreement with Dumoulin et al. (2000), and other imaging studies of MT/V5, for both the left and right hemispheres. The locations, measured as x, y, and z coordinates, of source 1 for M-OP and M-CL were similar (Table 3)

and relative to EYES, appeared to be located more superiorly and, compared with RADIAL, posteriorly (Table 3). Statistical testing, however, indicated that these apparent differences in the location of source 1 were not significant. 3.3. Comparison of data from right and left hemispheres 1M peak latencies in both hemispheres were compared using the paired t-test in 9 subjects who showed clear 1M components in each hemisphere. There were no significant inter-hemispheric differences of latency (Table 1). We did not compare RMS values between the hemispheres, as sensor placement over each hemisphere may not have been perfectly symmetrical and hence the distance between MEG sensors and the brain may not have been the same across hemispheres. The dipole moments of source 1 in both hemispheres were compared using the paired t-test in 5 subjects who showed a reliable source 1 for all four stimulus types in each hemisphere. There was no significant interhemispheric difference. 4. Discussion The location of sources for various facial movements did not differ within themselves or relative to the radial background condition. Interestingly, despite this lack of difference in the source locations, there were differences in the behavior of 1M across conditions. While 1M showed a peak latency of around 160 ms, the facial motion conditions in general produced longer 1Ms relative to our general motion control – a finding that we have seen previously using eye movements (Watanabe et al., 2001). Additionally, responses to mouth movements were smaller than responses observed to the eye aversion or radial background motion controls. These differences in the behavior of 1M suggest that MT/V5 and its surrounds possess multiple response characteristics. The latency and RMS data for the eye and radial background movements were consistent with our previous study (Watanabe et al., 2001) and will not be discussed further. We found no clear amplitude difference between the two mouth movement conditions, unlike the ERP

34 study of Puce et al. (2000, 2003), who reported an N170 component, corresponding to our 1M, which was significantly earlier in latency and larger in amplitude to mouth opening relative to mouth closing movements. There are a number of possible reasons for the observed differences between the ERP and this MEG study. Theoretically, MEG will detect tangential neural sources located just beneath the MEG sensors, and EEG will detect not only radial dipoles located near the electrodes but also more distant sources by volume current conduction, effectively detecting sources directed both tangentially and radially. The direction of the net dipole generated in STS was mainly radial, enabling it to be better detected with ERP methods, whereas the direction of the net dipole generated in MT/V5 was mainly tangential. As there are potentially multiple generators in the temporoparietal cortex that temporally overlap and may partially overlap in space, it could prove difficult for MEG, or indeed ERP methods, to clearly identify them. Additionally, neural activity, as detected by MEG, might be much smaller in STS than in MT/V5, so its contribution is masked by the net dipole generated by MT/V5 activity. This is more likely, given that in ERP studies the activity in response to general motion controls is smaller than that observed for facial motion (Puce et al., 2000, 2003). There was a slight difference between amplitude results using two different factors, i.e. there was no significant difference of dipole moment among M-OP, M-CL, and EYES, although the RMS was significantly larger for EYES than M-OP and M-CL. The RMS included the activities of all four sources, but the dipole moment of source 1 showed that activity was restricted to MT/V5. Therefore, this finding indicated that MT/V5 activities in response to eyes and mouth movements were not significantly different, but activities in the other 3 regions were significantly larger for eyes than mouth movements. This is an interesting finding, and results using dipole moment were compatible with an ERP study (Puce et al., 2000) reporting that N170 amplitudes for eye and mouth movements were not significantly different.

The source locations for mouth opening, mouth closing, and eye aversion movements or radial background movements did not differ despite significant differences in response latency. MEG source locations indicate activity centroids, and hence, it may be difficult for MEG to detect different activated regions when regions partially overlap. Similarly, partial voluming effects in PET and fMRI could also make it difficult to detect such small differences in partially overlapping activated regions. References Bonda, E., Petrides, M., Ostry, D. and Evans, A. (1996) Specific involvement of human parietal systems and the amygdala in the perception of biological motion. J. Neurosci., 16: 3737–3744. Dumoulin, S.O., Bittar, R.G., Kabani, N.J., Baker Jr., C.L., Le Goualher, G., Bruce, P.G. and Evans, A.C. (2000) A new anatomical landmark for reliable identification of human area V5/MT: a quantitative analysis of sulcal patterning. Cereb. Cortex, 10: 454–463. Kaneoke, Y., Bundou, M., Koyama, S., Suzuki, H. and Kakigi, R. (1997) Human cortical area responding to stimuli in apparent motion. Neuroreport, 8: 677–682. Kourtzi, Z. and Kanwisher, N. (2000) Cortical regions involved in perceiving object shape. J. Neurosci., 20: 3310–3318. Miki, K., Watanabe, S., Kakigi, R. and Puce, A. (2004) Magnetoencephalographic study of occipitotemporal activity elicited by viewing mouth movements. Clin. Neurophysiol., 115: 1559–1574. Puce, A., Allison, T., Bentin, S., Gore, J.C. and McCarthy, G. (1998) Temporal cortex activation in humans viewing eye and mouth movements. J. Neurosci., 18: 2188–2199. Puce, A., Smith, A. and Allison, T. (2000) ERPs evoked by viewing facial movements. Cogn. Neuropsychol., 17: 221–239. Puce, A., Syngeniotis, A., Thompson, J.C., Abbott, D.F., Wheaton, K.J., Castiello, U. (2003) The human temporal lobe integrates facial form and motion: evidence from fMRI and ERP studies. Neuroimage, 19: 861–869. Scherg, M. and Buchner, H. (1993) Somatosensory evoked potentials and magnetic fields: separation of multiple source activities. Physiol. Meas., 14 (Suppl 4A): A35–A39. Watanabe, S., Kakigi, R., Koyama, S. and Kirino, E. (1999a) It takes longer to recognize the eyes than the whole face in humans. Neuroreport, 10: 2193–2198. Watanabe, S., Kakigi, R., Koyama, S. and Kirino, E. (1999b) Human face perception traced by magneto- and electroencephalography. Cogn. Brain Res., 8: 125–142. Watanabe, S., Kakigi, R. and Puce, A. (2001) Occipitotemporal activity elicited by viewing eye movements: a magnetoencephalographic study. Neuroimage, 13: 351–363. Watanabe, S., Kakigi, R. and Puce, A. (2003) The spatiotemporal dynamics of the face inversion effect: a magneto- and electroencephalographic study. Neuroscience, 116: 879–895.


35

Chapter 5

Optimal methods of stimulus presentation and frequency analysis in P300-based brain–computer interfaces for patients with severe motor impairment Ryuji Neshigea,*, Nobuki Murayamab, Kazuya Tanoueb, Hiroaki Kurokawab and Tomohiko Igasakib a Neshige Neurological Clinic, 38-17 Tyuou-machi, Kurume, Fukuoka 830-0023 (Japan) Department of Electrical and Computer Engineering, Kumamoto University, 2-39-1 Kurokami, Kumamoto City, Kumamoto 860-8555 (Japan)

b

1. Introduction The P300-based brain–computer interface (BCI) was originally reported by Farwell and Donchin (1988) as a mental prosthesis. The system utilizes the fact that rare events in the oddball paradigm elicit the P300 component of the event-related potential (ERP). Subsequently, using the same method as Farwell and Donchin (1988), Donchin et al. (2000) reported, assessing the speed of a P300-based BCI, that a high correct response rate (95%) was obtained in about 14–19 s when the P300 amplitude was analyzed. With their method, 16 letters are arranged 4 × 4, and one letter serves as the target. When the luminance of each row and column was elevated at random to serve as the stimulus, P300 was induced in response to stimulation by the row and column that contained the target. The letter located at the point of intersection between this row and column was defined

*Correspondence to: Ryuji Neshige, Neshige Neurological Clinic, 38-17 Tyuou-machi, Kurume, Fukuoka 830-0023, Japan. Tel: +81-0942-36-6855; Fax: +81-0942-36-6856; E-mail: [email protected]

as the target. Their method focused too much on the speed of communication. The stimulus presentation time (100 ms) and the interval of stimulus presentation (25 ms) are quite short, and it seems to be difficult to apply this method clinically. Considering that the peak latency of visual P300 was around 400 ms (Neshige et al., 1991), it is questionable as to which components were analyzed by the method reported by Donchin et al. (2000) Here, we have presented new communication methods that may be applied to patients with motor disabilities. 2. Basic experiments 2.1. Methods 2.1.1. Subjects Healthy, right-handed university students (7 males and 1 female, 21–24 years old) were enrolled in the experiment. 2.1.2. Stimulation Each subject sat on a chair within a shield room. The subject was instructed to gaze at an image shown on

36 a 17-inch display placed one meter in front of them. Four pictorial symbols (rice, toilet, cup, and X) representing the choices of “want to eat,” “want to use toilet,” “want to drink water,” and “don’t want to do anything,” respectively, were presented (Fig. 1). Each symbol was presented against a black background. The symbols were depicted in green. The illumination level before the subject’s eyes was set to 68–70 lx. A stimulus symbol was presented for 300 ms with inter-stimulus interval of 1500 ms. The three methods shown in Fig. 1 were used for stimulus presentation. The background color change method (BCCM) displayed all symbols simultaneously during both the non-stimulation and stimulation phases and changed the color of the background of one of the 4 symbols to purple during the stimulation phase. The peripheral area presentation method (PAPM) presented no symbol during the

stimulation

non-stimulation phase and placed the symbol in the same location as used for BCCM during the stimulation phase. The central area presentation method (CAPM) presented no symbol during the non-stimulation phase and placed the symbol at the center of the display during the stimulation phase. One session of the experiment was performed in a randomized order. Each subject performed 4 sessions of the experiment. The methods of presentation were selected at random. In all sessions of the experiment, the cup served as the target stimulus. The subjects were instructed to pay close attention to the picture of the cup presented. In BCCM alone, we prepared 3 different sets of presentations, each of which consisted of 3 different pictorial symbols and X. Subjects selected one pictorial symbol as the target. When a personal computer (PC) selected X, another set of presentations was displayed

non-stimulation

stimulation

A

B

C

1500msec

300msec

Fig. 1. Diagrams of three visual presentation methods, (A) background color change method (BCCM), (B) peripheral area presentation method (PAPM), (C) central area presentation method (CAPM), Background color change or each picture was respectively displayed for 300 ms during thepresentation period with a stimulus interval of 1500 ms.

37 automatically and the experiment continued until the PC selected one symbol. 2.1.3. Recording EEGs were recorded with 11 electrodes (F3, Fz, F4, C3, Cz, C4, P3, Pz, P4, O1, and O2) arranged according to the International 10–20 system. The electrode linking both earlobes served as the reference electrode. The bandpass filter was set to 0.5–60 Hz. The interelectrode impedance was kept below 30 kg/cm2. To detect artifacts caused by blinking and eye movement, electrooculograms (EOG) were simultaneously recorded with bipolar electrodes placed at the middle of the forehead and in the inferolateral side of the left eye. Trials on which the EEG or EOG potential exceeded 100 μV were excluded from analysis. Sampling was done at intervals of 1 ms. For each stimulus image presentation, EEGs were recorded for the period from 100 ms before presentation to 1024 ms after presentation. 2.1.4. Analysis For amplitude analysis, the mean potential during the 100 ms period before stimulation served as the baseline. The positive peak appearing 230–440 ms after stimulation was defined as P300. The amplitude from the baseline to the P300 peak was defined as the P300 amplitude. The P300 amplitude response was compared for each of the different symbols. If the P300 amplitude following presentation of the target stimulus was maximal, the response was rated as correct. The correct response rates for 1 through 10 were calculated and compared among the different methods of stimulus presentation. During all experimental sessions, each symbol was presented 10 or more times. A total of 10 summations were calculated. For frequency analysis, fast Fourier transform (FFT) was performed at 1024 points on the averaged EEGs ranging from the starting point of stimulation to the post-stimulation period (0–1023 ms). The power spectrum was summed for a randomly selected range of frequencies and compared for each one of the different symbols. If the power spectrum was maximal following presentation of the target stimulus, the response was rated as correct. The frequency range for the power

spectrum was changed between 1 and 10 Hz in 1 Hz steps. The correct response rate was compared among different frequency ranges. 2.2. Results 2.2.1. P300 amplitude analysis The ERPs after averaging 1 did not allow the P300 response to the cup (target response) to be distinguished from responses to the other stimuli (non-target responses) with any method of stimulation. However, as the number of averages was increased to 5, it became possible to distinguish a target response from a nontarget response. In topographies of the target responses recorded after averaging 10 times in all subjects, we compared the amplitude at each site for the P300 and performed normalization with a maximum of 1 and a minimum of 0. Since intense responses were seen primarily at Pz, only EEGs recorded from Pz were analyzed. With each presentation method, the P300 amplitude in response to target stimulus was significantly greater than that in response to non-target stimuli (p < 0.01 to p < 0.001, Student’s t-test). With BCCM, the correct response rate was low when the average number was 4 or less, but the rate was higher than about 70% when the average number was 5 or more. The correct response rate was the highest with BCCM, as compared with PAPM and CAPM, where the correct response rate was below 60%. In subsequent analyses, only data obtained with BCCM were analyzed. 2.2.2. Frequency analysis The mean ± SD of the correct response rate after summation 6 was as high as 96.88 ± 8.27% when the frequency range was 2–5 Hz. This was significantly higher than the P300 amplitude analysis (75.00 ± 21.65%, p < 0.05, Student’s t-test). Figure 2 shows the results of this analysis for 2 subjects. The left side of the figure shows the data from a 22-year-old female. In the ERPs averaged 10 times, the solid line (response to the cup) had a higher amplitude than the thin lines (responses to other stimuli) (Fig. 2A, left). Thus, the ERP to the target was

38 toilet no-action

eat drink

A

−20

Amplitude (μV)

−30

30

20 0

0

1024 Time[msec]

Power (μV2/Hz)

B

1024 Time[msec]

10 1

10−6 0

25

0

50

25

50

Frequency[Hz]

Frequency[Hz] 2−5Hz Power (μV2/Hz)

C

15

5

0

0 eat

toilet

drink

no action

eat

toilet

drink

no action

Fig. 2. The discrimination between target and non-target responses using a frequency analysis of 2 subjects for the 2–5 Hz range. Left: 22-year-old female, right: 24-year-old male. (A) ERPs obtained by BCCM at 10 averaging times. From the P300 response of a 22-year-old female, we could easily distinguish target (thick line) from non-target (left) but not for a 24-year-old male (right). (B) Each response mentioned above was converted into the frequency domain by a FFT method and a summation of the 2–5 Hz power spectrum was calculated. (C) In both subjects, a summation of the 2–5 Hz power spectrum of target (drink) showed a maximum value as compared to the non-targets.

distinguishable from those to non-target responses, and the response was rated as correct. This averaged wave was subjected to frequency analysis using FFT, to calculate the power spectrum (Fig. 2B, left). From the calculated power spectra, the sum total of powers for the 2–5 Hz range was calculated (Fig. 2C, left). This analysis revealed that the power spectrum in response to the target stimulus was very large compared to the

responses to other stimuli. The right side of Fig. 2 shows the results for a 24-year-old male. In this case, the P300 amplitude in response to the target stimulus cannot be identified, since the ERP to the target stimulus cannot be distinguished from those to the other stimuli (Fig. 2A, right). Alternatively, the 2–5 Hz power spectrum was larger in response to the target stimulus than to non-target stimuli (Fig. 2C, right).

39 3. Clinical experiments 3.1. Methods 3.1.1. Subjects Clinical experiments involved 17 healthy volunteers (mean age: 25 ± 9 years) and 3 patients with amyotrophic lateral sclerosis (ALS) (no. 1: age 48, no. 2: age 49, no. 3: age 58). Both the patients and their family members were willing to participate in the experiments. All subjects had normal intelligence and were able to easily understand the experimental methods. In no. 1 and no. 2, verbal communication was not possible because of severe dysarthria and the experiments were performed with their heads immobilized on the headrest because their muscle weakness of the neck was very noticeable. Although no. 1, no. 2, and no. 3 did not use a ventilator, their vital capacity was below 1000 ml. The following three clinical experiments were performed, based on the results of the basic experiment, of which the best method was found to be BCCM, using a 2–5 Hz power spectrum for analysis. The healthy volunteers were analyzed only by their P300 maximum amplitude and the patients were analyzed both by P300 maximum amplitude and ERP frequency. 3.1.2. “Yes or No” paradigm (Fig. 3A) Four short sentences (“Yes very much,” “Yes,” “No,” and “Don’t know”) were presented on the display. The subject was instructed to select one of these four sentences as the target. The background color of each sentence was changed to serve as the stimulus. The total frequency of stimulus presentation was 28. Each sentence was presented 7 times at random and care was taken to avoid the same sentence being presented in succession. These conditions were also used for the following paradigms. 3.1.3. Fifty-sound paradigm (Fig. 3B) A systematic table of 50 Japanese sounds was presented on the display. The subject was instructed to select one target letter. Determination of row: While the systematic table of 50 Japanese sounds was on the display, the

background color for each row was changed at random to serve as the stimulus. The total frequency of stimulus presentation was 50. Each row was presented 5 times. Determination of letters: A row was selected by the PC to contain the target. After X was added to the end of this row, the row was presented at the center of the display. The background color for one letter in this row was changed at random to serve as the stimulus. The total frequency of stimulus presentation was 5. In view of the results in healthy volunteers where the letters were arranged vertically, the letters were arranged horizontally in the experiment involving patients. The X served as the target for cases where no target was presented on the display or if the PC made an erroneous answer. 3.1.4. Pictorial symbol paradigm (Fig. 3C) Sixteen pictorial symbols, including X, were arranged on the display. Each row (in the vertical direction) was composed of 4 symbols and each column (in the horizontal direction) was made up of 5 symbols. Stimulation was performed as in a fifty-sound paradigm experiment. The subject was instructed to select one symbol as the target. The background color for each row, composed of 4 symbols, was changed to serve as the stimulus. For the row determined by the PC to contain the target, X was added to the 4 symbols, which were then arranged horizontally while changing the background color for each symbol to serve as the stimulus. 3.1.5. Task The subject was instructed to count 1, 2, 3, …, mentally at a rate of about one count per second, which was not to be synchronous with the rate of stimulus presentation. He/she was instructed to discontinue counting when the target was presented, and to resume counting from 1 after a short break. Recording and analysis: Fz, Cz, Pz, and Oz were used as electrodes with a linked ears reference. Of the averaged ERPs, the wave satisfying the following requirements was determined online by the PC to contain the target: (1) showing a positive peak 300–600 ms after the stimulus, or (2) the wave with the maximal power spectrum for the 2–5 Hz range after FFT.

40 A

yes or no

Yes yes very much

no don’t know

B

fifty-sound

Raw (healthy volunteer)

C

pictorial symbol

picture

Letter (patient)

Fig. 3. Presentations of stimuli on the displays for the mental prosthesis in 3 experiments.

Healthy volunteers underwent all three experiments mentioned above. The patients received experiments 1 and 2. 3.2. Results 3.2.1. “Yes and No” paradigm The correct response rate was 100% both for the healthy volunteers by P300 amplitude analysis and 3 patients by frequency analysis. In the analysis of healthy volunteers alone, P300s in response to the

target (latency 461 ± 36 ms, amplitude 11.8 ± 3.2 μV) were detected from the Pz electrode. These potentials had a higher amplitude than non-target responses in all volunteers. Thus, the potentials recorded at Pz had a higher amplitude than those recorded at the other electrodes. 3.2.2. Fifty-sound paradigm In healthy volunteers, the correct response rate in the “determination of row” during the first session was 86% and the rate in the “determination of letter” was 77%.

41 In the second and third sessions, the correct response rate in the determination of row was 100% and for the determination of letter was 87%. In the determination of row, the maximum potentials in response to the target were recorded at Pz, with a mean latency of 499 ± 41 ms and a mean amplitude of 14.4 ± 4.1 μV. In the determination of letter, maximum potentials in response to the target were recorded at Pz, with a mean latency of 483 ± 43 ms and a mean amplitude of 13.0 ± 3.2 μV. In most records from the patient group, it was difficult to identify the peak latency. However, when a power spectrum for the 2–5 Hz range was analyzed, the correct response rate was 100% for no. 1 and no. 3. In no. 2, a wrong response was recorded only for the “determination of letter” during the second session. 3.2.3. Pictorial symbols The correct response rate was 100% for both the “determination of row” and the “determination of a single pictorial symbol.” In the “determination of row,” maximum potentials in response to the target were recorded at Pz, with a mean latency of 439 ± 49 ms and mean amplitude of 8.5 ± 2.3 μV. In the “determination of a single pictorial symbol,” maximum potentials were recorded at Pz, with a mean latency of 436 ± 46 ms and mean amplitude of 7.9 ± 2.4 μV. 4. Discussion In the basic experiments, BCCM (keeping all symbols presented during both the non-stimulation and stimulation phases and changing the background color during the stimulation phase) was found to provide the most effective means of stimulus presentation. It seems that BCCM makes it easier for the subject to pay attention to the target, without being confused with non-target stimuli, when compared to the other methods of presentation, since the BCCM keeps each symbol presented at a certain location. As a method of analysis for P300, to the authors knowledge, there have been few reports that have investigated in detail the frequency range that is most effective for ERP analysis. In the basic experiment, a

2–5 Hz power spectrum was found to be the most effective. The α waves sometimes disturb the P300 component. If the average number is low, α waves cannot be completely eliminated. If α waves with a high amplitude remain after about 300 ms following non-target stimuli, the PC may determine it as the P300 amplitude, leading to a wrong response. This also seems to be a factor responsible for the low accurate response rate. When using the 2–5 Hz power spectrum, the effects of α waves can be eliminated completely. Kolev et al. (2003) reported that δ and θ wave components of the P300 in response to target stimulus are significantly larger than those in response to non-target stimuli. The frequency range of these components is included in the 2–5 Hz range. This is probably the reason why the correct response rate was higher for the 2–5 Hz range than for the range of δ or θ waves. Unlike the study by Kolev et al. (2003), which involved analysis of only a fixed frequency range (δ waves, θ waves, etc.), we analyzed each frequency range in detail. Therefore, we may say that the 2–5 Hz range is more suitable than δ or θ waves in order to identify ERPs induced by the target. Our device for communication involves several issues that require particular mention. The first issue pertains to the method for guiding the subject to concentrate his/her consciousness on the target. To facilitate conscious concentration, we asked the subject to count from 1 and to stop counting after the target has appeared, and then to resume counting from 1 following a break. It is a paradigm overlapping the “emitted P300” (Ruchkin and Sutton, 1978), by which P300 can be induced even when no stimulus (counting) is presented. The correct response rate, as analyzed by the PC, was improved using our paradigm. Second, we should confirm that P300 can be induced clearly by the standard oddball paradigm before this device is applied clinically because P300 sometimes has a low voltage, where ERPs are easily distorted by activities not related to stimulus presentation (such as alternating current and α activity) and also because P300 is sometimes induced by non-target stimuli and sometimes not induced by target stimulus (Roschke et al., 1996). Third, a problem with this method is the long time required for evaluation.

42 Individuals often have many thoughts they wish to communicate. It is necessary to increase the number of selections used in one session, but increasing the number will take more time and cause increased stress on the subject. We therefore devised methods to allow a smooth clinical application. For the fifty-sound paradigm, 2 min were required for one letter. To resolve this problem, it appears necessary to modify the method by transmitting 2 or 3 letters and utilize the template for communication installed in the PC to find an optimum letter. In patient no. 1, pronunciation of each word was recorded while the patient was still able to speak. If the patient loses the ability to speak in the future, the target estimated by the PC will be reproduced in the patient’s recorded voice. Even this way, longer times are necessary. After we asked the patients with ALS what they wanted to communicate, we realized there were not so many things they were eager to convey. Therefore, to shorten the time needed for communication, we also devised “yes or no,” “clinical BCCM,” and “pictorial symbols” paradigms where there were fewer selections. Although using the device is still

time-consuming, it appears quite useful for patients with ALS who have lost all ability to communicate other than directly via the brain. References Donchin, E., Spencer, K.M. and Wijesinghe, R. (2000) The mental prosthesis: assessing the speed of a P300-based brain-computer interface. IEEE Trans. Rehab. Eng., 2: 174–179. Farwell, L.A. and Donchin, E. (1988) Talking off the top of your head: toward a mental prosthesis utilizing event-related brain potentials. Electroencephalogr. Clin. Neurophysiol., 70: 510–523. Kolev, V., Demiralp, T., Yordanova, J., Ademoglu, A. and IsogluAlkac, U. (2003) Time-frequency analysis reveals multiple functional components during oddball P300. Neuroreport, 8: 2061–2065. Neshige, R., Kuroda, Y., Kakigi, R., Fujiyama, F., Matoba, R., Yarita, M., Lüders, H. and Shibasaki, H. (1991) Event-related brain potentials as indicators of visual recognition and detection of criminals by their use. Forensic Sci. Int., 51: 95–103. Roschke, J., Mann, K., Wagner, P., Grozinger, M., Fell, J. and Frank, C. (1996) An approach to single trial analysis of eventrelated potentials based on signal detection theory. Int. J. Psychophysiol., 22: 155–162. Ruchkin, D.S. and Sutton, S. (1978) Emitted P300 potentials and temporal uncertainty. Electroencephalogr. Clin. Neurophysiol., 45: 268–277.


43

Chapter 6

An integrated approach to face and motion perception in humans Shozo Tobimatsua,*, Yoshinobu Gotoa, Takao Yamasakia, Reimi Tsurusawaa,b and Takayuki Taniwakia,c a

Department of Clinical Neurophysiology, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582 (Japan) b Department of Pediatrics, Faculty of Medicine, Fukuoka University, Fukuoka 814-0184 (Japan) c Department of Neurology, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582 (Japan)

1. Introduction There are two major parallel pathways in humans; the parvocellular (P) and magnocellular (M) pathways. The former is responsible for carrying information on the form and color of an object because of its ability to detect stimuli with high spatial frequencies and color, while the latter plays an important role in detecting motion due to its ability to respond to high temporal stimuli (Livingstone and Hubel, 1988; Tobimatsu et al., 2000). We have been studying the functions of the Pand M-pathways with evoked potentials by manipulating the characteristics of the visual stimulus (Tobimatsu et al., 1995, 1999, 2000; Tobimatsu and Kato, 1998; Tobimatsu, 2000). Information on the characteristics of a face is first processed in the fusiform gyrus (V4), and the information is carried by the P-pathway (Vuilleumier et al., 2003). Information on the motion

*Correspondence to: Shozo Tobimatsu, M.D., Ph.D., Department of Clinical Neurophysiology, Graduate School of Medical Sciences, Neurological Institute, Faculty of Medicine, Kyushu University, Fukuoka 812-8582, Japan. Tel: +81-92-642-5541; Fax: +81-92-642-5545; E-mail: [email protected]. ac.jp

of an object is processed in MT/V5 and the information is carried by the M-pathway (Rizzolatti and Matelli, 2003). There are several advantages and disadvantages of the electrophysiological and neuroimaging methods when we investigate the human brain function. However, both are objective and quantifiable methods that can be used effectively in association with psychophysics to study both normal and abnormal visual function. We herein report the neural mechanisms of face and motion perception using psychophysical threshold measurements, event-related potentials (ERPs), and functional magnetic resonance imaging (fMRI).

2. Methods Ten to twenty subjects participated in each study. Informed consent was obtained after the experimental procedures had been fully explained. The Ethics Committee of the Faculty of Medicine, Kyushu University Graduate School, approved our experimental protocols. ERPs elicited by facial and motion stimuli were recorded at multiple scalp sites in normal subjects.

44 The recording parameters have been reported in detail (Arakawa et al., 1999). 2.1. Face stimuli A photograph of a face was filtered to alter the spatial frequency components and used to investigate how the low spatial frequency (LSF) and high spatial frequency (HSF) components of the face contribute to the identification and recognition of a face. First, the HSF of

a photograph of a face was filtered out by increasing the mosaic levels (up to 64 levels) using Adobe Photoshop 5.0 software. By doing so, the psychophysical threshold mosaic levels for facial perception and identification were determined (Fig. 1, upper panel). The photographs at subthreshold, threshold, and suprathreshold mosaic levels were selected as stimuli for eliciting ERPs for each subject. The familiarity of a face was determined by asking the subjects to identify the person in the unfiltered photographs.

Subthreshold

Threshold

Suprathreshold

Neutral

Angry

Sad

Horizontal

Optic Flow (radial-in)

Optic Flow (radial-out)

Familiar face

Chernoff's face

Motion

Fig. 1. Facial stimuli used in this study. In the upper panel, the high spatial frequency components of the face are filtered out by increasing the mosaic level. By increasing mosaic levels, it is difficult to perceive and identify the face. This picture is an example of familiar faces (Prime Minister Koizumi). In the middle panel, the high spatial frequency components of the face are pronounced in a simple drawing of the face (Chernoff’s face). By modulating the angles of the eyebrow and mouth, neutral, angry, and sad faces are easily made. In the lower panel, horizontal and radial optic flow stimuli are shown which were used to evoke motion perception. Four hundred white dots on a dark background were animated at a 60 Hz frame rate. Horizontal motion stimuli consisted of leftward or rightward movement of the dots. Radial motion consisted of dots moving in a radial pattern out from a focus of expansion on the horizontal meridian, 5⬚ to the left or right of center.

45 To study the perception of facial expressions, the angles of lines that make up the eyebrow and mouth of a Chernoff’s face, i.e. a simple drawing of the face with rich HSF components, were altered to make neutral, angry, or sad faces (Fig. 1, middle panel). The N100 component of the ERP in the occipital area and the N170 component in the posterior temporal region were the main components analyzed (Bentin et al., 1996). 2.2. Motion stimuli Stimuli with coherent horizontal (HO) motion and stimuli with radial optic flow (OF) motion were used to study motion perception (Fig. 1, lower panel). First, the psychophysical thresholds for detecting HO and OF motions were determined in healthy young and elderly subjects, and in patients with mild cognitive impairment (MCI) and Alzheimer’s disease (AD). Second, the motion coherence thresholds for HO and radial OF motions were determined using a left/right two alternative forced-choice discrimination technique (Mapstone et al., 2003). The perceptual threshold was defined by the percentage of coherent motion in the stimuli ((coherently moving dots)/(coherently moving dots + random dots) × 100) yielding 82% correct responses. The thresholds were related to the Weibull fit to psycophysical responses (Mapstone et al., 2003). From these results, a coherence level of 90% for HO and OF motions was used to elicit ERPs in normal subjects because it was sufficiently suprathreshold. Third, fMRIs were measured while normal subjects viewed the motion stimuli. A box-car design was employed and statistical analysis was performed by SPM99 (Taniwaki et al., 2003). 3. Results 3.1. Effects of low and high spatial frequency components of the face Unfiltered face stimuli (suprathreshold level) clearly evoked the N100 component in the occipital region and N170 component at posterior temporal sites (data not shown). The latencies of N170 were significantly shorter and their amplitudes significantly larger when

the mosaic level was decreased for familiar and unfamiliar faces. The differences in the latencies for N100 and N170 for familiar and unfamiliar faces were reduced by decreasing the mosaic levels, and the reduction was more pronounced for familiar faces. The Chernoff’s faces also elicited the N170 component at the temporal sites (Fig. 2A). The latencies of N170 for neutral and angry faces were significantly shorter and their amplitudes were signficantly larger than those for objects (Fig. 2A, upper column). Interestingly, a slow negative shift of the wave was observed over a 230–450 ms time period for angry faces but not for a neutral face (Fig. 2A, lower column). This negative shift was elicited when the presentation time of the Chernoff’s faces was set at 300 ms and less marked at 200 ms (data not shown). 3.2. Effects of motion direction Coherence visual motion thresholds were obtained for all subjects. The mean threshold for HO motion was significantly lower than that for OF motion in the young and elderly normal subjects, and in patients with MCI and AD. The mean coherence thresholds for OF motion increased in the following order; elderly, MCI, AD. In the ERP study, there were two major components elicited by motion stimuli; an N170 component and a P200 component (Fig. 2B, upper column). In the parietal area, the N170 component was evoked by HO motion but less so by OF motion, and showed an adaptational effect for random motion (Fig. 2B, lower column). On the other hand, the P200 component was elicited only by the OF motion and did not show an adaptational effect (Fig. 2B, lower column). fMRI studies showed that both HO and OF motion stimuli activated MT/V5 (data not shown), and interestingly, the superior parietal lobule was also activated by OF motion. 4. Discussion Investigations of the psychophysical thresholds are useful for evaluating the behavior of subjects. ERPs have excellent temporal resolution while fMRIs offer

46 A. ERPs to face and object stimuli

B. ERPs to motion stimuli random horizontal radial(in) radial(out)

Neutral Angry Object

N170

N170

5 μV

0

200

400

600 ms

1μV

P200

0

200

400

600 ms

horizontal radial(in) radial(out)

Neutral Angry N170

5 μV

1μV

P200

Negative shift

0

200

400

600 ms

0

200

400

600 ms

Fig. 2. Event-related potentials to face stimuli recorded at T6 (A) and motion stimuli obtained at P4 (B). The N170 component in response to face is maximal at posterior temporal regions bilaterally (T5 or T6, International 10–20 system). The N170 in response to object stimulus has a significantly longer latency and lower amplitude compared with that of the face stimuli (A, upper column). A slow negative shift of the wave can be seen over the 230–450 ms time period in the angry faces compared to a neutral face (A, lower column). Motion stimuli also evoke N170 but this component is maximal over the parietal region (P3 or P4, International 10–20 system) (B). N170 in response to either horizontal or random motion is much higher in amplitude compared with radial optic flow (B, upper column). In contrast, the later component, P200, is only recorded in response to radial optic flow. N170 evoked by HO shows an adaptational effect for random motion while P200 elicited by OF reveals no adaptational effect (B, lower column).

47 excellent spatial resolution. Hence, these three methods were combined to evaluate facial and motion perception in humans non-invasively in this study. 4.1. Face perception Information on different components of the face is transmitted mainly by the P-pathway and processed in the fusiform gyrus (V4) (Vuilleumier et al., 2003). Direct recordings from human V4 have demonstrated that a surface-negative potential (N200) is evoked by faces but not by the other types of stimuli (Allison et al., 1994, 1999). Scalp-recorded ERPs have shown that the N170 component is a face-specific potential, and that it is predominant in the posterior temporal cortex (Bentin et al., 1996). More specifically, it was considered to be generated in the occipitotemporal sulcus lateral to V4 (Bentin et al., 1996). In agreement with this, the N170 component recorded in this study was the largest over the posterior temporal region. Our finding that the latencies of N170 were significantly shorter and their amplitudes significantly larger with lower mosaic levels for both familiar and unfamiliar faces suggests that the HSF components are related to face perception for both familiar and unfamiliar faces. We also found that the difference in the latencies for N100 and N170 decreased for familiar and unfamiliar faces as the mosaic levels decreased, and that the reduction was more pronounced for familiar faces. These findings indicate that familiarity can facilitate the cortico-cortical processing of facial perception. Stimulation by Chernoff’s faces also elicited the N170 component at the temporal sites. Our findings of significantly shorter latencies and larger amplitudes for the N170 components with neutral and angry faces than for objects suggest that the recognition of facial expressions occurs during 230–450 ms after the appearance of face, and that the HSF components of the face are crucial for the recognition of facial expressions. 4.2. Motion perception Information about a moving stimulus is carried by the M-pathway and is processed in MT/V5 (Rizzolatti and Matelli, 2003). The lower thresholds for HO motion

for OF for each test group suggested that aging and visuospatial impairment affect motion perception. These findings are consistent with a recent study by Mapstone et al. (2003) and partly support our VEP finding that patients with AD have a temporal frequency deficit (Tobimatsu et al., 1994). A recent magnetoencephalographic study has shown that human V5 is activated by HO motion made up of random dot kinematograms (Nakamura et al., 2003). The response latencies at 100% coherence level ranged from 175 to 250 ms in both hemispheres. Our results showed that HO motion elicited an N170 component in the parietal area and had an adaptational effect for random motion while a P200 component was elicited by OF without an adaptational effect. Our observations suggest that the perception of HO and OF are segregated and sequentially processed. In agreement with this suggestion, fMRI studies showed that both HO and OF motion stimuli activated MT/V5 with the superior parietal lobule activated by only OF motion. These findings are in accord with the recent view of the importance of the parietal lobe for OF motion perception (Ptito et al., 2001). In conclusion, our integrated approach provided useful information on spatial and temporal processing of face and motion non-invasively. Further studies will determine the roles of such visual channels for processing face and motion. 5. Acknowledgement This study was supported in part by Grant-in-Aid for the 21st Century COE Program and Grant-in-Aid for Scientists, No 16390253 and No 16200005 from the Ministry of Education, Culture, Sports, Science and Technology in Japan. References Allison, T., Ginter, H., McCarthy, G., Nobre, A.C., Puce, A., Luby, M. and Spencer, D.D. (1994) Face recognition in human extrastriate cortex. J. Neurophysiol., 71: 821–825. Allison, T., Puce, A., Spencer, D.D. and McCarthy, G. (1999) Electrophysiological studies of human face perception. I: Potentials generated in occipitotemporal cortex by face and non-face stimuli. Cereb. Cortex, 9: 415–430. Arakawa, K., Tobimatsu, S., Kato, M. and Kira, J. (1999) Parvocelluar and magnocellular visual processing in spinocerebellar

48 degeneration and Parkinson’s disease: an event-related potential study. Clin. Neurophysiol., 110: 1048–1057. Bentin, S., Allison, T., Puce, A., Perez, E. and McCarthy, G. (1996) Electrophysiological studies of face perception in humans. J. Cogn. Neurosci., 8: 551–565. Livingstone, M. and Hubel, D. (1988) Segregation of form, color, movement, and depth: anatomy, physiology, and perception. Science, 240: 740–749. Mapstone, M., Steffenella, T.M. and Duffy, C.J. (2003) A visuospatial variant of mild cognitive impairment. Getting lost between aging and AD. Neurology, 60: 802–808. Nakamura, H., Kashii, S., Nagamine, T., Matsui, Y., Hashimoto, T., Honda, Y. and Shibasaki, H. (2003) Human V5 demonstrated by magnetoencephalography using random dot kinematograms of different coherence levels. Neurosci. Res., 46: 423–433. Ptito, M., Kupers, R., Faubert, J. and Gjedde, A. (2001) Cortical representation of inward and outward radial motion in man. Neuroimage, 14: 1409–1415. Rizzolatti, G. and Matelli, M. (2003) Two different streams from the dorsal visual system: anatomy and functions. Exp. Brain Res., 153: 146–157. Taniwaki, T., Okayama, A., Yoshiura, T., Nakamura, Y., Goto, Y., Kira, J. and Tobimatsu, S. (2003) Reappraisal of the motor role of basal ganglia: a functional magnetic resonance imaging study. J. Neurosci., 23: 3432–3438.

Tobimatsu, S. (2002) Neurophysiologic tools to explore visual cognition. Electroencephalogr. Clin. Neurophysiol., S54: 261–265. Tobimatsu, S. and Kato, M. (1998) Multimodality visual evoked potentials in evaluating visual dysfunction in optic neuritis. Neurology, 50: 715–718. Tobimatsu, S., Hamada, T., Okayama, M., Fukui, R. and Kato, M. (1994) Temporal frequency deficit in patients with senile dementia of the Alzheimer type: a visual evoked potential study. Neurology, 44: 1260–1263. Tobimatsu, S., Tomoda, H. and Kato, M. (1995) Parvocellular and magnocellular contributions to visual evoked potentials in humans: stimulation with chromatic and achromatic gratings and apparent motion. J. Neurol. Sci., 134: 73–82. Tobimatsu, S., Shigeto, H., Arakawa, K. and Kato, M. (1999) Electrophysiological studies of parallel visual processing in humans. Electroencephalogr. Clin. Neurophysiol., S49: 103–107. Tobimatsu, S., Celesia, G.G., Haug, B.A., Onofrj, M., Sartucci, F. and Porciatti, V. (2000) Recent advances in clinical neurophysiology of vision. Electroencephalogr. Clin. Neurophysiol., S53: 312–322. Vuilleumier, P., Armony, J.L., Driver, J. and Dolan, R.J. (2003) Distinct spatial frequency sensitivities for processing faces and emotional expressions. Nature Neurosci., 6: 624–631.


49

Chapter 7

The role of the basal ganglia and cerebellum in cognitive impairment: a study using event-related potentials Koichi Hirataa,*, Hideaki Tanakaa, Xiao-Hui Zenga,b, Akinori Hozumia and Mio Araia a Department of Neurology, Dokkyo University School of Medicine, Tochigi 321-0293 (Japan) Neurological Sciences Institute, Oregon Health and Sciences University, 505 NW 185th Avenue, Beaverton, OR 97006 (USA)

b

1. Introduction Recent studies have indicated the involvement of the basal ganglia and cerebellum in cognitive function. Cognitive dysfunction in Parkinson’s disease (PD) is considered to have a particular relationship with frontal lobe activity (Piccirili et al., 1989). Mental changes related to frontal lobe dysfunction are present even in mild PD. In addition, PD deficits have been attributed by some to behavioral disturbances; one type that affects cognitive function in PD is commonly described as failure of maintaining or switching the set (Lees and Smith, 1983; Flowers and Robertson, 1985; Taylor et al., 1990; Hozumi et al., 2000; Zeng et al., 2002). In contrast, learning impairment caused by abnormal functioning of the basal ganglia in PD has also been described (Kimura, 1995). The involvement of the cerebellum in cognitive function has been indicated, in addition to its well-known

*Correspondence to: Dr. Koichi Hirata, Department of Neurology, Dokkyo University School of Medicine, 880 Kitakobayashi, Mibumachi, Shimotsuka-gun, Tochigi 3210293, Japan. E-mail: [email protected]

role in motor function. Employing various neuropsychological tests, previous reports have described impairment of generalized intellect (Kish et al., 1994), spatial cognition (Wallesch and Horn, 1990; Kish et al., 1994) and frontal lobe function (Grafman et al., 1992; Appollonio et al., 1993; Kish et al., 1994) in patients with cerebellar degeneration and stroke. Cerebellar activation is often evident in tests of cognitive planning (Kim et al., 1994), shifting attention (Allen et al., 1997), sustained attention (Pardo et al., 1991), and semantic retrieval (Sakurai et al., 1993), and in functional neuroimaging studies such as PET fMRI and EEG LORETA (Arai et al., 2003; Tanaka et al., 2003). Event-related potentials (ERPs) – P3 (P300) in particular – can be used to test cognitive performance in order to evaluate mental state. P3 is considered an electrophysiological parameter of the cognitive processes of attention, memorizing, and discrimination of stimuli (Goodin et al., 1978). In addition to the P3 oddball paradigm, a cued continuous performance task (CPT) paradigm, which was developed to quantify sustained attention and to validate frontal brain function in 1956 (Rosvold et al., 1956), has been recently modified for use in ERP studies (Strik et al., 1998). In this cued CPT, the cue induces the subject to anticipate a motor reaction, which is executed after the target (Go),

50 and suppressed after non-target stimuli (NoGo). Furthermore, the NoGo stimuli have been interpreted as a reflection of a response inhibition process (Pfefferbaum et al., 1985). The purpose of this study was to determine the profile and difference of cognitive impairment in disturbance of basal ganglia and cerebellum using ERPs and a battery of neuropsychological tests.

2. Methods 2.1. Subjects Twenty idiopathic Parkinson’s disease (PD) patients, 13 patients with cortical cerebellar atrophy (CCA) and 20 age-matched normal controls (NC) participated (Table 1). The PD presented at least two of the cardinal features (tremor, rigidity, and bradykinesia). Disease severity ranged between I and III on the Hoehn–Yahr scale. The PD enrolled had no history of any other neurological disease or of drug abuse. Those with any evidence of focal lesions on CT and/or MRI were excluded. Antiparkinsonian drugs continued to be administered. All the patients were evaluated by the Minimental state examination (MMSE), which measures the severity of intellectual deterioration, a cutoff score of less than 24 being the index for cognitive impairment.

All CCA patients satisfied following criteria: (1) The primary symptom was that of cerebellar ataxia without pyramidal signs, extrapyramidal signs, or autonomic nervous dysfunction. (2) Cerebellar atrophy without brainstem atrophy was confirmed by neuroimaging, with no abnormal signs within cerebral and other brain structures being observed. (3) Other causes of cerebellar atrophy were excluded, including long-term alcohol addiction, anticonvulsant poisoning, hypothyroidism, and malignant tumors. (4) Absence of a positive family history. Normal control subjects represented healthy individuals who visited our clinic for routine medical check ups and were found to be free of any neurological and psychiatric diseases. Furthermore, neurological examination, including MRI, was normal in these subjects. Neither the control subjects nor CCA subjects were taking any central nervous system (CNS)-active drugs at the time of testing. The control participants had neither neurological nor psychiatric disease and were not taking any centrally active medication. None of them complained of subjective memory impairment. No clinical evidence of depression was present. We studied all subjects after having given their informed consent, which was approved by institutional

TABLE 1 BACKGROUND OF PARTICIPATING SUBJECTS

No. of subjects (N) Sex (male/female) Age (years, mean ± SD) Duration of disease (years, mean ± SD) Duration of education (years, mean ± SD) MMSE (mean ± SD) WCST CA (mean ± SD) PEN (mean ± SD) p < 0.05.

*

PD

CCA

Normal

20 11/9 60.6 ± 10.8 – 11.3 ± 3.9 28.3 ± 2.0

13 7/6 59.1 ± 10.0 4.08 ± 2.3 12.6 ± 2.6 27.7 ± 2.5

20 11/9 61.4 ± 12.3 8.4 ± 7.6 11.4 ± 2.3 29.2 ± 1.5

3.6 ± 2.2 6.2 ± 5.0

3.8 ± 1.0 4.7 ± 2.1

1.8 ± 1.4* 8.7 ± 5.0

51 review boards of Dokkyo University School of Medicine. 2.2. Neuropsychological testing The new modified Wisconsin card sorting test (WCST; Kashima et al., 1985) was used to evaluate the difficulty in the change of category or “set” (perseveration of a higher level) in the PD, CCA, and NC. The WCST is one of the few tests that can detect a clear deficit, specific to patients with frontal lobe dysfunction, but for many patients the original WCST was too difficult and distressing. A simpler and less ambiguous modification was therefore made. The two main points modified were the order of the reaction cards and the process of giving instructions. This new version has been used successfully in Japan. To evaluate this test, the maximum classification score and the perseverative errors reported by Nelson (1976) were used. 2.3. Stimuli and experimental conditions A conventional auditory oddball task and visual CPT paradigm were utilized for ERP recordings. The subject lay down on a bed in a sound-attenuated and dimly lit room, and was presented with a sequence of binaural stimuli via earphones for the auditory oddball paradigm. This comprised 100 ms tone bursts (10 ms rise/fall time) of 80 dB at 1.5 s intervals; in random sequence, 20% of the tones with a pitch of 2000 Hz (rare “target” tones), and 80% of the tones with a pitch of 1000 Hz (frequent “non-target” tones). The subject was instructed to close the eyes and silently count the target tones without using their fingers for counting; at the end of the session, they were asked to report the count (“counting performance value”); this was acknowledged without information about errors. Subjects were given a brief practice session to ensure that they understood the instructions and could discriminate the tones. Stimuli were presented until 20 artifact-free targets were collected. After a short break, the session was repeated. For the CPT paradigm, the subject was seated in a comfortable chair with the head fixed in a forehead– chin rest 60 cm in front of a 14-inch computer monitor. Five Japanese vowels such as [ ] (a), [ ] (i),

[ ](u), [ ](e), [ ](o) with hiragana (syllabograms) letters were used. The size of the letters on the monitor was 5 cm × 5 cm, and they were presented sequentially in random order. Each letter was presented for 200 ms in the center of the monitor, and the inter-trial interval was random at 1610–2490 ms. The letter [ ] was presented as a signal to prepare a motor response. Subjects were instructed to press a button using the index finger of their right hand as fast as possible whenever the letter [ ] was followed by the same letter (Go condition). When the other letters [ ], [ ], [ ], or [ ] followed an [ ], the subject was instructed not to press a button (NoGo condition) (Fig. 1). In each session, we used 200 letters; 15 (7.5%) letters for the Go condition and 32 (16%) letters for the NoGo condition (Fig. 1). 2.4. Recording of event-related potentials The electroencephalogram was recorded using 20 channels according to the International 10/20 system (Fp1/2, Fz, F3/4, F7/8, T3/4, C3/4, Cz, P3/4, Pz, T5/6, O1/2, Oz) with an electrode behind the ear as the reference. The signals were processed with an automatic artifact rejection system and computed into series of potential distribution maps using a BioLogic Brain Atlas. The ERPs recorded over 1024 ms post-stimulus were averaged online at 250 samples/s, using an automatic artifact rejection for signals higher than 136 μV. The ERPs for the two sessions were separately averaged for the conventional oddball and the CPT task. The recording was performed at 14.00 hours in all subjects to exclude the effects of circadian rhythm. 2.5. Spatial analysis In the first step, ERP voltages were transformed into reference-independent values by recomputation of the voltages vs. average reference. In order to quantify the amount of spatial relief or hilliness of the topographic fields, the scalp electric potential power or global field power (GFP) as defined by Lehmann and Skrandies (1980) was calculated. Based on these GFP curves, latencies of the P3 component of each ERP recording were determined. P3 latency was defined as the

52

Fig. 1. CPT paradigm.

maximum GFP peak within 280–480 ms. Amplitudes of ERP components were measured at these latencies and displayed on topographic maps. 2.6. Statistical evaluation Differences in parameters derived from ERPs and psychological tests were examined for statistical significance using an unpaired sample t test. Doubleended p values are reported. These descriptive p values were considered relevant if they appeared in nearly regular patterns associated with numerically relevant differences. Data are expressed as mean ± standard deviation. 3. Results 3.1. Neuropsychological examination In the WCST examination, PD showed significant decreases in the categories evaluated and increases in perseverative errors as compared to the NC. On the other hand, CCA showed no significant change in the WCST as compared to the NC. MMSE scores did not differ significantly between the 3 groups (Table 1).

3.1.1. P3 GFP latency and strength (global field power (GFP)) Figure 2 gives the differences of P3 GFP strength and latency between patients with PD, CCA and controls, for target tones in oddball paradigm, Go and NoGo conditions in CPT paradigm. P3 GFP latency did not show any differences between three groups for the Go condition in CPT or for target stimuli in 2-tones paradigm. P3 GFP strength increased for the auditory oddball paradigm target and for the Go condition of CPT paradigm in PD compared with the controls. The P3 GFP strength significantly decreased in PD group for the NoGo condition compared with the controls. In CCA patients, GFP peak latency was prolonged and GFP strength was attenuated in the NoGo condition compared with the control subjects, but there were no differences in auditory oddball task or in Go condition compared with the control subjects (Fig. 2). 4. Discussion The purpose of the present study was to investigate, electrophysiologically, disturbance of frontal lobe function resulting from basal ganglia or cerebellum dysfunction.

53

Fig. 2. The differences of P3 GFP strength and latency between patients with PD, CCA, and controls for target tones in oddball paradigm Go and NoGo conditions.

Various studies have established that patients with PD develop neuropsychological deficits across a range of cognitive function commonly attributed to frontal lobe dysfunction (Taylor et al., 1986; Brown and Marsden, 1988; Cooper et al., 1991). These include planning, selective attention, set shifting, and behavior learning, which has been attributed to frontal dopamine deficiency or to frontal-subcortical circuit dysfunction (Lees and Smith, 1983; Dagher et al., 2001). Increased P3 GFP amplitude for target tone in oddball and Go in CPT paradigm in patients with PD suggest that they have insufficient processing resources (Hozumi et al.,

2000; Zeng et al., 2002). In addition, our results showed a disturbance of inhibitory function in patients with PD during the NoGo condition when performing the CPT paradigm. Inhibition is an important contributory process in successful selective attention, and selective attention has been shown to be disrupted by frontal lobe dysfunction. Both human and animal studies have demonstrated that prefrontal damage disrupts inhibitory modulation (Edinger et al., 1975; Skinner and Yingling, 1977). It has been suggested that prefrontal lobe has a critical role in the process of inhibitory control (Knight et al., 1999; Metzler and Parkin, 2000). The recently introduced Go/NoGo task reflects spatial brain electrical changes in relation to execution and inhibition of a motor response elicited with a CPT; execution of the “push” and “wait” tasks evokes “Go” and “NoGo” P3 components. The WCST is considered an attentional set shift task, which is sensitive to frontal dysfunction, and has been accepted for evaluating frontal lobe dysfunction. It is one of the few tests that detect a clear deficit, specific to patients with frontal lobe dysfunction, but for many patients the original WCST was too difficult and distressing. Therefore, a simpler and less ambiguous version was made. The two main points modified were the order of the reaction cards and the process of giving instructions. This new version has been used successfully in Japan (Kashima et al., 1985). A few authors (Tachibana et al., 1995; Kamitani et al., 1998; Yamaguchi et al., 1998) have reported ERP findings in spinocerebellar degeneration (SCD) such as CCA and these results are controversial. Such differences in P3 findings may be due to the use of different paradigms for P3 measurement and/or conducting such studies in heterogeneous samples of SCD patients. Our results showed no significant difference in the P3 component of the auditory oddball paradigm between CCA patients and control subjects. On the other hand, CCA patients showed abnormalities only under NoGo condition. A reason for the failure to find any difference between patients and controls with respect to the P3 component of the auditory oddball paradigm might lie in the main source; mainly temporal lobe (Tarkka and Stokic, 1998). It is also worth pointing out the existence of a cerebellar projection to higher order areas in

54 the prefrontal cortex (Leiner et al., 1986), but not in the temporal cortex. Considered together, the results of these previous and present studies suggest that damage restricted to the cerebellum is functionally associated with impairment of inhibition of prepared motor response, and they provide evidence for frontal lobe dysfunction, especially in premotor areas. Cerebellar dysfunction in CCA patients would be expected to influence certain aspects of cognitive dysfunction. However, it is often difficult to identify such abnormality with ordinary psychological tests such as WCST. Our study using ERP and CPT paradigms demonstrated the advantages of these techniques in detecting abnormalities caused by isolated impairment of the cerebellum, and also that such impairment is characterized by dysfunction of the inhibitory system. Both PD and CCA have inhibitory function disturbance, which may be based on frontal lobe, especially prefrontal lobe, dysfunction. However, in PD it could be based on insufficient processing, and in CCA it could not be detected using ordinary psychological tests such as WCST, and this is the essential psychophysiological difference between the two conditions. 5. Conclusion Our studies suggest that basal ganglia and cerebellum have different roles in cognitive function related to frontal lobe function. 6. Acknowledgements and funding This study was supported in part by a Smoking Research Foundation. References Allen, G., Buxton, R.B., Wong, E.C. and Courchesne, E. (1997) Attentional activation of the cerebellum independent of motor involvement. Science, 275: 1940–1943. Appollonio, I.M., Grafman, J., Schwartz, V., Massaquoi, S. and Hallett, M. (1993) Memory in patients with cerebellar degeneration. Neurology, 43: 1536–1544. Arai, M., Tanaka, H., Pascual-Marqui, R.D. and Hirata, K. (2003) Reduced brain electric activities of frontal lobe in cortical cerebellar atrophy. Clin. Neurophysiol., 114: 740–747.

Brown, R.G. and Marsden, C.D. (1988) Internal versus external cues and the control of attention in Parkinson’s disease. Brain, 111: 323–345. Cooper, J.A., Sagar, H.J., Harvey, N.S., Jordan, N. and Sullivan, E.V. (1991) Cognitive impairment in early, untreated Parkinson’s disease and its relationship to motor disability. Brain, 114: 2095–2122. Dagher, A., Owen, A. M., Boecker, H. and Brooks, D.J. (2001) The role of the hippocampus in planning: a PET activation study in Parkinson’s disease. Brain, 124: 1020–1032. Edinger, H.M., Siegel, A. and Troiano, R. (1975) Effect of stimulation of prefrontal cortex and amygdala diencephalic neurons. Brain Res., 97: 272–282. Flowers, K.A. and Robertson, C. (1985) The effect of Parkinson’s disease on the ability to maintain a mental set. J. Neurol. Neurosurg. Psychiatr., 48: 517–529. Goodin, D.S., Squires, K.C. and Starr, A. (1978) Long latency event-related components of the auditory evoked potential in dementia. Brain, 101: 635–648. Grafman, J., Litvan, I., Massaquoi, S., Stewart, M., Sirigu, A. and Hallett, M. (1992) Cognitive planning deficit in patients with cerebellar atrophy. Neurology, 42: 1493–1496. Hozumi, A., Hirata, K., Tanaka, H. and Yamazaki, K. (2000) Perseveration for novel stimuli in Parkinson’s disease, an evaluation based on event-related potentials topography. Mov. Disord., 15: 835–842. Kamitani, T., Kuroiwa, Y. and Wang, L. (1998) Visual event-related potentials in patients with spinocerebellar degeneration. In: I. Hashimoto and R. Kakigi (Eds.), Recent Advances in Human Neurophysiology. Elsevier, Amsterdam, pp. 515–526. Kashima, H., Kato, M. and Handa, T. (1985) A modified Wisconsin card sorting test – a comparison of the chronic schizophrenics to the patients with frontal lesions. Folia Psychiatr. Neurol. Jpn., 39: 97. Kim, S.G., Ugurbil, K. and Strick, P.L. (1994) Activation of a cerebellar output nucleus during cognitive processing. Science, 265: 949–951. Kimura, M. (1995) Role of basal ganglia in behavioral learning. Neurosci. Res., 22: 353–358. Kish, S.J., El-Awar, M., Stuss, D., Nobrega, J., Currier, R., Aita, J.F., Schut, L., Zoghbi, H.Y. and Freedmann, M. (1994) Neuropsychological test performance in patients with dominantly inherited spinocerebellar ataxia, relationship to ataxia severity. Neurology, 44: 1738–1746. Knight, R.T., Staines, W.R., Swick, D. and Chao, L.L. (1999) Prefrontal cortex regulates inhibition and excitation in distributed neural networks. Acta Psychol., 101: 157–178. Lees, A.J. and Smith, E. (1983) Cognitive deficits in the early stages of Parkinson’s disease. Brain, 106: 257–270. Lehmann, D. and Skrandies, W. (1980) Reference-free identification of components of checkerboard evoked multichannel potential fields. Electroencephalogr. Clin. Neurophysiol., 48: 609–621. Leiner, H.C., Leiner, A.L. and Dow, R.S. (1986) Does the cerebellum contribute to mental skills? Behav. Neurosci., 100: 443–454. Metzler, C. and Parkin, A.J. (2000) Reversed negative priming following frontal lobe lesions. Neuropsychologia, 38: 362–379. Nelson, H.E. (1976) A modified card sorting test sensitive of frontal lobe defects. Cortex, 12: 313–324.

55 Pardo, J.V., Fox, P.T. and Raichle, M.E. (1991) Localization of a human system for sustained attention by positron emission tomography. Nature, 349: 61–64. Pfefferbaum, A., Ford, J.M., Weller, B.J. and Kopell, B.S. (1985) ERPs to response production and inhibition. Electroencephalogr. Clin. Neurophysiol., 59: 85–103. Piccirili, M., D’Alessandro, P., Finali, G., Piccinin, G.L. and Agostini, L. (1989) Frontal lobe dysfunction in Parkinson’s disease, prognostic value for dementia? Eur. Neurol., 29: 71–76. Rosvold, H.E., Mirsky, A.F., Sarason, I., Bransome, E.D. and Beck, L.H. (1956) A continuous performance test of brain damage. J. Consult. Psychol., 20: 343. Sakurai, Y., Momose, T., Iwata, M., Watanabe, T., Ishikawa, T. and Kanazawa, I. (1993) Semantic process in kana word reading, activation studies with positron emission tomography. Neuroreport, 4: 327–330. Skinner, J.E. and Yingling, C.D. (1977) Central gating mechanisms that regulate event-related potentials and behavior. Prog. Clin. Neurophysiol., 1: 30–69. Strik, W.K., Fallgatter, A.J., Brandeis, D., Pascual-Marqui, R.D. (1998) Three-dimensional tomography of event-related potentials during response inhibition, evidence for phasic frontal lobe activation. Electroencephalogr. Clin. Neurophysiol., 108: 406–413.

Tachibana, H., Aragane, K. and Sugita, M. (1995) Event-related potentials in patients with cerebellar degeneration, electrophysiological evidence for cognitive impairment. Cogn. Brain Res., 2: 173–180. Tanaka, H., Harada, M., Arai, M. and Hirata, K. (2003) Cognitive dysfunction in cortical cerebellar atrophy correlates with impairment of the inhibitory system. Neuropsychobiology, 47: 206–211. Tarkka, I.M. and Stokic, D.S. (1998) Source localization of P300 from oddball, single stimulus, and omitted-stimulus paradigms. Brain Topogr., 11: 141–151. Taylor, A.E., Saint-Cyr, J.A. and Lang, A.E. (1986) Frontal lobe dysfunction in Parkinson’s disease. Brain, 109: 845–883. Taylor, A.E., Saint-Cyr, J.A. and Lang, A.E. (1990) Memory and learning in early Parkinson’s disease, evidence for a “frontal lobe syndrome.” Brain Cogn., 13: 211–232. Wallesch, C.W. and Horn, A. (1990) Long-term effects of cerebellar pathology on cognitive functions. Brain Cogn., 14: 19–25. Yamaguchi, S., Tsuchiya, H. and Kobayashi, S. (1998) Visuospatial attention shift and motor responses in cerebellar disorders. J. Cogn. Neurosci., 10: 95–107. Zeng, X.H., Hirata, K., Tanaka, H., Hozumi, A. and Yamazaki, K. (2002) Insufficient processing resources in Parkinson’s disease, evaluation using multimodal event-related potentials paradigm. Brain Topogr., 14: 299–311.


57

Chapter 8

High-frequency oscillatory activities during selective attention in humans Isamu Ozakia,*, Yukoh Yaegashib, Masayuki Babac and Isao Hashimotod a

Faculty of Health Sciences, Aomori University of Health and Welfare, Aomori (Japan) b Department of Social Science, Akita Keijoh College, Oodate (Japan) c Department of Neurological Sciences, Hirosaki University School of Medicine, Hirosaki (Japan) d Human Information Systems Laboratory, Kanazawa Institute of Technology, Tokyo (Japan)

1. Introduction

2. Subjects and methods

The frequency of spontaneous electroencephalogram (EEG) is known to change with the level of vigilance: the low-frequency alpha wave is dominant at rest and the high-frequency gamma oscillation (mainly 30–40 Hz) appears at a high vigilance level. However, little attention has been given to changes in very high-frequency (>100 Hz) activities. Recent studies on somatosensory evoked magnetic fields or potentials disclosed that the initial cortical response contains a high-frequency (HF) component at around 600 Hz (Curio et al., 1994; Hashimoto et al., 1996; Ozaki et al., 1998), and the power of HF component changes during a sleep–waking cycle (Yamada et al., 1988; Hashimoto et al., 1996); however, there have been few reports on HF component changes during attentive tasks. We therefore tested whether EEG activities at extreme HF ranges alter in selective attention to stimuli using somatosensory evoked potentials (SEPs).

Seven normal adults (3 men, 4 women) participated in the experiments. The mean age was 29 years (range 20–45 years). The subject sat relaxed in a comfortable reclining chair in a quiet, air-conditioned, and electrically shielded room. During the recording session, the subject was encouraged to minimize muscle and eye blink interference and was kept awake. All the subjects gave their informed consent. Brief electric shock (0.2 ms duration) was applied to the right median nerve at the wrist at a rate of 5 stimuli/s. Intensity was adjusted so as to induce a small muscular twitch in the thenar muscles. SEPs were recorded from the frontal scalp at Fz (International 10–20 system), the ipsilateral central scalp at C4, and six electrodes covering the left centro-parietal scalp at FC3 (3 cm anterior to C3), C3, CP3 (3 cm posterior to C3), and 2 cm medial to FC3, C3, or CP3 (see Fig. 1). The right ear served as a reference. SEPs of 100 ms (5 ms before stimuli) were digitized at a 20 kHz sampling rate. In attentive sessions, we omitted the stimuli in a random sequence, and the subject was instructed to count omitting stimuli in their head. In neglect sessions, the subject listened to their favorite music to try to ignore the electric stimuli. For each run, about 1000 trials

*Correspondence to: Dr. Isamu Ozaki, Faculty of Health Sciences, Aomori University of Health and Welfare, 58-1 Mase, Hamadate, Aomori 030-8505, Japan. Tel/Fax: +81-17-765-2070; E-mail: [email protected]

58

Fig. 1. The position of the 8 electrodes over the scalp. Channel 7 (Ch 7): Fz; Ch 1: 3 cm anterior and 2 cm medial to C3; Ch 2: 3 cm anterior to C3; Ch 3: 2 cm medial to C3; Ch 4: C3; Ch 5: 3 cm posterior and 2 cm medial to C3; Ch 6: 3 cm posterior to C3; Ch 8: C4.

were averaged with a Dantec Keypoint electromyograph. Data were stored on floppy disks and converted to IBM format for later off-line analysis. We recorded 5–7 runs for attentive or neglect sessions. Then, grand averaging of attentive or neglect sessions was performed in each subject. The original filter setting was 0.5–2000 Hz. In off-line analysis, using a Hamming window, digital bandpass filtering of 300–1000 Hz was performed to separate high frequency oscillations (HFOs) from the underlying initial cortical response such as the N20 potential. Using a maximum entropy method (Ulrych, 1972), a power spectrum with a 10 ms period was calculated as the logarithm of power squared in base 10 for each frequency at every 1 ms interval. A time–frequency analysis of HF components was then obtained. For statistical analysis of HF components, we performed 3-way (channel × subject × condition) ANOVA. 3. Results Figure 2 shows SEP traces recorded from the left parietal scalp in a representative subject. There were no significant changes in the original wide-band traces between attentive and neglect conditions. However, for the post-filtered traces, HF activities at 300–400 Hz were dominant under attentive conditions. They appeared after 30 ms and lasted up to 60 ms under

Fig. 2. Effect of attention on SEPs to right median nerve stimulation in a 23-year-old woman. Top panel: the original SEP traces recorded from the left parietal scalp are compared between attentive and neglect conditions. The thick trace shows the attentive condition, and the thin trace, the neglect condition. Middle and bottom panels: the postfiltered traces of 8 channels for the attentive condition (middle) and neglect condition (bottom) are superimposed. Note that although HFOs overlying N20 seem unchanged, high-frequency activity at 300–400 Hz appears after 30 ms and lasts up to 60 ms in the attentive condition (small arrows).

attentive conditions, although HFOs overlying N20 seemed unchanged. Although the HF activities appeared at around 60 ms under neglect conditions, they did not last long compared with those under attentive conditions. In Fig. 3, the results of time–frequency analysis for the left parietal scalp recording in the same subject as in Fig. 1 are illustrated. HF activity at approximately 350 Hz was discerned under attentive conditions. HF activity also appeared at around 30 ms and lasted up to 55 ms in other recording channels under attentive conditions. For HFOs overlying N20–P20 and central P22 potentials, there were no obvious changes between attentive and neglect conditions in the subjects examined.

59 and a mean appearance time of 40 ms with 95% confidence interval between 35 and 45 ms. Figure 4 shows the time–frequency analysis of the left fronto-central scalp recording of the grand averaged data for 7 subjects. The dominant HF activity at around 350 Hz was discerned under attentive conditions. This appeared at around 30 ms and lasted up to 65 ms. This finding was consistent in the other recording channels. 4. Discussion

Fig. 3. Time–frequency representation of channel 6 at the left parietal scalp (obtained from the same subject as in Fig. 1). Upper panel: attentive session. Lower panel: neglect session. Ordinate – frequency; abscissa – time after the onset of stimuli. Note that relative signal energy powers at around 350 Hz are dominant between 30 and 55 ms in the attentive condition.

On the other hand, the signal strengths of HF activities following HFOs overlying N20–P20 components were increased under attentive conditions compared to neglect conditions. Although there was inter-individual variety, the frequency range of the HF activities was between 300 and 450 Hz and the appearance time was between 30 and 60 ms. For statistical analysis of attention-related HF activities, we therefore focused on 300–450 Hz frequencies and the period between 30 and 60 ms. We then obtained the maximal power from 7 subjects to be analyzed. Three-way (channel × subject × condition) ANOVA showed a significant main effect for subject and condition: for subject, F (6, 42) = 85.879, p = 5.74 × 10−22 and for condition, F (1, 42) = 172.001, p = 1.92 × 10−16. For the interaction term, the subject and the condition interaction term (F (6, 42) = 7.972, p = 9.04 × 10−6) alone was found significant. By analyzing the maximal signals of HF activities under attentive conditions for all 7 subjects, we obtained the characteristics of attention-related HF activity: a mean frequency of 355 Hz with 95% confidence interval between 335 and 376 Hz

We have found that, following HFOs overlying the N20–P20 component, HF activity at approximately 350 Hz appeared in human SEPs. Although there was inter-individual variation, the HF activity changed significantly due to the conditions; in other words, attention augments HF activity. This attention-related HF activity was found at 35–45 ms after stimulation and lasted up to 60 ms. HF activity following N20–P20 or N20m has attracted little attention in research into SEPs or somatosensory evoked fields (SEFs). However, in a recent SEP study on cortical HFOs by Restuccia’s

Fig. 4. Time–frequency representation of channel 2 at the left fronto-central scalp (obtained from the grand averaged data). Upper panel: attentive session. Lower panel: neglect session. Ordinate – frequency; abscissa – time after the onset of stimuli. Note that relative signal energy powers at around 350 Hz are dominant between 30 and 65 ms in the attentive condition.

60 group, they showed HF activity after 30 ms (see Fig. 1 in Restuccia et al., 2003). In SEF research by Haueisen et al. (2001), HF activity after 30 ms was also discerned in their Fig. 2. We therefore believe that HF activity after 30 ms appears consistently in SEP or SEF recordings. Significant inter-subject variation in the appearance time suggests that HF activity after 30 ms in SEPs or SEFs reflects responses not directly evoked by sensory stimuli but induced by neural circuit activity. Recent animal experiments showed that HF activity (so-called ripples) at 80–200 Hz overlies spontaneous EEG in the cat brain (Grenier et al., 2001) as well as SEPs in the rat cortex (Jones and Barth, 1999). In a study on the correlation of neural activity with HF activity, Grenier et al. (2001) found that the firing pattern of the first rhythmic burst (FRB) cell is correlated with HF activity (ripples) recorded from the cortical surface of area 7 of the cat brain. The FRB cell is known to give rise to highfrequency (300–600 Hz) spike bursts recurring at fast rates such as 30–50 Hz (Steriade, 2001). We therefore suppose that HF activity after 30 ms recorded in SEPs can be analogous to the ripples in the cat brain. The mechanism underlying augmentation of HF activity during attention is undetermined as yet. By analyzing single trial EEG during auditory attentive tasks, Winterer et al. (1999) found that EEG periods with fast reaction time performance are accompanied with a large amplitude of N100 response, and that there is an inverse relationship between the reaction time and noise power obtained by subtracting the averaged response from accumulated EEG strength. This suggests that a high vigilance level or good performance is correlated with a large amplitude evoked response and is accompanied by high levels of brain noise. Winterer et al. (1999) proposed a theory of stochastic resonance to explain this paradoxical phenomenon. When a specific input signal at or near the threshold level reaches the sensory system of a neural network, signals that are irrelevant to the specific sensory signal will augment output signals of the sensory system as stochastic resonance. In other words, a particular noise level can induce a multistable, excitable system to respond to a weak input signal. We therefore suppose that attention-related

HF activity may reflect this augmentation of neural noise, resulting in the facilitation of sensory information processing in a neural network as stochastic resonance. 5. Acknowledgement This research was partly supported by a Grant for Special Research Project, Aomori University of Health and Welfare. References Curio, G., Mackert, B.-M., Burghoff, M., Koetitz, R., AbrahamFuchs, K. and Härer, W. (1994) Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system. Electroencephalogr. Clin. Neurophysiol., 91: 483–487. Grenier, F., Timofeev, I. and Steriade, M. (2001) Focal synchronization of ripples (80–200 Hz) in neocortex and their neuronal correlates. J. Neurophysiol., 86: 1884–1898. Hashimoto, I., Mashiko, T. and Imada, T. (1996) Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex. Electroencephalogr. Clin. Neurophysiol., 100: 189–203. Haueisen, J., Schack, B., Meier, T., Curio, G. and Okada, Y. (2001) Multiplicity in the high-frequency signals during the shortlatency somatosensory evoked cortical activity in humans. Clin. Neurophysiol., 112: 1316–1325. Jones, M.S. and Barth, D.S. (1999) Spatiotemporal organization of fast (>200 Hz) electrical oscillations in rat vibrissa/barrel cortex. J. Neurophysiol., 82: 1599–1609. Ozaki, I., Suzuki, C., Yaegashi, Y., Baba, M., Matsunaga, M. and Hashimoto, I. (1998) High frequency oscillations in early cortical somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 108: 536–542. Restuccia, D., Della Marca, G., Valeriani, M., Rubino, M., Paciello, N., Vollono, C., Capuano, A. and Tonali, P. (2003) Influence of cholinergic circuitries in generation of high-frequency somatosensory evoked potentials. Clin. Neurophysiol., 114: 1538–1548. Steriade, M. (2001) Impact of network activities on neuronal properties in corticothalamic systems. J. Neurophysiol., 86: 1–39. Ulrych, T.J. (1972) Maximum entropy power spectrum of truncated sinusoids. J. Geophys. Res., 77: 1396–1400. Winterer, G., Ziller, M., Dorn, H., Frick, K., Mulert, C., Dahhan, N., Herrmann, W.M. and Coppola, R. (1999) Cortical activation, signal-to-noise ratio and stochastic resonance during information processing in man. Clin. Neurophysiol., 110: 1193–1203. Yamada, T., Kameyama, S., Fuchigami, Y., Nakazumi, Y., Dickins, Q.S. and Kimura, J. (1988) Changes of short latency somatosensory evoked potential in sleep. Electroencephalogr. Clin. Neurophysiol., 70: 126–136.


61

Chapter 9

Event-related components of laser evoked potentials (LEPs) in pain stimulation: recognition of infrequency, location, and intensity of pain Masutaro Kanda* Department of Brain Pathophysiology, Human Brain Research Center, Kyoto University Graduate School of Medicine, Kyoto 606-8507 and Department of Neurology, Takeda General Hospital, Kyoto 601-1495 (Japan)

1. Introduction Painful stimulation with short-pulse CO2 laser was introduced by Mor and Carmon (1975) which, unlike electric shock, activates nociceptive receptors selectively and generates pure pain sensation (Bromm et al., 1984). Information about the activation is conducted through both small myelinated Aδ- and unmyelinated C-fibers and reaches the cerebral cortex via the spinothalamic tract (Bromm and Treede, 1987). Somatosensory evoked potentials (SEPs) following CO2 laser stimulation, known as laser evoked potentials (LEPs), are mainly composed of 3 components (N1, N2, and P2) (Miyazaki et al., 1994; Xu et al., 1995). The peak latencies of these components were reported as about 150 ms for N1, 220 ms for N2, and 330 ms for P2 (Miyazaki et al., 1994). LEPs are influenced by attention modulations and cognitive tasks. It was reported that N1, N2, and P2 of

*Correspondence to: Masutaro Kanda, M.D., Department of Neurology, Takeda General Hospital, 28-1 Moriminamimachi, Fushimi-ku, Kyoto 601-1495, Japan. Tel: +81-75-572-6331; Fax: +81-75-571-8877; E-mail: [email protected]

LEPs were modulated by attention, the components of which are arousal, vigilance, alertness or sustained attention, selective or focused attention, and executive attention (Beydoun et al., 1993; Miyazaki et al., 1994; García-Larrea et al., 1997; Legrain et al., 2002; Lorenz and García-Larrea, 2003). It was also reported that event-related potentials (ERPs) were evoked following P2 by using cognitive tasks when recording LEPs. In this chapter, I have focused on the previous studies in which ERPs of LEPs were evoked using an oddball task (Kanda et al., 1996), a point localization task (Kanda et al., 1999), and a pain intensity assessment (PIA) task (Kanda et al., 2002). 2. Oddball paradigm In an oddball paradigm, a distinct “target” stimulus is presented infrequently and at random intervals within a series of frequent “regular” stimuli (Sutton et al., 1965). Brain activities are recorded as P300 or P3 in response to target stimulus, which is the characteristic endogenous positive wave (Donchin et al., 1978, 1986). Towell and Boyd (1993) demonstrated a positive potential following P2 in response to the target CO2 laser stimulus in an oddball paradigm. However, the field

62 distribution of that potential, effects of different response tasks, and differences from other stimulus modalities remained unclear. Therefore, the aim of this study was to clarify the characteristics of the endogenous, cognitive component of LEPs by employing an oddball paradigm. In 12 healthy subjects, CO2 laser stimuli were frequently delivered to the ulnar side of the left-hand dorsum while rarely to the radial side of the same hand. The stimulated sides for frequent and rare were exchanged with each other from session to session. The subjects were instructed to either count mentally or press a button in response to the “target” rare stimuli delivered at a probability (p) of 0.2. Likewise, electric stimuli for the target were applied to either the second or fourth digit of the left hand at p = 0.2, while those for frequent were applied to the other of these two digits. For auditory stimulation, the 2000 Hz tone was presented binaurally at p = 0.2 as target stimulus, while the 1000 Hz tone was frequently presented. Electroencephalograms (EEGs) were recorded from electrodes placed on the scalp; Fpz, F3, Fz, F4, T3, C3, Cz, C4, T4, P3, Pz, P4, O1, Oz, and O2 according to the International 10–20 system. For each session, responses to frequent stimuli and target stimuli were separately averaged and time-locked to the stimulus onset. In the oddball paradigm using laser stimulation, N2 and P2 of LEPs were recorded maximally at Cz regardless of oddball conditions (frequent or rare) and response tasks (mental count or button press). Neither of the two components showed differences in the latency, amplitude, or scalp topography between the oddball conditions or between response tasks. Only in response to the target stimuli, another positive component (593 ± 31 ms, 10.6 ± 3.8 μV for the count and 560 ± 54 ms, 10.5 ± 1.8 μV for the button press responses) was recorded maximally at Pz following P2, called as “laser P3.” “Electric P3” was recorded in response to target electric stimuli (425 ± 69 ms, 11.2 ± 3.7 μV for the count and 417 ± 58 ms, 12.2 ± 5.5 μV for the button press responses) and “auditory P3” in response to the target auditory stimuli (351 ± 31 ms, 11.7 ± 4.8 μV for the count, and 336 ± 38 ms, 13.4 ± 5.9 μV for the button press responses), both of which were maximal at Pz. There was no statistical difference in amplitude or scalp topography among laser, electric, and auditory P3.

The longer latency of laser P3 compared with electric or auditory P3 can be mainly explained by its slower impulse conduction of Aδ-fiber through which peripheral activation by laser stimulus was conducted. Since laser P3 was recorded only after the target presentation regardless of the type of response task and showed the same scalp topography as that of electric or auditory P3, it is suggested that laser P3 is mainly related to the categorization process and shares the same mechanism of auditory or electric P3 reported to be generated in multiple brain areas (Nishitani et al., 1998). As neither N2 nor P2 differed in latency, amplitude or scalp topography between oddball conditions or response tasks, it is postulated that these two are mainly pain-related components of LEPs. 3. Localization of the pain spot A great advantage of using CO2 laser stimulus is that the stimulus spot can be chosen irrespective of the peripheral nerve pathways. It is therefore expected that when recording LEPs, the subject would, at least to some extent, perform a discriminative task to identify the location of the pain spot for each CO2 laser stimulus; if so, this corresponds to “point localization” that forms a cortically dependent sensory function (Corkin et al., 1970; Bassetti et al., 1993; Kim and Choi-Kwon, 1996). If this is the case, LEPs might contain ERPs that reflect the process of point localization of the pain spot. Therefore, the aim of this study was to demonstrate ERPs during the point localization of pain caused by CO2 laser stimulation. Painful CO2 laser stimuli were delivered to the dorsum of either hand in 16 healthy subjects. While the stimulus spot (pain spot) was shifted for each stimulus, the subject was requested to identify the stimulated spot as accurately as possible and to use a pointer in the non-stimulated hand to indicate the corresponding spot on a picture of a hand projected onto a screen (localization condition). For the control condition, the subject was instructed to point to a single predetermined spot, regardless of the location of the stimulated spot (control motor task condition). In the control rest condition, neither point localization nor the motor task was requested. EEGs were recorded from 21 electrodes

63 placed on the scalp: Fp1, Fpz, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz, and O2 according to the International 10–20 system. They were referenced to the linked earlobes, and were averaged time-locked to the stimulus onset for each task separately. Under the control rest condition, N2 and P2 of LEPs were recorded. During the control motor task condition, a steep negative slope (NS′) was recorded at the frontocentral region following P2, indicating movementrelated cortical potentials (MRCPs) (Neshige et al., 1988). Exclusively during the localization condition, a positive peak (647 ± 89 ms, 5.6 ± 2.9 μV for the left and 634 ± 58 ms, 5.7 ± 2.7 μV for the right-hand stimulation) was identified and called the “localizationrelated potential (LP)”, which was maximal at the midline centro-parietal area and symmetrically distributed over the scalp. Neither the latency nor amplitude was significantly different between the stimulated hands. It is suggested that the LP is related to the somatotopic point localization of the pain spot. Its scalp distribution, showing midline centro-parietal distribution, suggests that it is most likely related to activation of the superior parietal cortices bilaterally, since it was reported that a superior-posterior stroke caused cortical sensory syndrome consisting of an isolated loss of discriminative sensation (stereognosis, graphesthesia, and joint position sense) involving one or two parts of the body (Bassetti et al., 1993). Another possibility is that LP as well as oddball P3 might be generated from multiple brain areas including those located in the bilateral temporal lobes and the second somatosensory cortex, because it was reported that auditory P3, which shows midline centro-parietal distribution, was generated by the contribution of multiple structures including the mesial temporal, superior temporal, and inferior parietal regions on both hemispheres (Nishitani et al., 1998). 4. Pain intensity assessment (PIA) The visual analogue scale (VAS) has been commonly used as a self-rating score of subjective pain intensity in clinical as well as experimental settings (Price et al., 1983). There have been no reports referring to a component of LEPs related to the measurement of pain

intensity, although by using strong electric shocks as painful stimuli, Becker et al. (2000) found late brain potentials relating to the measurement of pain intensity. In this study, therefore, LEPs were recorded using CO2 laser stimuli of various intensities, while the subject was required to measure on VAS the subjective intensity of pain evoked by the stimuli. In 12 healthy subjects, three kinds of CO2 laser stimuli were delivered to the left-hand dorsum at irregular intervals of 4–6 s, while the irradiation duration of each stimulus was randomly set to 40, 60, or 80 ms. For the VAS, a 50-cm long horizontal line was drawn on a screen placed 1.5 m in front of the subject. The line was labeled “no pain” at the left end and “the most intense pain imaginable” at the right. The subject was requested to assess the intensity of each pain stimulus and to move a pointer in their right hand to point to the VAS scale according to the subjective feeling of pain sensation (PIA condition). For the control condition, the subject was asked to move the pointer to the midpoint of the VAS line irrespective of the pain intensity (control motor task condition). In the control rest condition, neither PIA nor a motor response was required. EEGs were recorded from 21 electrodes placed on the scalp: Fp1, Fpz, Fp2, F7, F3, Fz, F4, F8, T3, C3, Cz, C4, T4, T5, P3, Pz, P4, T6, O1, Oz, and O2 according to the International 10–20 system. They were referenced to the linked earlobes, and were averaged time-locked to the stimulus onset for each type of stimulus as well as for each task condition. The VAS scores were 2.8 ± 0.5/10 for a stimulus of 40 ms duration, 4.8 ± 0.8/10 for 60 ms, and 6.1 ± 0.9/10 for 80 ms, and showed a highly significant positive correlation with the stimulus duration. Following the N2 and P2 of LEPs which were affected by stimulus duration but not modulated by task conditions, a positive peak was identified exclusively under the PIA condition regardless of the stimulus intensity and was called the “intensity assessment-related potential (IAP)”. The peak latencies of IAP were 642 ± 64 ms for a stimulus duration of 40 ms, 612 ± 92 ms for 60 ms, and 619 ± 76 ms for 80 ms, and their amplitudes were 8.2 ± 4.2, 7.1 ± 5.7, and 9.4 ± 5.6 μV, respectively. The IAP was maximal at the midline parietal area and symmetrically distributed over the scalp. Neither the

64 latency nor amplitude of the IAP was significantly different among three different stimulus intensities. Regarding N2 and P2, the latency of N2 was not significantly different among three conditions or among three stimulus durations, while that of P2 was significantly correlated with the stimulus duration. The amplitudes of both components were significantly correlated with the stimulus duration. It is the existence of an actual stimulus, regardless of its intensity, that operates the psychophysical processes involved in the VAS for the sensory intensity dimension of pain. As IAP did not differ in the latency, amplitude, or scalp distribution among the three different levels of stimulus intensity, it was indicated that the single major factor in the generation of IAP is the existence of pain stimulus but not its given or perceived intensity. These features of IAP fulfill most of the requirements for an endogenous component of ERPs (Donchin et al., 1978, 1986). From its scalp distribution, it can be assumed that the assessment of pain intensity involves multiple areas in both hemispheres. 5. Discussion Characteristic features of laser P3, LP, and IAP, and are shown in Table 1. For the tasks, laser P3 is related to the categorization process, LP is the precise localization,

and IAP is the intensity assessment of pain stimulus. The probability of the stimulus was one of the major factors for laser P3, but not for LP or IAP. These differences suggest that the three ERPs can be distinguished from each other. However, the possibility that LP and IAP are P3-like potentials cannot be rejected, because the waveform, latency, and scalp distribution were similar among the 3 ERPs (Table 1). Previous reports on LEPs using cognitive tasks are listed in chronological order (Table 2). Towell and Boyd (1993) reported laser P3 for the first time, and six authors have employed an oddball paradigm. Zaslansky et al. (1996) reported that P2 is involved in an oddball component, although other studies using an oddball paradigm showed that laser P3 corresponded to the oddball component (Towell and Boyd, 1993; Kanda et al., 1996; Legrain et al., 2002, 2003a, b). While we introduced the localization task that demonstrated the LP, two other authors used similar tasks. Valeriani et al. (2000) reported an early potential (eP) preceding N1 unmasked by a localization task, and Bentley et al. (2004) identified N1 enhancement with a localization task. Although we introduced a pain intensity assessment task that demonstrated the IAP, no other report using a similar task has been reported. In an oddball paradigm, Legrain et al. (2002) studied the LEPs of attended and unattended hands for

TABLE 1 TABLE 2 CHARACTERISTIC FEATURES OF LASER P3, LP, AND IAP Laser P3 Task Categorization Precise localization Intensity assessment Probability (%) Waveform Latency (ms) Amplitude (μV) Scalp distribution Maximum Symmetry Extension

LP

PREVIOUS REPORTS ON LEPs USING COGNITIVE TASKS

IAP Authors

Task Oddball paradigm Oddball paradigm Oddball paradigm Localization task Localization task Pain intensity assessment task Oddball paradigm for attended hand and unattended hand 3-stimulus oddball paradigm 3-stimulus oddball paradigm Localization task

Yes No No 20 Positive 580 10.6

No Yes No 100 Positive 630 5.7

No No Yes 100 Positive 625 8.2

Towell and Boyd (1993) Kanda et al. (1996) Zaslansky et al. (1996) Kanda et al. (1999) Valeriani et al. (2000) Kanda et al. (2002) Legrain et al. (2002)

Pz Yes Parietal

Cz or Pz Yes Parietal

Pz Yes Parietal

Legrain et al. (2003a) Legrain et al. (2003b) Bentley et al. (2004)

65 frequent and “target” rare stimuli, and compared each LEP component obtained in these 4 conditions. Their results were as follows: N1 and N2 were modulated by the direction of spatial attention, P2 was affected by the probability of the stimulus regardless of the spatial attention, and an additional parietal P600 was induced by attended rare stimuli, which could be seen as P3b. To study P3a or novelty P3, Legrain et al. (2003a, b) recorded LEPs in a three-stimulus oddball paradigm, in which frequent and “target” rare laser stimuli were given to one hand while “distractor” rare laser stimuli were given to the other hand. It was shown that targets and distractors elicited a late positive complex (LPC) around 465–500 ms and concluded that distractor LPC corresponds to P3a indexing an involuntary orientation of attention toward an unexpected new/deviant event. Taking these reports into account, it is suggested that

laser P3, LP, and IAP, which had peak latencies at about 600 ms, may share the same property with P3b but not P3a. In summary, components of LEPs and their modulation of attention and cognition are schematically shown in Fig. 1. Valeriani et al. (2000) reported an eP unmasked by using a localization task. N1 and N2 are influenced by various attention modulations (Beydoun et al., 1993; Miyazaki et al., 1994; García-Larrea et al., 1997; Legrain et al., 2002; Lorenz and García-Larrea, 2003). P2 is affected by the stimulus probability (Legrain et al., 2002). Distractor LPC corresponds to P3a (Legrain et al., 2003a, b). Laser P3, LP, and IAP are cognitive components generated in multiple brain areas, which share the same time range with P3b activities. This is also consistent with the notion that most neuronal processes specific to pain in the

Fig. 1. Schematic representation of LEP components. The solid line shows the LEP influenced by attention modulations and cognitive tasks while the interrupted line shows the LEP without these effects. A localization task unmasks a positive early potential (eP). Both N1 and N2 increased in amplitude by paying attention to the laser stimuli. The amplitude of P2 is larger after rare stimuli compared with that after frequent stimuli, independent of paying attention to the stimuli. Both target and distractor stimuli elicited a late positive complex (LPC) around 465–500 ms. Distractor LPC corresponds to P3a that indexes involuntary orientation of attention toward an unexpected new/deviant event. Laser P3 is recorded in the oddball paradigm, localizationrelated potential (LP) in the point localization task condition, and pain intensity assessment potential (IAP) in the task to assess pain intensity by means of the visual analogue scale (VAS), which appears at about 600 ms, corresponding to the time range with P3b activities.

66 cortices are likely completed before the generation of these ERPs. References Bassetti, C., Bogousslavsky, J. and Regli, F. (1993) Sensory syndromes in parietal stroke. Neurology, 43(10): 1942–1949. Becker, D.E., Haley, D.W., Urena, V.M. and Yingling, C.D. (2000) Pain measurement with evoked potentials: combination of subjective ratings, randomized intensities, and long interstimulus intervals produces a P300-like confound. Pain, 84(1): 37–47. Bentley, D.E., Watson, A., Treede, R.D., Barrett, G., Youell, P.D., Kulkarni, B. and Jones, A.K. (2004) Differential effects on the laser evoked potential of selectively attending to pain localisation versus pain unpleasantness. Clin. Neurophysiol., 115(8): 1846–1856. Bromm, B. and Treede, R.D. (1987) Human cerebral potentials evoked by CO2 laser stimuli causing pain. Exp. Brain Res., 67(1): 153–162. Bromm, B., Jahnke, M.T. and Treede, R.-D. (1984) Responses of human cutaneous afferents to CO2 laser stimuli causing pain. Exp. Brain Res., 55: 158–166. Beydoun, A., Morrow, T.J., Shen, J.F. and Casey, K.L. (1993) Variability of laser-evoked potentials: attention, arousal and lateralized differences. Electroencephalogr. Clin. Neurophysiol., 88(3): 173–181. Corkin, S., Milner, B. and Rasmussen, T. (1970) Somatosensory thresholds–contrasting effects of postcentral-gyrus and posterior parietal-lobe excisions. Arch. Neurol., 23(1): 41–58. Donchin, E., Ritter, W. and McCallum, W.C. (1978) Cognitive psychophysiology: the endogenous component of the ERP. In: E. Callaway, P. Tueting and S.H. Koslow (Eds.), Event-Related Brain Potentials in Man. Academic Press, New York, pp. 349–411. Donchin, E., Karis, D., Bashore, T.R., Coles, M.G.H. and Gratton, G. (1986) Cognitive psychophysiology and human information processing. In: M.G.H. Coles, E. Donchin and S.W. Porges (Eds.), Psychophysiology: Systems, Processes, and Applications. Guilford Press, New York, pp. 244–267. García-Larrea, L., Peyron, R., Laurent, B. and Mauguière, F. (1997) Association and dissociation between laser-evoked potentials and pain perception. Neuroreport, 8(17): 3785–3789. Kanda, M., Fujiwara, N., Xu, X., Shindo, K., Nagamine, T., Ikeda, A. and Shibasaki, H. (1996) Pain-related and cognitive components of somatosensory evoked potentials following CO2 laser stimulation in man. Electroencephalogr. Clin. Neurophysiol., 100(2): 105–114. Kanda, M., Shindo, K., Xu, X., Fujiwara, N., Ikeda, A., Nagamine, T. and Shibasaki, H. (1999) Cortical mechanisms underlying point localization of pain spot as studied by event-related potentials following CO2 laser stimulation in man. Exp. Brain Res., 127(2): 131–140. Kanda, M., Matsuhashi, M., Sawamoto, N., Oga, T., Mima, T., Nagamine, T. and Shibasaki, H. (2002) Cortical potentials related to assessment of pain intensity with visual analogue scale (VAS). Clin. Neurophysiol., 113(7): 1013–1024.

Kim, J.S. and Choi-Kwon, S. (1996) Discriminative sensory dysfunction after unilateral stroke. Stroke, 27(4): 677–682. Legrain, V., Guérit, J.M., Bruyer, R. and Plaghki, L. (2002) Attentional modulation of the nociceptive processing into the human brain: selective spatial attention, probability of stimulus occurrence, and target detection effects on laser evoked potentials. Pain, 99(1–2): 21–39. Legrain, V., Bruyer, R., Guérit, J.M. and Plaghki, L. (2003a) Nociceptive processing in the human brain of infrequent taskrelevant and task-irrelevant noxious stimuli. A study with eventrelated potentials evoked by CO2 laser radiant heat stimuli. Pain, 103(3): 237–248. Legrain, V., Guérit, J.M., Bruyer, R. and Plaghki, L. (2003b) Electrophysiological correlates of attentional orientation in humans to strong intensity deviant nociceptive stimuli, inside and outside the focus of spatial attention. Neurosci. Lett., 339(2): 107–110. Lorenz, J. and García-Larrea, L. (2003) Contribution of attentional and cognitive factors to laser evoked brain potentials. Neurophysiol. Clin., 33(6): 293–301. Miyazaki, M., Shibasaki, H., Kanda, M., Xu, X., Shindo, K., Honda, M., Ikeda, A., Nagamine, T., Kaji, R. and Kimura, J. (1994) Generator mechanism of pain-related evoked potentials following CO2 laser stimulation of the hand: scalp topography and effect of predictive warning signal. J. Clin. Neurophysiol., 11: 242–254. Mor, J. and Carmon, A. (1975) Laser emitted radiant heat for pain research. Pain, 1: 233–237. Neshige, R., Lüders, H. and Shibasaki, H. (1988) Recording of movement-related potentials from scalp and cortex in man. Brain, 111(Pt 3): 719–736. Nishitani, N., Nagamine, T., Fujiwara, N., Yazawa, S. and Shibasaki, H. (1998) Cortical-hippocampal auditory processing identified by magnetoencephalography. J. Cogn. Neurosci., 10(2): 231–247. Price, D.D., McGrath, P.A., Rafii, A. and Buckingham, B. (1983) The validation of visual analogue scales as ratio scale measures for chronic and experimental pain. Pain, 17(1): 45–56. Sutton, S., Braren, M., Zubin, J. and John, E.R. (1965) Evokedpotential correlates of stimulus uncertainty. Science, 150(3700): 1187–1188. Towell, A.D., Boyd, S.G. (1993) Sensory and cognitive components of the CO2 laser evoked cerebral potential. Electroencephalogr. Clin. Neurophysiol., 88(3): 237–239. Valeriani, M., Restuccia, D., Le Pera, D., Fiaschetti, L., Tonali, P. and Arendt-Nielsen, L. (2000) Unmasking of an early laser evoked potential by a point localization task. Clin. Neurophysiol., 111(11): 1927–1933. Xu, X., Kanda, M., Shindo, K., Fujiwara, N., Nagamine, T., Ikeda, A., Honda, M., Tachibana, N., Barrett, G., Kaji, R., Kimura, J. and Shibasaki, H. (1995) Pain-related somatosensory evoked potentials following CO2 laser stimulation of foot in man. Electroencephalogr. Clin. Neurophysiol., 96: 12–23. Zaslansky, R., Sprecher, E., Katz, Y., Rozenberg, B., Hemli, J.A. and Yarnitsky, D. (1996) Pain-evoked potentials: what do they really measure? Electroencephalogr. Clin. Neurophysiol., 100(5): 384–391.


67

Chapter 10

Novelty-related brain response and its clinical applications Shuhei Yamaguchi* Departments of Neurology, Hematology and Immunology, Shimane University School of Medicine, Izumo, Shimane 693-8501 (Japan)

1. Introduction Involuntary attention shift to unexpected or novel events (i.e. an orienting response) is a fundamental biological mechanism for survival and evolution of human beings. The neural systems for novelty processing have been investigated with regard to two different aspects of stimulus novelty, i.e. perceptual novelty and contextual novelty (Ranganath and Rainer, 2003). Perceptual novelty has been mainly investigated via cellular mechanisms for reduced neural activity with stimulus repetition. This phenomenon is known as adaptation and is observed commonly in various areas of the brain. On the other hand, contextual novelty has been studied extensively by scalp-recorded event-related potentials (ERPs). When an unexpected event occurs in a familiar environment or context, attention is automatically oriented toward the stimulus. ERPs and recent imaging techniques have contributed to understanding the neural basis of novelty processing. Progress in exploration of neural mechanism for novelty processing has been reviewed recently (Soltani and Knight, 2000;

* Correspondence to: Shuhei Yamaguchi, M.D., Ph.D., Departments of Neurology, Hematology and Immunology, Shimane University School of Medicine, Izumo, Shimane 693-8501, Japan. Tel: +81-853-20-2196; Fax: +81-853-20-2194; E-mail: [email protected]

Yamaguchi, 2004). This chapter will present new findings obtained from functional magnetic resonance imaging (fMRI) study using a high-field MRI machine and will summarize the several clinical applications of the novelty-related ERP. 2. Neural systems for novelty processing The orienting response generates a P3 ERP (novelty P3), which has a frontocentral scalp distribution and peaks approximately 250–350 ms after novel stimulus onset. The novelty P3 is studied using a task paradigm in which novel or deviant stimuli are presented randomly within a stream of infrequent target and frequent non-target (standard) stimuli. Data from ERP recordings in patients with focal brain lesion, intracranial ERP recordings, and current source analysis of ERPs provide converging evidence for a distributed neural network for novelty processing (Knight and Scabini, 1998; Friedman et al., 2001). Recent fMRI studies have also made significant contributions to exploring more precise cortical and sub-cortical networks (Kiehl et al., 2001). In a recent event-related fMRI study, we have explored brain areas related to novelty processing, and the time course of their activation, to characterize habituation properties of brain responses to novel events (Yamaguchi et al., 2004a). We adopted a visual novelty oddball paradigm, in which subjects were required to

68 direct their attention to one visual hemifield. Visual stimuli were flashed randomly in either hemifield, and subjects were instructed to allocate their attention to one hemifield and to detect targets in that one. Novel stimuli were delivered in either the attended or unattended hemifield. In agreement with previous studies, novel stimuli generated activation in distributed brain regions, involving the middle frontal gyrus, cingulate gyrus, precuneus, superior and inferior parietal lobules, middle temporal gyrus, cuneus, lingual gyrus, fusiform gyrus, and hippocampus. When the novel stimuli were presented in the unattended hemifield, most brain regions such as the fusiform gyrus showed reduced activation, except for the prefrontal gyrus and hippocampus. They were activated by unattended novel stimuli to the same degree as attended novel stimuli (Fig. 1). Involuntary capture of attention is

the central mechanism of the orienting response. Thus, the prefrontal–hippocampal activation to both attended and unattended novel stimuli indicates that these structures are most critical for the orienting response. Furthermore, when looking at the blood oxygen level dependent (BOLD) responses in the prefrontal and hippocampal regions, the activity showed a rapid decrease within the first few occurrences of novel stimuli (Fig. 2). These responses to attended and unattended novel stimuli were comparably habituated. In contrast, the signals in other brain regions were relatively constant. These findings indicate that prefrontal and hippocampal regions are involved in rapid automatic detection and habituation to unexpected environmental events and are key elements of the orienting response in humans.

Fig. 1. MRI images (left panel) show brain activations by novel stimuli presented in the attended (left column) and unattended (right column) visual field. The right panel shows time course plots of signal changes for attended (solid line) and unattended (dashed line) novel stimuli in the left fusiform gyrus (FG), left hippocampus (Hip), and right superior/middle frontal gyrus (SFG/MFG).

69

Fig. 2. Changes in the time series of hemodynamic response curve for first five trials in response to attended (solid line) and unattended (dashed line) novel stimuli in the hippocampus and prefrontal cortex. Note the rapid reduction of hemodynamic response as a function of stimulus number in both regions.

3. Clinical applications 3.1. Aging and gender effects When novelty P3 measures are applied to neurobiological assessment of cognitive function in clinical population, it is important to assess the effect of normal aging to its amplitude and latency. There have been many reports that P3 amplitude and latency are affected by aging (Polich, 1996). We measured auditory and visual P3s to target and novel stimuli in “normal” subjects aged 35–76 years old, who had no neurological or psychiatric disorders. As seen for, the latency and amplitude of the P3 to target stimuli, and also of the novelty P3 showed clear aging effects. In addition, our study demonstrated gender differences in aging effects on P3 measures; i.e. the positive correlation between age and P3 latency was stronger for the male subjects than for female subjects (Fig. 3). The reason for this difference was investigated by comparing anatomical changes in MR images. There was a significant increase

in the number of silent ischemic lesions in aged male subjects (Fig. 4). Since the incidence of silent brain infarction in elderly increases along with aging and being male is a significant risk factor for silent brain infarction (Kobayashi et al., 1997), the difference in P3 measures may be explained partially by these silent brain lesions, which could affect cognitive functions sub-clinically. Profiles including anatomical MRI should thus be checked carefully in the selection of normal subjects before control ERP data are collected. 3.2. Novelty seeking in Parkinson’s disease The brain system for novelty processing has also been discussed in terms of dimensions of personality (Cloninger, 1987), and they are linked to variability in neurotransmitter systems in individual brains. Among several personalities, novelty seeking is reported to be associated with the dopamine system. Since the pathological deficiency of dopamine is a cardinal feature of Parkinson’s disease (PD), the reduced novelty-seeking

70

Fig. 3. The correlation between age and latency of the target and novelty P3 in male and female subjects. Male subjects show a stronger aging effect.

Fig. 4. The left panel shows the incidence of silent brain infarction as a function of age in male and female subjects. Male subjects showed higher incidence of silent brain infarction in the oldest group. The right panel shows deteriorative effects of silent brain infarction on the latency of target and novelty P3.

71 behavior is expected in some patients with PD (Tomer and Aharon-Peretz, 2004). We recorded novelty P3 ERPs using a novelty oddball task in PD patients without dementia, and compared them with age-matched controls (Tsuchiya et al., 2000). The PD patients showed prolonged P3 latency to novel stimuli, whereas their P3 latency to target stimuli was not different from that in controls. In addition, the PD patients manifested amplitude reduction over frontal scalp sites compared with controls. The prolonged latency and frontal reduction of novelty P3 correlated with a poor performance in the Wisconsin card sorting test. These results suggest that the orienting response in PD patients to novel events is impaired and the prefrontal–striatal dopamine system is involved in their reduced novelty-seeking behavior. 3.3. Apathy after stroke Apathy is often observed after stroke and is defined as reduced motivation and lack of initiative and exploration. Decreased novelty seeking is also a behavioral feature of apathy. Many clinical observations have demonstrated that the frontal cortex is a neural structure responsible for apathy state after stroke (Stuss et al., 2000). In addition, sub-cortical structures including caudate nucleus, thalamus, and basal ganglia are also reported to be related to apathy. Frontal lobe lesions cause decreased novelty-seeking behavior associated with reduced novelty P3 amplitude (Daffner et al., 2000). We recorded novelty P3 ERP in patients with sub-cortical stroke with or without apathy (Yamagata et al., 2004). Apathetic state was quantified by a selfassessment scale. There were no differences in lesion location between the apathy and non-apathy group, but the global cognitive function was significantly impaired in the apathy group compared to the nonapathy group. The ERP data showed that the novelty P3 latency was significantly prolonged, and its amplitude was reduced over the frontal site in the apathy group (Fig. 5). The apathy scale was correlated with the novelty P3 latency and amplitude at the frontal site. Thus, the study suggests that apathy after sub-cortical stroke is associated with impaired neural processing of novel events within the frontal–sub-cortical system and that the novelty P3 is a useful physiological

measure for assessing apathy after stroke. The advantage of the novelty P3 recording over target P3 is that the paradigm for recording novelty P3 does not require any behavioral output to a novel stimulus itself. This feature of automaticity is clearly suitable for assessing subjects with a lack of psychic initiative to environments. 3.4. Dementia It is well known that the target P3 is a reliable physiological index of cognitive impairments in various neurological disorders exhibiting dementia (Goodin and Aminoff, 1992). However, the changes in the novelty P3 remained to be determined in patients with dementia. The data from PD or sub-cortical stroke patients suggest contributions of the frontal lobe to the changes in the novelty P3 and its relationship with cognitive and behavioral abnormality. So it is predicted that disorders affecting frontal lobe function may show significant alterations in the novelty P3. We recorded ERPs to target and novel stimuli in a novelty auditory oddball task in patients with Alzheimer-type dementia (AD), vascular dementia (VD), and age-matched controls (Yamaguchi et al., 2000). Global cognitive impairments were the same in AD and VD. The amplitude, latency, and scalp topography of the target P3 were comparably affected by both AD and VD. However, the amplitude of the novelty P3 was markedly reduced in VD, but not in AD, and the scalp topographies were different in the three groups. The amplitude was maximal at frontal sites in controls, at central sites in AD, and at parietal sites in VD (Fig. 6). The target P3 latency was prolonged in both AD and VD, whereas the novelty P3 latency was only prolonged in VD. We also performed discriminant analysis for AD and VD only by means of novelty P3 measures. AD was discriminated satisfactorily from VD by using the novelty amplitude at Cz and the ratio of the amplitudes at Fz and Pz as independent variables. Some contradictory data were reported regarding the novelty P3 in AD (Daffner et al., 2001). They demonstrated marked reduction in the novelty P3 amplitude in mild AD patients compared to normal controls in a visual novelty oddball task. It is not clear whether the

72

Fig. 5. Topographies of the novelty P3 in patients with and without apathy after stroke (left panel). Apathetic group shows posterior shift of the peak amplitude. The degree of apathy was correlated with the novelty P3 latency and amplitude in sub-cortical stroke patients (right panel).

discrepancy is due to difference in stimulus modality or patient characteristics. In our patient population, apathy state was apparently more severe in VD compared to AD. Although further studies including AD patients with varying degrees of dementia and apathy are necessary to address these issues, our study suggests that the response to novel stimuli may be relatively resistant to the changes in AD, at least in the early stage of the disease, because there was no correlation between the intelligence score and the novelty P3 latency or amplitude. Thus, dissociated P3 responses to target and novel stimuli in demented subjects seem preferable for the diagnosis of AD rather than VD.

3.5. Drug study We have recently demonstrated the usefulness of novelty P3 recording for monitoring drug effects on some cognitive functions (Yamaguchi et al., 2004b). Cholinesterase inhibitors and estrogen replacement therapy have been proven to be effective in improving some cognitive functions in patients with AD. In contrast to AD, few effective treatments have been reported for VD. A herbal medicine termed Choto-san is a kampo (Japanese herbal) prescription, administered to older patients suffering from physical weakness and subjective symptoms such as headache, a heavy feeling of the head, vertigo, hot flashes,

73

Fig. 6. Topographies of the target and novelty P3 in the groups of controls, AD and VD patients. Note differences only in the novelty P3 topography among three groups.

tinnitus, insomnia, and painful tension of the shoulder. Since these symptoms often occur after stroke, Chotosan was also expected to be effective in cognitive impairments after stroke. One double-blind study demonstrated its effectiveness on several cognitive and behavioral impairments after stroke (Terasawa et al., 1997). We measured P3 ERPs to assess the effect of Choto-san administration on stroke patients with mild cognitive impairments. Choto-san was given for 12 weeks to ten chronic stroke patients. P3 ERPs were recorded in a novelty auditory oddball paradigm before and after drug administration. Twelve-week administration of Choto-san significantly improved mini mental state examination and verbal fluency test scores. P3 latency to target sounds was shortened in association with reduced reaction time to the sounds after drug administration. Furthermore, P3 amplitude to novel sounds was enlarged and its topography shifted from central to frontal sites after 12-week Choto-san administration. These results indicate that Choto-san improves electrophysiological indices related to attention and decision-making, associated

with improvement of neuropsychological test scores in stroke patients with mild cognitive impairments. This study also suggests that the novelty oddball paradigm provides useful tools for assessing electrophysiological modulation by drug administration. 3.6. Danger detection whilst walking Novelty detection is critical both in avoiding danger and in adapting to environmental change. In daily life, danger detection plays an important role in avoiding collisions or falls whilst walking. Few studies have been conducted to assess objective measures of danger detection. We developed a paradigm for recording the novelty P3 as an index for the ability of danger detection during an imaginary walk. First, in order to make a stimulus movie, we took a scenery film while walking in various situations such as a sidewalk, narrow alley, or mountain path. Then we edited the film so that a variety of obstructing objects occasionally intruded into the scene with real sounds. They consisted of a car, bicycle, ball, animal, puddle, and

74

Fig. 7. Topographies of the P3 potential to target and obstructive stimuli. The P3 to obstructing stimuli distributes maximally over the frontal scalp site.

rock. We also inserted an animal cartoon as a target stimulus in the movie. The subjects sat in front of a cathode ray tube (CRT) and watched the movie. They were asked to detect all target stimuli in the movie, but ignore other occasionally intruding stimuli. ERPs were measured to target and obstructing stimuli separately. All subjects detected all targets with a high accuracy of more than 95%. Positive ERP components were elicited to both target and obstructing stimuli in the latency range of 270–360 ms after the stimulus onset. The target P3 amplitude was maximal at Pz, whereas the P3 amplitude to obstructing objects was maximal at Fz with the mean latency of 320 ms (Fig. 7). The P3 component to bimodal obstructing stimuli shares common characteristics with a novelty P3 in a conventional novelty oddball task. This paradigm would be useful in assessing the ability of danger detection whilst walking in various neurological disorders as well as in aged people. References Cloninger, C.R. (1987) A systematic method for clinical description and classification of personality variables. Arch. Gen. Psychiatry, 44: 573–588. Daffner, K.R., Mesulam, M.M., Scinto, L.F., Acar, D., Calvo, V., Faust, R., Chabrerie, A., Kennedy, B. and Holcomb, P. (2000) The central role of the prefrontal cortex in directing attention to novel events. Brain, 123: 927–939.

Daffner, K.R., Rentz, D.M., Scinto, L.F., Faust, R., Budson, A.E. and Holcomb, P.J. (2001) Pathophysiology underlying diminished attention to novel events in patients with early AD. Neurology, 56: 1377–1383. Friedman, D., Cycowicz, Y.M. and Gaeta, H. (2001) The novelty P3: an event-related brain potential (ERP) sign of the brain’s evaluation of novelty. Neurosci. Biobehav. Rev., 25: 355–373. Goodin, D.S. and Aminoff, M.J. (1992) Evaluation of dementia by event-related potentials. J. Clin. Neurophysiol., 9: 521–525. Kiehl, K.A., Laurens, K.R., Duty, T.L., Forster, B.B. and Liddle, P.F. (2001) Neural sources involved in auditory target detection and novelty processing: an event-related fMRI study. Psychophysiology, 38: 133–142. Knight, R. and Scabini, D. (1998) Anatomic bases of event-related potentials and their relationship to novelty detection in humans. J. Clin. Neurophysiol., 15: 3–13. Kobayashi, S., Okada, K., Koide, H., Bokura, H. and Yamaguchi, S. (1997) Subcortical silent brain infarction as a risk factor for clinical stroke. Stroke, 28: 1932–1939. Polich, J. (1996) Meta-analysis of P300 normative aging studies. Psychophysiology, 33: 334–353. Ranganath, C. and Rainer, G. (2003) Neural mechanisms for detecting and remembering novel events. Nature Rev. Neurosci., 4: 193–202. Soltani, M. and Knight, R.T. (2000) Neural origins of the P300. Crit. Rev. Neurobiol., 14: 199–224. Stuss, D.T., Reekum, R.V. and Murphy, K.J. (2000) Differentiation of states and causes of apathy. In: J.C. Borod (Ed.), The Neuropsychology of Emotion. Oxford University Press, New York, pp. 340–363. Terasawa, K., Shimada, Y., Kita, T., Yamamoto, T., Tosa, H., Tanaka, N., Saito, Y., Kanaki, E., Goto, S., Mizushima, N., Fujioka, M., Takase, S., Seki, H., Kimura, I., Ogawa, T., Nakamura, S., Araki, G., Marayama, I., Maruyama, Y. and Takaori, S. (1997) Choto-san in the treatment of vascular dementia: a double-blind, placebo-controlled study. Phytomedicine, 4: 15–22. Tomer, R. and Aharon-Peretz, J. (2004) Novelty seeking and harm avoidance in Parkinson’s disease: effects of asymmetric dopamine deficiency. J. Neurol. Neurosurg. Psychiatr., 75: 972–975. Tsuchiya, H., Yamaguchi, S. and Kobayashi, S. (2000) Impaired novelty detection and frontal lobe dysfunction in Parkinson’s disease. Neuropsychologia, 38: 645–654. Yamagata, S., Yamaguchi, S. and Kobayashi, S. (2004) Impaired novelty processing in apathy after subcortical stroke. Stroke, 35: 1935–1940. Yamaguchi, S. (2004) Neural network for novelty processing. In: M. Hallett, L.H. Phillips II, D.L. Schomer and J.M. Massey (Eds.), Advances in Clinical Neurophysiology. Elsevier, Amsterdam, pp. 631–637. Yamaguchi, S., Tsuchiya, H., Yamagata, S., Toyoda, G. and Kobayashi, S. (2000) Event-related brain potentials in response to novel sounds in dementia. Clin. Neurophysiol., 111: 195–203. Yamaguchi, S., Hale, L., D’Esposito, M. and Knight, R.T. (2004a) Rapid prefrontal-hippocampal habituation to novel events. J. Neurosci., 24: 5356–5363. Yamaguchi, S., Matsubara, M. and Kobayashi, S. (2004b) Event-related brain potential changes after Choto-san administration in stroke patients with mild cognitive impairments. Psychopharmacology, 171: 241–249.


75

Chapter 11

Cross-modal plasticity in the blind revealed by functional neuroimaging Norihiro Sadato* Department of Cerebral Research, National Institute for Physiological Sciences/The Graduate University for Advanced Studies, Okazaki 444-8585 (Japan)

1. Introduction

2. Historical background

In recent years, non-invasive cerebral functionalimaging techniques, such as positron emission tomography (PET) and functional magnetic resonance imaging (fMRI), have becoming indispensable in our quest to understand higher brain functions. Local neural activity, especially synaptic activity, increases in parallel with the glucose metabolism in a particular region of the brain. In turn, the regional cerebral blood flow (rCBF) parallels the glucose metabolism, which is mediated by the oxygen supply to the region (Raichle, 1987). Thus, changes in local neural activity can be inferred by measuring changes in rCBF. These methods measure the cerebral blood flow while a subject executes a particular task, which is compared to the blood flow while the subject is in a resting state. The distribution of differences in activity between the active and resting states is then visualized, indicating the regions of the brain that are involved in a particular task.

2.1. Neural activity and cerebral blood flow

*Correspondence to: Norihiro Sadato, M.D., Ph.D., Department of Cerebral Research, National Institute for Physiological Sciences, 38 Nishigonaka, Myodaiji, Okazaki 444-8585, Aichi, Japan. Tel: +81-564-55-7841; Fax: +81-564-55-7786; E-mail: [email protected]

The Italian physiologist Mosso (1881) made the first reference to the relationship between cerebral blood flow and neural activity. Mosso measured the pulsation of the cerebral cortex in a patient whose cranial bone was partially missing after neurosurgery. Because this pulsation showed local increases that occurred simultaneously with mental activity, he concluded that the regional cerebral circulation (rCBF) changed following psycho-neuronal activity. Roy and Sherrington (1890) used animal studies to deduce that the increased metabolism associated with local activity in the brain causes an increase in blood flow to the location. Fulton (1928) found that a patient with arteriovenous malformation in the occipital lobe complained that he heard a murmur inside his head. This murmur, which was caused by the difference in arteriovenous blood pressure, was proportional to the blood flow. Fulton noted that this sound was stronger when the patient was reading than when he was just looking, and concluded that the rCBF correlated with the intensity of mental activity. Thus, it has been known for some time that cerebral activity can be measured by the changes in the rCBF. However, limitations in measurement techniques meant that this knowledge could only be put into practice at a much

76 later date. Kety (1951) developed a method of quantifying the rCBF in experimental animals. Rapid progress in medical imaging methods since the 1970s has enabled researchers to develop a non-invasive way to measure cerebral blood flow in humans.

Medical imaging techniques use electromagnetic waves to visualize the human body. Information about the inside of the body is obtained using electromagnetic waves with a wavelength that is longer (radio waves) or shorter (X-rays or gamma rays emitted by an isotopic tracer) than that of visible light. Such information includes morphological and functional data: the former is found primarily through X-ray imaging and the latter by means of nuclear medicine. Measurements of cerebral blood flow in humans were first made possible by nuclear medicine techniques. These techniques are based on labeling a substance with a radioisotope. The substance is injected into the body and accumulates in regions of the brain in proportion to the regional blood flow; its levels are then measured from outside the body. Measurements of cerebral blood flow in humans were first conducted in the 1960s using 85Kr gas (Lassen et al., 1963). In 1973, Hounsfield invented X-ray computerized tomography (CT) (Hounsfield, 1973). The tomographical reconstruction techniques of CT were used to construct positron emission tomography (PET) (Ter-Pogossian et al., 1975). PET is a tomographical technique based on the measurement of gamma rays (emitted when a positron is annihilated) and the calculation of the distribution of the positronemitting tracer in the body. Using the appropriate tracer, various physiological and biochemical measurements can be conducted in addition to the measurement of cerebral blood flow.

resonance (NMR) of the hydrogen atom. The phenomenon of NMR was discovered independently by Bloch (1946) and Purcell et al. (1946), and was developed principally in the field of chemistry. A report in the early 1970s revealed that this technique was useful in differentiating between malignant and benign tumors, which is of great importance to medical diagnosis (Damadian, 1971). This yielded a prime opportunity to create medical imaging systems based on the NMR phenomenon; subsequently, Lauterbur (1973) invented the MRI. When a hydrogen atom is placed in a uniform static magnetic field, it absorbs (resonance) and emits (relaxation) a radio wave with a specific frequency (the phenomenon of NMR). By placing a coil in parallel to the static magnetic field, this phenomenon can be detected as a gradually decaying alternating current, which is the MR signal. The positional information embedded in this MR signal is captured based on the principles of CT. The image obtained primarily reflects differences in the distribution density and the speed of relaxation of the hydrogen atoms, which in turn reflects the different composition of tissues in the body. By changing the data-acquisition parameters, images that emphasize various contrasts between tissues can be obtained. Compared with X-ray imaging, MRI has several advantages. First, because the radio waves have less energy than X-rays (approximately 10−12), the probability of MRI causing tissue damage is lower. In addition, while X-rays are best suited to detecting heavy atoms, which are present only in small quantities in the body (for example, calcium contained in the bones), MRI is suitable for detecting hydrogen atoms, which are abundant throughout the body (primarily as water). For this reason, MRI is particularly useful for imaging neural tissue, which is tightly protected by the cranial bone and the spine.

2.3. MRI

2.4. Detection of rCBF by MRI: fMRI

Compared with the medical application of short electromagnetic waves, MRI is a relatively recent development. Information about the inside of the body is visualized using radio waves with long wavelengths. MRI is an imaging technique that utilizes the nuclear magnetic

Due to its high-contrast resolution, MRI was initially used clinically to image the anatomical details of the brain. At the beginning of the 1990s, however, the visualization of changes in rCBF was made possible through the use of blood oxygen as an endogenous

2.2. Medical imaging

77 contrast medium, which paved the way for fMRI (Ogawa and Lee, 1990). fMRI is known as a “blood oxygen level-dependent” (BOLD) method, because it principally enhances small signals produced by local changes in the balance of intravascular blood oxygenation that appear during enhanced neural activity. It has long been known that oxyhemoglobin and deoxyhemoglobin have different magnetic properties (Pauling and Coryell, 1936), and that the presence of deoxyhemoglobin in the blood vessels produces a patchiness of the local perivascular magnetic field. The uneven local magnetic field causes the NMR signal to decrease compared to the signal in a homogeneous magnetic field. During enhanced neural activity, there is an increase in cerebral blood flow, supplying oxygen above and beyond the local demand of the neural tissue, which in turn lowers the local deoxyhemoglobin level. As a result, the NMR signal is augmented (Ogawa and Lee, 1990). The advantage of this method is that changes in the cerebral blood flow to the entire brain can be recorded at intervals of a few seconds, thereby providing much more data than PET. Presently, changes in the rCBF can be measured every second, with a spatial resolution of a few millimeters. 3. The study of brain plasticity following loss of vision 3.1. Braille reading Braille is a tactile letter system, which consists of a series of raised dots that can be read with the fingers, and is a well-known sensory substitution for the blind. Braille symbols are formed within units of space known as Braille cells. A full Braille cell consists of six raised dots arranged in two parallel columns, each having three dots. Sixty-three combinations are possible using one or more of these six dots. A single cell can be used to represent a letter of the alphabet, number, punctuation mark, or even a whole word. Braille is not a language; rather, it is a code by which languages can be written and read. The Braille system has its roots in the military field. In the early nineteenth century, it was invented as the tactile “night writing” code for sending military messages that could be read on the battlefield

without light. The system used 12 raised dots to represent sounds, and was too complicated to be of practical use. In 1821, Louis Braille, who was blind from the age of 4 years, realized how useful this system of raised dots could be and simplified it from the original 12 to 6 dots. Braille has now become a worldwide standard. It is not only an effective means of communication, but is also a proven avenue for achieving and enhancing literacy for the blind. 3.2. Activation studies with Braille tasks Braille reading requires the conversion of simple tactile information into meaningful patterns that have lexical and semantic properties (Sadato et al., 1998). The perceptual processing of Braille might be mediated by the somatosensory system, whereas visual letter identity is routinely accomplished within the visual system. Adult subjects with early blindness of peripheral origin showed a higher glucose metabolic rate in striate and prestriate areas than sighted subjects at rest as well as during tactile or auditory stimulation (Wanet-Defalque et al., 1988). This finding raised the possibility that the visual cortices of the blind might participate in the processing of non-visual information, although they failed to reveal specific task-related activation. Tactile imagery or Braille reading in blind subjects caused task-related activation in occipital electroencephalography (EEG) leads (Uhl et al., 1991), suggesting that somatosensory input is redirected to the occipital area. PET with O-15 water revealed that the primary visual cortex is activated when congenital and early-onset (165 μV were automatically rejected. Groups of 20 responses were averaged, and 2–4 repetitions subsequently averaged together for each subject, after ensuring they were free from major artefacts. In the overall mean responses of each subject, the peak latencies of the N1 and P2 defections (the largest negative and positive peaks within the first 300 ms) were measured from stimulus onset, and their peak amplitudes measured from the baseline, less than 20 ms after the stimulus. The responses of all subjects in each group were averaged together for illustration. 2.3. Statistical analysis In the preliminary study, the amplitude and latency of responses recorded at Fz to the onset of random noise and the transition from random noise to IRN were compared using paired t-tests. In the second experiment, amplitude and latency values were compared between

92 TABLE 1 MEANS AND STANDARD DEVIATIONS OF N1 AND P2 AMPLITUDE (μV) AND LATENCY (ms) AT Fz Wide-band

Low-pass

High-pass

1. Noise onset N1 (μV) (ms) P2 (μV) (ms)

6.7 ± 5.1 111.0 ± 19.0 9.2 ± 4.8 194.0 ± 25.0

5.6 ± 2.7 105.0 ± 14.0 9.8 ± 3.6 194.0 ± 27.0

7.0 ± 3.4 105.0 ± 9.0 6.4 ± 2.8 194.0 ± 23.0

2. Transition to IRN N1 (μV) (ms) P2 (μV) (ms)

8.4 ± 4.1 109.0 ± 9.0 6.6 ± 1.3 219.0 ± 21.0

7.4 ± 3.3 108.0 ± 7.0 6.7 ± 2.8 220.0 ± 19.0

6.8 ± 2.1 116.0 ± 6.0 5.4 ± 1.1 224.0 ± 22.0

6 stimulus conditions and 6 recording sites using 2-way repeated measures ANOVA. Additionally, in order to compare the antero-posterior scalp distribution of amplitudes in each condition, amplitude measures at the 4 midline sites (Fpz, Fz, Cz, and Pz) were expressed as a proportion of their sum, and the condition × electrode interaction term in ANOVA was examined. When the main effect or interaction term was significant, paired t-tests were performed as appropriate. Probabilities of less than 0.05 were regarded as significant. 3. Results In the preliminary study, consistent responses to the onset of random noise and to the transition from random noise to IRN were recorded in every subject. At Fz, the N1 measured on average 9.1 μV to noise onset and 10.6 μV to IRN. For each stimulus the P2 measured approximately 8 μV. Mean peak latencies were around 105 ms for the N1 and 180–210 ms for the P2. In paired t-tests there were no significant differences in amplitude between the two stimulus conditions, but the P2 was significantly shorter in latency in the responses to noise onset (p < 0.01). In the second experiment, consistent responses were recorded to noise onset and to the transition from noise to IRN in all 3 filter conditions, wide-band,

low-pass, and high-pass (Table 1). Comparing the 6 conditions (2 stimulus × 3 filter conditions), there was a significant main effect on N1 latency (F = 3.692, p = 0.004), P2 amplitude (F = 5.191, p = 0.0002), and P2 latency (F = 9.363, p = 0.0000001) but not on N1 amplitude (F = 0.550, p = 0.738). P2 amplitude tended to be lower at the transition to IRN as compared with the onset of random noise, and to the onset of high-pass as compared with wide-band filtered noise, but none of the 6 conditions differed significantly in paired t-tests. P2 latency was significantly longer in the low-pass and high-pass IRN conditions as compared with the corresponding noise onsets (p = 0.05), but no significant inter-condition differences were found for N1 latency. Inspection of the waveforms (Fig. 1) suggested a tendency for the responses to the transition from noise to IRN to be more anteriorly distributed. In order to compare their antero-posterior distribution between stimulus conditions, in each subject the amplitudes at the 4 midline electrodes (Fpz, Fz, Cz, and Pz) were expressed as a proportion of their sum, excluding the lateral electrodes C4 and C3, and a repeated measures ANOVA performed. There was then a significant condition × electrode interaction for the N1 (F = 2.828, p = 0.0008). It is apparent from Fig. 1 that this was due to a tendency for relatively larger N1 amplitudes to be recorded at Fpz and Fz in the IRN conditions.

Fig. 1. Group mean waveforms of 6 subjects to the onset of random noise and the transition to iterated rippled noise; wideband, low-pass, and high-pass filtered sounds. Note the tendency for the N1 peaking at about 100 ms to be relatively larger at Fpz and Fz in the “noise to IRN” conditions.

94 The pattern for P2 amplitude could not be reliably determined, since the peak was occasionally negative with respect to the baseline at Fpz, possibly on account of residual eye-movement artefacts. 4. Discussion The AEP findings are in broad agreement with the MEG data of Krumbholz et al. (2003), confirming that the transition from noise to IRN evokes a consistent N1–P2 complex with a scalp distribution significantly anterior to that of the corresponding potentials following the onset of random noise. Additionally, it has been demonstrated that the IRN response is largely independent of the frequency band of the noise. Specifically, the most remarkable finding was that the transition to IRN with a pitch corresponding to that of a 200 Hz sinusoid was a potent stimulator of cortical activity even in the high-pass condition, where there would have been little spectral energy below 2 kHz (24 dB of attenuation at 1 kHz). Since it is generally believed that only the first few partials of an harmonic series can be resolved by the cochlea (Patterson et al., 1996), this substantiates the opinion of earlier authors (e.g. Patterson et al., 1996; Yost, 1996; Krumbholz et al., 2003), that the subjective pitch of IRN and the associated cortical responses are due to the detection of temporal periodicity, rather than the spectral ripples. In any case, it is arguable that the cochlea is not calibrated to recognize equally spaced (in Hz) spectral frequencies, so the extraction of “periodicity” pitch from a harmonic series lacking the lower partials is more likely to be due to processing in the temporal domain. In an earlier study, we (Jones and Perez, 2001) similarly examined the effect of high- and low-pass filtration on the C-potentials to frequency change of a complex tone, believed to represent processing in the spectral domain. In contrast to the present study these were markedly attenuated by the filter, indicated that the underlying neuronal populations have restricted frequency response ranges. The population underlying the IRN response, on the other hand, appears to respond equally well to frequencies below and above

2 kHz, and there was no evidence that a larger population was activated by wide-band IRN. Krumbholz et al. (2003) interpreted the response to the transition from random noise to IRN as a “pitch onset response”. This, of course, refers specifically to the psychoacoustic phenomenon of “periodicity” or “residue” pitch which is independent of the spectral frequencies present. On the basis of earlier MEG studies, Pantev et al. (1989, 1996) and Langner et al. (1997) hypothesized that the periodicity of sounds might be represented in the human supratemporal auditory cortex. This is clearly borne out by the AEP and MEG responses to IRN. However, against the suggestion of Langner et al. (1997) that the same region of cortex may contain spectral and periodicity “maps” with orthogonally oriented frequency and periodicity gradients, the neural populations concerned appear not to be coterminous. On the basis of single equivalent dipole modelling, Krumbholz et al. (2003) concluded that the N1 dipole for IRN was 12.4 mm more anterior, 6.0 mm more medial, and 10.9 mm more inferior than that to the onset of random noise. Of course, this must be interpreted in the light of the many assumptions underlying dipole modelling, and it is entirely possible that the activated neuronal populations may be widespread, partially overlapping one another. It also remains to be confirmed whether there truly exists an orderly “map” of sound periodicity in the supratemporal cortex. Such a map has been found in the low frequency area of the primary auditory cortex of the Mongolian gerbil (Schulze et al., 2002). References Bilsen, F.A. (1966) Repetition pitch: monaural interaction of a sound with the repetition of the same, but phase shifted sound. Acustica, 17: 295–300. Gutschalk, A., Patterson, R.D., Rupp, A., Uppenkamp, S. and Scherg, M. (2002) Sustained magnetic fields reveal separate sites for sound level and temporal regularity in human auditory cortex. Neuroimage, 15: 207–216. Howard 3rd, M.A., Volkov, I.O., Abbas, P.J., Damasio, H., Ollendieck, M.C. and Granner, M.A. (1996) A chronic microelectrode investigation of the tonotopic organization of human auditory cortex. Brain Res., 724: 260–264.

95 Jones, S.J. (2003) Sensitivity of human auditory evoked potentials to the harmonicity of complex tones: evidence for dissociated processes of spectral and periodicity analysis. Exp. Brain Res., 150: 506–514. Jones, S.J. and Perez, N. (2001) The auditory “C-process”: analysing the spectral envelope of complex sounds. Clin. Neurophysiol., 112: 965–975. Jones, S.J., Longe, O. and Vaz Pato, M (1998) Auditory evoked potentials to abrupt pitch and timbre change of complex tones: electrophysiological evidence of ‘streaming’? Electroencephalogr. Clin. Neurophysiol., 108: 131–142. Jones, S.J., Vaz Pato, M. and Sprague, L. (2000) Spectro-temporal analysis of complex tones: two cortical processes dependent on retention of sounds in the long auditory store. Clin. Neurophysiol., 111: 1569–1576. Krumbholz, K., Patterson, R.D., Seither-Preisler, A., Lammertmann, C. and Lütkenhöner, B. (2003) Neuromagnetic evidence for a pitch processing center in Heschl’s gyrus. Cereb. Cortex, 13: 765–772. Langner, G., Sams, M., Heil, P. and Schulze, H. (1997) Frequency and periodicity are represented in orthogonal maps in the human auditory cortex: evidence from magnetoencephalography. J. Comp. Physiol. A, 181: 665–676. Liang, L., Lu, T. and Wang, X. (2002) Neural representations of sinusoidal amplitude and frequency modulations in the primary auditory cortex of awake primates. J. Neurophysiol., 87: 2237–2261. Morel, A., Garraghty, P.E. and Kaas, J.H, (1993) Tonotopic organization, architectonic fields and connections of auditory cortex in macaque monkeys. J. Comp. Neurol., 335: 437–459.

Pantev, C., Hoke, M., Lütkenhöner, B. and Lehnertz, K. (1989) Tonotopic organization of the auditory cortex: pitch versus frequency representation. Science, 246: 486–488. Pantev, C., Elbert, T., Ross, B., Eulitz, C. and Terhardt, E. (1996) Binaural fusion and the representation of virtual pitch in the human auditory cortex. Hear. Res., 100: 164–170. Patterson, R.P., Handel, S., Yost, W.A. and Datta, A.J. (1996) The relative strength of tone and noise components of iterated rippled noise. J. Acoust. Soc. Am., 100: 3286–3294. Picton, D.W., Alain, C., Otten, L. and Ritter, W. (2000) Mismatch negativity: different water in the same river. Audiol. Neuro-otol., 5: 111–139. Schulze, H., Hess, A., Ohl, F.W. and Scheich, H. (2002) Superposition of horseshoe-like periodicity and linear tonotopic maps in auditory cortex of the Mongolian gerbil. Eur. J. Neurosci., 15: 1077–1084. Vaz Pato, M. and Jones, S.J. (1999) Cortical processing of complex tone stimuli: mismatch negativity at the end of a period of rapid pitch modulation. Cogn. Brain Res., 7: 295–306. Wallace, M.N., Shackleton, T.M. and Palmer, A.R. (2002) Phaselocker responses to pure tones in the primary auditory cortex. Hear. Res., 172: 160–171. Yost, W.A. (1996) Pitch strength of iterated rippled noise. J. Acoust. Soc. Am., 100: 3329–3335. Yost, W.A. (1997) Pitch strength of iterated rippled noise when the pitch is ambiguous. J. Acoust. Soc. Am., 101: 1644–1648.


97

Chapter 14

The mystery of photopsias, visual hallucinations, and distortions Gastone G. Celesia* Loyola University of Chicago, 3016 Heritage Oak Lane, Oak Brook, IL 60523 (USA)

Can we unravel the mystery of visual hallucinations and other positive spontaneous visual phenomena (PSVP)? Can we find the source of visual percepts arising endogenously from the human brain? Modern neuro-imaging and physiological methods shed light on the pathophysiology of these phenomena and will be reviewed therein. PSVP are visual events occurring in isolation, in the absence of altered consciousness, dementia, or psychosis (Vaphiades et al., 1996; Celesia, 2005). PSVP include phosphenes, photopsias, visual hallucinations, visual distortions, kinetopsias, palinopsia, polyopia, and visual allesthesia. Phosphenes are unstructured lights such as flashes, sparkles, zigzag lines or rainbows, black and white or colored, static or moving. Photopsias are structured images such as geometric figures (triangles, cubes, pyramids, etc.) or other simple pictures often recurring in a repetitive pattern. Kinetopsia from “kineto” motion and “opsis” vision, is an image moving in space. Palinopsia is visual perseveration or the recurrent appearance of a visual image after the stimulus has disappeared. Polyopia is the simultaneous perception of multiple objects or structures. The same object is

*Correspondence to: Gastone G. Celesia M.D., Loyola University of Chicago, 3016 Heritage Oak Lane, Oak Brook, IL 60523, USA. Tel: +1 630 968-2199; Fax: +1 630 968-2179; E-mail: [email protected]

seen multiple times either side by side or in circles. Visual distortions refer to the abnormal perception of figures and objects and can be further subdivided into macropsia, micropsia, pelopsia, teleopsia, and metamorphopsia. Macropsia refers to the appearance of an enlargement in size of all objects seen (Fig. 1, left illustration), whereas micropsia refers to the appearance of a reduction in size of all the objects seen. Pelopsia refers to the perception that objects are closer while teleopsia refers to the perception that the objects are farther away than they are. Metamorphopsia is the distortion of the objects or figures seen (Fig. 1, right illustration). Visual hallucinations consist of complex scenes including people, animals, furniture, landscapes, vehicles that may be stationary or mobile, colored or black and white, and perceived, at least temporarily, as real. Visual hallucinations as part of PSVP are isolated without associated mental disturbances such as dementia, psychosis, or altered state of consciousness. Ophthalmologists refer to these visual hallucinations as the Charles Bonnet syndrome (Bonnet, 1796). PSVP can be due to lesions in any part of the visual system from the retina to the occipital cortex (Table 1). Thus none of the PSVP have localizing value. It is also not unusual that more than one PSVP occur in the same patient; Jacob (1980) reported visual allesthesia and palinopsia in a 24-year-old woman with a right parieto-occipital arteriovenous malformation. We have noted the co-occurrence of visual hallucinations and

98

Fig. 1. Two illustrations from Lewis Carroll’s, Alice’s adventures in Wonderland, illustration by John Tenniel. Left illustration: Alice is very small and meet a puppy “An enormous puppy was looking down at her with large round eyes, and feebly stretching out one paw, trying to touch her.” Lewis Carroll suffered from migraine (Podoll and Robinson, 1999) and his description of Alice may be a manifestation of the macropsia and visual hallucinations seen in his migrainous auras. Right illustration: Alice stretched tall, an example of metamorphopsia.

photopsias in several patients with lesions in optic radiations and or occipital lobe. Another controversy in the literature is the statement that PSVP related to central nervous system pathology are predominantly due to right hemispherical lesions (Pailas et al., 1965; Vallar and Perani, 1986; Harvey et al., 1994). Such statement ignores the fact that reports on patients with destructive brain lesions are biased by the exclusion of patients with left-side damage due to the associated language deficits that prevent accurate testing. In our study of hemianopic patients (Vaphiades

et al., 1996) despite this built-in bias, we found no statistically significant differences between right- and left-side lesions. PSVP have a prevalence of around 8–9%. Perhaps the most frequently described PSVP are the phosphenes, photopsias, and metamorphopsias of the migraine aura. Russels et al. (1995) in a study of the Danish population reported an overall lifetime prevalence of migraine of 18%, with migraine aura without headaches occurring in 8% of males and 16% of females. Twenty-six of 2110 subjects in the Framingham

99 TABLE 1 LOCATION OF LESION IN 614 CASES OF PSVP Lesion site

Retina Optic nerve Chiasma/tract Optic radiation Occipital cortex Normal subjects Total number of subjects with PSVP Total number of patients

Total (number)

Total (percent)

84 47 5 3 469 6

1.2 0.7 0.1 0.04 6.5 0.1

614 7211

8.5 100.0

study had migraine aura without headaches, a prevalence of 1.23% (Wijman et al., 1998). Lipton et al. (2002) studied the prevalence of migraine in Philadelphia County and reported a prevalence of 13% in adults. There is a remarkable agreement in all these studies that in the Western world the prevalence is approximately 13% with 3% of patients experiencing visual auras and about 1.2% having visual auras without migraine. Charles Bonnet syndrome (visual hallucinations or other PSVP in the presence of ophthalmologic dysfunction) has been reported in a variety of ocular diseases from cataracts to glaucoma, from retinitis pigmentosa to optic neuritis (Gold and Rabin, 1989; Nesher et al., 2001; Volpe et al., 2001; Burke, 2002). In summary, the most frequent cause of PSVP is migraine followed by ophthalmologic diseases associated with impairment of visual acuity equal or worse than 20/200. There are three possible mechanisms causing PSVP (Fig. 2): (1) excitation; (2) release phenomenon; (3) spreading depression (SD). Excitation can be physiological or pathological. Physiological excitation occurs at the retinal, visual pathways, lateral geniculate nucleus (LGN), and cortical level. Stimulation of the eyeball by deformation, pressure, saccades, electrical and magnetic stimulation produces phosphenes and photopsias by activating

retinal neuronal circuitry and stimulating ganglion cells (Grüsser and Landis, 1991; Lindenblatt and Silny, 2002). Phosphenes were noted in subjects exposed to highdose X-rays. ERGs can be recorded to X-ray stimulation and the amplitude of the b-wave is related to the intensity of the X-ray stimulus (Doly et al., 1980a, b). Astronauts in space have reported seeing flashes of light in the darkness presumably due to high-energy particles (cosmic ray) stimulating the retina photoreceptors (Casolino et al., 2003). Phosphenes and photopsias have been evoked by electrical stimulation of the optic tract, LGN and optic radiations during neurosurgery in humans. Electrical stimulation of the striate cortex and area V2 evokes simple visual forms, while stimulation of surrounding extrastriate areas evokes intermediate visual forms and stimulation of temporal regions elicits visual hallucinations (Penfield and Jasper, 1954; Lesser et al., 1998; Lee et al., 2000). Pathological excitation. Epilepsy is the prototype of a pathological excitatory phenomenon. Phosphenes, photopsias, palynopsia, and polyopia have been described as an aura preceding partial complex seizures and attributed to an epileptic focus localized in the occipital cortex. Visual hallucinations and visual distortions (metamorphopsia, micropsia, macropsia, teleopsia, and pelopsia) are observed at the onset of partial complex seizures originating in the temporal lobe (Penfield and Jasper, 1954). Release phenomenon may produce PSVP. A release phenomenon is essentially a disinhibition of neuronal structures. As a result of disinhibition, there is an increase in the excitability of the disinhibited neurons and an increase in spontaneous activity causing PSVP. fMRI obtained in patients experiencing visual hallucinations correlated with increased cerebral activity in extrastriate cortex (Fftyche et al., 1998). The type of PSVP (from phosphenes to visual hallucinations) will then depend on the specific visual cortical areas activated by the deafferentation. Burke (2002) suggested that visual “hallucinations can be regarded as local paroxysms in the sensory system.” PSVP in macular degeneration, glaucoma, cataracts, macular holes, and other retinopathies (Charles Bonnet syndrome) can be explained as a release phenomenon.

100

PSVP Pathophysiology

Excitation

Pathological

Physiological

Deformation Phosphenes

Exposure to High dose X-ray

Exposure to Cosmic Ray (astronauts)

Electrical or TMS Stimulation CNS Structures

Epilepsy

Release Phenomenon

Charles Bonnet Syndrome with Ocular pathology

Charles Bonnet Syndrome with CNS Pathology

Spreading Depression

Migraine or Migraine equivalent

Fig. 2. Mechanisms underlying PSVP.

Spreading depression (SD). SD is a depression of cerebral activity that slowly spread to adjacent structures. SD consists of a propagating negative potential with amplitude of 10–30 mV lasting 0.5–1 min (Gorji, 2001). The phenomenon is associated with slow negative potentials, cellular dysfunction, and increase in release of K+ and H+, inducing a refractory period to further stimulation that may last minutes or hours. The wave of depression propagates slowly (3–5 mm/min) in all directions. The visual aura of migraine has been attributed to SD as demonstrated by cerebral blood flow measurements of spreading hypoperfusion (Cao et al., 1999; Gorji, 2001; James et al., 2001). Hadjikhani et al. (2001) recorded fMRI in 3 migraneurs during their visual aura. They reported an increase in blood oxygen level-dependent (BOLD) signals first in the extrastriate

visual areas with spreading at the speed of 3.5 ± 1.1 mm/min over the occipital cortex. These changes were associated with clinical scintillations and were followed by a 15 ± 3 min decrease in MR signal. This decrease was associated with clinical scotoma. Many artists including Pablo Picasso suffered from migraine. Are some of their pictures influenced by their experience as migraneurs? Is SD responsible for artistic expression? Are Pablo Picasso paintings of cubism period a representation of metamorphopsia seen during its migraine’s aura? We will never know for sure but it may be interesting to speculate about the influence of diseases on artistic creativity. In the differential diagnosis of PSVP, we need to consider associated symptoms and signs. Foremost, we need to exclude visual hallucinations and other visual phenomena that arise in patients with delirium

101

PSVP

PSVP with Psychosis

Schizophrenia

Dementia (AD, etc,)

Paranoid States

Affective Psychoses

PSVP with Delirium

Alcohol or Drugs

Systemic or CNS Infections

Agitated Delirium With hemianopia

PSVP with Clear Sensorium

Charles Bonnet Syndrome

Physiologic

Ocular Pathology

Migraine or Migraine equivalent

Epilepsy

Central Visual Pathways Lesions

Fig. 3. Algorithm for PSVP diagnosis.

or psychosis (Fig. 3). Approximately 2% of elderly patients referred to a psychiatric clinic were eventually diagnosed to have Charles Bonnet syndrome related to eye pathology (Berrios and Brooks, 1984; NortinWillison and Munir, 1987). Thus, a careful psychiatric history and mental status examination are required before we conclude that a patient suffers from PVSP. Alcohol, drug withdrawal, or intoxications are the most frequent causative agents of delirium. Delirium can also be seen in systemic infections with high temperature or in infections of the nervous system. However, occasionally visual hallucinations, and delirium are seen in patients with acute insults to the occipital lobe. We have termed this syndrome agitated delirium with hemianopia (Vaphiades et al., 1996). The syndrome is characterized by agitation, confusion, visual hallucinations, and is always associated

with homonymous hemianopia. This syndrome is indicative of specific anatomical lesions involving the mesial aspect of the occipital lobe, the parahippocampal gyrus, and the hippocampus (Medina et al., 1974; Vaphiades et al., 1996). PSVP have been under-reported and underrecognized both by ophthalmologists and neurologists (Siatkowski et al., 1990; Vaphiades et al., 1996). Often individuals with these benign PSVP are treated with neuroleptics and undergo other unnecessary psychiatric treatments. There is some controversy on the definition of Charles Bonnet syndrome and specifically whether it should be limited to ocular pathology or be inclusive of all visual pathways’ lesions. Chaudhuri (2000) defines the syndrome by the triad of “visual hallucinations, visual sensory deprivation, and preserved

102 cognitive status” and includes in the syndrome any visual sensory deprivation whether caused by ocular or central visual pathway deficits. The term Charles Bonnet syndrome applied to both ocular and central vision lesions may be beneficial to physicians and patients in understanding the benign nature of these phenomena and focus on the treatment of the underlying ocular or central pathology. Visual hallucinations in psychosis are associated with auditory hallucinations or with sound, whereas PSVP are always silent. Bonnet (1796) was the first to observe that visual hallucinations in his syndrome were not associated with sound: “because the men and women did not speak, and no sound affected the ear.” Vaphiades et al. (1996) similarly remarked on the absence of sound with PSVP limited to the affected hemifield. Thus the presence of “silent” visual hallucinations or other PSVP should alert the clinician to the possible presence of visual dysfunction. References Berrios, G.E. and Brooks, P. (1984) Visual hallucinations and sensory delusions in the elderly. Br. J. Psychiatr., 144: 662–664. Bonnet, C. (1796) Essai Analytique sur les Faculté de l’Ame. Philibert, Copenhagen, pp. 426–429. (Reprinted by G. Olms Verlag, Hildesheim, 1973, 552 pp.) Burke, W. (2002) The neural basis of Charles Bonnet hallucinations: a hypothesis. J. Neurol. Neurosurg. Psychiatr., 73: 535–541. Cao, Y., Welch, K.M.A., Aurora, S. and Vikingstad, E.M. (1999) Functional MRI-BOLD of visually triggered headache in patients with migraine. Arch. Neurol., 56: 548–554. Casolino, M., Bidoli, V., Morselli, A., Narici, L., DePascale, M.P., Picozza, P., Reali, E., Spacvoli, R., Mazzenga, G., Ricci, M., Spillantini, P., Buezio, M., Bonvicini, V., Vacchi, A., Zampa, N., Castellini, G., Sannita, W.G., Carlson, P., Galper, A., Korotkov, M., Popov, A., Vavilov, N., Avdeev, S. and Fugelsang, C. (2003) Space travel – dual origins of light flashes seen in space. Nature, 422: 680. Celesia, G.G. (Ed.) (2005) Disorders of Visual Processing; Handbook of Clinical Neurophysiology, Vol. 5. Elsevier, Amsterdam. Chaudhuri, A. (2000) Charles Bonnet Syndrome: an example of cortical dissociation syndrome affecting vision. J. Neurol. Neurosurg. Psychiatr., 69: 704–705. Doly, M., Isabelle, D.B., Vincent, P., Gaillard, G. and Meyniel, G. (1980a) Mechanisms of the formation of X-ray induced phosphenes. I. Electrophysiological investigation. Radiat. Res., 82: 93–105. Doly, M., Isabelle, D.B., Vincent, P., Gaillard, G. and Meyniel, G. (1980b) Mechanisms of the formation of X-ray induced phosphenes. II. Photochemical investigation. Radiat. Res., 82: 430–440.

Fftyche, D.H., Howard, R.J., Brammer, M.J., David, A., Woodruff, P. and Williams, S. (1998) The anatomy of conscious vision: an fMRI study of visual hallucinations. Nature Neurosci., 1: 738–742. Gold, K. and Rabin, P.V. (1989) Isolated visual hallucinations and the Charles Bonnet syndrome: a review of the literature and presentation of six cases. Compr. Psychiatry, 30: 90–98. Gorji, A. (2001) Spreading depression: a review of the clinical relevance. Brain Res. Rev., 38: 33–60. Grüsser, O.J. and Landis, T. (1991) Visual Agnosia and Other Disturbances of Visual Perception and Cognition. CRC Press, Boca Raton, 610. pp. Hadjikhani, N., Sanchez Del Rio, M., Wu, O., Schwartz, D., Bakker, D., Fischl, B., Kwong, K.K., Cutrer, F.M., Rosen, B.R., Tootell, R.B., Sorensen, A.G. and Moskowitz, M.A. (2001) Mechanisms of migraine aura revealed by functional MRI in human visual cortex. Proc. Natl. Acad. Sci. USA, 98: 4687–4692. Harvey, M., Milner, A.D. and Roberts, R.C. (1994) Spatial bias in visually-guided reaching and bisection following right cerebral strokes. Cortex, 30: 343–350. Jacob, L. (1980) Visual allesthesia. Neurology, 30: 1059–1063. James, M.F., Smith, J.M., Boniface, S.J., Huang, C.L. and Leslie, R.A. (2001) Cortical spreading depression and migraine: new insights from imaging? Trends Neurosci., 24: 266–271. Lee, H.W., Hong, S.B., Seo, W., Tae, D.W. and Hong, S.C. (2000) Mapping of functional organization of human visual cortex. Electrical cortical stimulation. Neurology, 54: 849–854. Lesser, R.P., Arroyo, S., Crone, N. and Gordon, B. (1998) Motor and sensory mapping of the frontal and occipital lobes. Epilepsia, 39: S69–S80. Lindenblatt, G. and Silny, J. (2002) Electrical phosphenes: on the influence of conductivity in homogeneities and small-scale structures of the orbita on the current density threshold of excitation. Med. Biol. Eng. Comput., 40: 354–359. Lipton, R.B., Scher, A.I., Kolodner, K., Liberman, J., Steiner, T.J. and Stewart, W.F. (2002) Migraine in the United States: epidemiology and patterns of health care use. Neurology, 58: 885–894. Medina, J.L., Rubino, F.A. and Ross, E. (1974) Agitated delirium caused by infarction of the hippocampus formation and fusiform and lingual gyri: a case report. Neurology, 24: 1181–1183. Nesher, R., Nesher, G., Epstein, E. and Assia, E. (2001) Charles Bonnet syndrome in glaucoma patients with low vision. J. Glaucoma, 10: 396–400. Norton-Willison, L. and Munir, M. (1987) Visual perceptual disorders resembling the Charles Bonnet syndrome: a study of 434 consecutive patients referred to a psychogeriatric unit. Fam. Pract., 4: 27–31. Paillas, J.E., Cossa, P., Darcourt, G. and Naquet, R. (1965) Etude sur l”épilepsie occipitale (a` propos de 36 observations de lésions occipitales verifiées chirurgicallement). Proceedings of 8th International Congress Neurology III, 193. Penfield, W. and Jasper, H. (1954) Epilepsy and the Functional Anatomy of the Human Brain. Little Brown, Boston, 896 pp. Podoll, K. and Robinson, D. (1999) Lewis Carroll’s migraine experiences. Lancet, 353: 1366. Russell, M.B., Rasmussen, B.K., Thorvaldsen, P. and Olesen, J. (1995) Prevalence and sex-ratio of the subtypes of migraine. Int. J. Epidemiol., 24: 612–618.

103 Siatkowski, R.M., Zimmer, B. and Rosenberg, P.R. (1990) The Charles Bonnet syndrome. Visual perceptive dysfunctionin sensory deprivation. J. Clin. Neuroophthalmol., 10: 215–218. Vallar, G. and Perani, D. (1986) The anatomy of unilateral neglect after right-hemisphere stroke lesions. A clinical CT-scan correlation study in man. Neuropsychologia, 24: 609–622. Vaphiades, M.S., Celesia, G.G. and Brigell, M.G. (1996) Positive spontaneous visual phenomena limited to the hemianopic

field in lesions of central visual pathways. Neurology, 47: 408–417. Volpe, N.J., Rizzo III, J.F. and Lessell, S. (2001) Acute idiopathic blind spot enlargement syndrome: a review of 27 new cases. Arch. Ophthalmol., 119: 59–63. Wijman, C.A., Wolf, P.A., Kase, C.S., Kelly-Hayes, M. and Beiser, A.S. (1998) Migrainous visual accompaniments are not rare in late life. The Framingham study. Stroke, 29: 1539–1543.


107

Chapter 15

Intraoperative neurophysiology of the corticospinal tract of the spinal cord Vedran Deletis* Institute for Neurology and Neurosurgery, St. Luke’s-Roosevelt Hospital, 1000 10th Avenue, New York, NY 10019-1147 (USA)

Intraoperative monitoring (IOM) of the functional integrity of the corticospinal tract (CT) has specific requirements, given the pharmacological influence of applied anesthetics and the limitations inherent to surgery on an unconscious patient. These factors present advantages and disadvantages for both IOM and the continued exploration of the physiology of the CT. With this in mind, several IOM methodologies have been developed which give us a unique opportunity to explore the physiology of the CT, especially within the spinal cord. The application of these methodologies expands what is possible during interoperative monitoring of the functional integrity of the motor system. These methodologies include: 1. Transcranial electrical stimulation (TES) or direct electrical stimulation (DES) TES or DES of the surgically exposed motor cortex and subcortical pathways with recording of electrical activity of fast corticospinal tract fibers (fCTF), from

*Correspondence to: Vedran Deletis M.D. Ph.D., Institute for Neurology and Neurosurgery, St. Luke’s-Roosevelt Hospital, 1000 10th Avenue, New York, NY 10019-1147, (USA). Tel: +1 (212)-636 3280; E-mail: [email protected]

the spinal cord. Recordings of this activity from the spinal cord are in the form of D and I waves while motor evoked potentials record activity from limb muscles (mMEPs) (Fig. 1). Combined recordings of the D wave and mMEPs during spinal cord surgery allow us to estimate better the functional integrity of the motor system involved in the generation of mMEPs. Combining these methodologies (D wave and mMEP monitoring) proves more advantageous than using each method alone and allows for more accurate intraoperative estimations of postoperative prognosis. For the majority of patients undergoing surgery for intramedullary spinal cord tumors, combined use of these methods has allowed us to prevent injury to the CT and can distinguish patients with transient motor deficits from those who have permanent postoperative motor deficits (Kothbauer et al., 1998). Furthermore, studies on the recovery of D wave amplitude following double pulse transcranial stimulation gives us the opportunity to explore the optimal interstimulus interval required for the train of stimuli used to elicit mMEPs (Novak et al., 2004). 2. The D wave collision technique This technique involves simultaneous TES of the motor cortex with concurrent stimulation of the CT from

108

B

A

C

D

Fig. 1. (A) Schematic illustration of electrode positions for transcranial electrical stimulation of the motor cortex according to the International 10–20 EEG system. The site labeled “6 cm” is 6 cm anterior to Cz. (B) Illustration of grid electrode overlying the motor and sensory cortexes. (C) Schematic diagram of the positions of the catheter electrodes (each with three recording cylinders) placed cranial to the tumor (control electrode) and caudal to the tumor to monitor the descending signal after passing through the site of surgery (left). In the middle are D and I waves recorded rostral and caudal to the tumor site. On the right is depicted the placement of an epidural electrode through a flavectomy/flavotomy when the spinal cord is not exposed. (D) Recording of muscle motor evoked potentials from the thenar and tibial anterior muscles after eliciting them with multipulse stimuli applied either transcranial or over the exposed motor cortex. (Modified with permission from Deletis et al., 2001.)

109

Fig. 2. Mapping of the spinal cord for the CT using D wave collision technique (see text for explanation). S1 – transcranial electrical stimulation (TES); S2 – spinal cord electrical stimulation (SCES); D1 – Control D wave (TES only); D2 – D wave after combined stimulation of the brain and spinal cord; R – D wave recording electrode in the spinal epidural space cranial and caudal to the spinal cord pathology. Below left: negative mapping results (D1 = D2). Below right: positive mapping results (D2 < D1).

the surgically exposed spinal cord caudal to a lesion (Fig. 2). This newly developed technique has allowed us to: (a) Intraoperatively map the anatomical position of the CT within the surgically exposed spinal cord. In 18 patients undergoing surgery for thoracic intramedullary spinal cord tumors, we elicited D waves by TES and recorded them cranial and caudal to the spinal cord tumor. Simultaneous to TES, we stimulated the surgically exposed spinal cord (caudal to the tumor) with a miniature bipolar hand-held probe (#5522.010 INOMED, GmbH, Germany). The tips of the probe were 1 mm apart delivering constant current stimuli up to 2 mA intensity, 0.5 ms duration, and 1 Hz repetition rate. The D wave elicited by TES collided with “anti D wave” elicited by spinal cord stimulation whenever the stimulating probe was in a close proximity to the CT. This collision resulted in diminished amplitude of the D wave recorded cranial to the lesion after collision (see Fig. 2; Deletis and Camargo, 2001).

Initial use indicates an impressive ability to selectively map the spinal cord for the CT. Using this method, the CT can be localized to within 1 mm. This is in concordance with the other CT mapping techniques used to stimulate cranial nerves’ motor nuclei within the brainstem, which show the same degree of selectivity (Deletis et al., 2000). (b) Semi-quantitatively determine the number of unaffected, desynchronized, and blocked (nonconducting) fCTF. Pathology of the spinal cord affects fCTF in three ways; some of the fibers passing through the lesion site are not affected, some of them become desynchronized and each fiber conducts D and I waves with a different conduction velocity. The third group of fCTF become blocked by the lesion and are non-functional and non-conducting (Fig. 3A). D and I waves from unaffected fCTF can be easily recorded caudal to the spinal cord lesion while these waves cannot be recorded from desynchronized or blocked

110 A

B

C

Fig. 3. Schematic of semi-quantitative calculation of unaffected, desynchronized, and blocked fast CT fibers. (see text for explanations). S1 – transcranial electrical stimulation (TES); S2 – stimulation of the spinal cord caudal to the lesion site; R1 – recording of the D wave cranial to the spinal cord lesion; R2 – recording of the D wave caudal to the spinal cord lesion.

fCTF. However, even desynchronized fCTF can elicit mMEPS (Fig. 4) indicating that these fibers remain functional. Before collision takes place (TES alone): the amplitude of the D wave recorded cranial to the lesion semi-quantitatively represents the total amount of fCTF approaching the lesion. The amplitude of the D wave recorded caudal to the lesion represents the amount of unaffected axons of the FCN transversing lesion (Fig. 3B). The amplitude difference between the cranial (R1) and caudal (R2) D waves represents the total amount of blocked and desynchronized fCTF since they do not contribute to the D wave amplitude caudal to the lesion and are not recorded (Fig. 3B).

After collision takes place (TES together with spinal cord stimulation): the amplitude of the cranial recorded D wave (R1) represents the amount of blocked (non-conducting) fCTF since stimulation of the spinal cord caudal to the lesion activates both unaffected and desynchronized fCTF (but not blocked fibers since they can not conduct through the lesion). Activation of desynchronized and unaffected fibers by caudal stimulation of the spinal cord, produces a volley which travels as an “anti-D wave” moving antidromically. This wave collides with the descending D wave elicited by TES (Fig. 3C) and diminishes its amplitude. By subtracting the D wave amplitude difference in Fig. 3B (R1 − R2) from the D wave

111

A

B Fig. 4. (A) Recording of a D wave cranial (upper trace) and caudal (lower trace) to an intramedullary spinal cord tumor. Note the well synchronized D wave cranial, in contrast to the desynchronized D wave caudal to the tumor. The arrow indicates the beginning of the stimulus artifact. (B) Very small amplitude D wave (due to extreme desynchronization) not suitable for monitoring caudal to a high cervical intramedullary tumor. Despite this desynchronized D wave, large muscle mMEPs were recorded from a small hand muscle elicited after a short train of 6 stimuli (to the right). (Modified with permission from Deletis and Kothbauer, 1998.)

amplitude in Fig. 3C (R1) one can semi-quantitatively determine the amount of desynchronized but still functioning fCTF: (non-conducting + desynchronized) − non-conducting = desynchronized. (c) To expand the benefits of the D wave monitoring method by further investigating the theoretical

background behind the monitoring of completely desynchronized D waves caudal to a spinal cord lesion. In patients in whom the D wave is extremely desynchronized, it is practically non-recordable (Fig. 4; Deletis and Kothbauer, 1998; Morota et al., 1997). However, electrical stimulation of the spinal cord caudal to the site of pathology

112 will activate these desynchronized CT axons and their antidromic activity will collide with the D wave elicited by TES. Therefore, the D wave will diminish in amplitude. If the conductivity of the CT transversing a lesion becomes compromised during surgery, this will result in an increased amplitude of the D wave cranial to the lesion (due to the smaller number of fCTF able to conduct a colliding volley toward the D wave approaching the lesion). Thus, this method will allow us to monitor the functional integrity of the desynchronized but still functional fCTF transversing a lesion when they are not recordable with existing methodologies of D wave recording. References Deletis, V. and Camargo, A.B. (2001) Interventional neurophysiological mapping during spinal cord procedures. Stereotact. Funct. Neuosurg., 77: 25–28.

Deletis, V. and Kothbauer, K. (1998) Intraoperative neurophysiology of the corticospinal tract. In: E. Stålberg, H.S. Sharma and Y. Olsson (Eds.), Spinal Cord Monitoring. Springer, Vienna, pp. 421–444. Deletis, V., Sala F. and Morota, N. (2000) Intraoperative neurophysiological monitoring and mapping during brain stem surgery: a modern approach. Operative Tech. Neurosurg., 2(2): 109–113. Deletis, V., Rodi, Z. and Amassian, V.E. (2001) Neurophysiological mechanisms underlying motor evoked potentials (MEPs) elicited by a train of electrical stimuli. 2. Relationship between epidurally and muscle recorded MEPs in man. Clin. Neurophysiol., 112: 445–452. Kothbauer, K., Deletis, V. and Epstein, F.J. (1998). Motor evoked potential monitoring for intramedullary spinal cord tumor surgery: correlation of clinical and neurophysiological data in a series of 100 consecutive procedures. Neurosurg. Focus, 4(5): 1–9. Morota, N., Deletis, V., Shlomi, C., Kofler, M., Cohen, H. and Epstein, F. (1997) The role of motor evoked potentials (MEPs) during surgery of intramedullary spinal cord tumors. Neurosurgery, 41: 1327–1366. Novak, K., Camargo, A.B., Neuwirth, M., Kothbauer, K., Amasian, V.E. and Deletis, V. (2004) The refractory period of fast conducting corticospinal tract neurons in man and its implications for intraoperative monitoring of motor evoked potentials. Clin. Neurophysiol., 115: 1931–1941.


113

Chapter 16

Generators of subcortical components of SEPs and their clinical applications Masahiro Sonoo* Department of Neurology, Teikyo University School of Medicine, Tokyo 173-8605 (Japan)

1. Introduction Somatosensory evoked potentials (SEPs) have always been a key subject of research in clinical neurophysiology, especially in the 1970s and 1980s, and numerous theories have been put forward regarding the origin of each SEP component. Efforts to clarify the mechanism by which SEP components are generated have enhanced our understanding of electrical phenomena occurring within the volume conductor of the human body. However, after most SEP component generators were determined and the theory of far-field potential was established (Dumitru and Jewett, 1993), investigators seem to have lost interest in this area, now considered “classical” methodology. SEPs should therefore be included in routine examinations and, if this is to occur, many studies on the clinical application of SEPs should have followed the initial research; unfortunately, however, there are relatively few such studies. The use of SEPs in diagnostic medicine is less well established than auditory brainstem responses (ABRs) and visual evoked potentials (VEPs), each playing definite roles

*Correspondence to: Masahiro Sonoo, Department of Neurology, Teikyo University School of Medicine, Kaga 2-11-1, Itabashi-ku, Tokyo 173-8605, Japan. Tel: +81 3 3964 1211; Fax: +81 3 3964 6397; E-mail: [email protected]

in the evaluation of hearing acuity and brainstem function or optic nerve lesion in multiple sclerosis, despite the potential and promising capacity of SEPs to evaluate the long somatosensory pathway through the central nervous system as well as the peripheral nerves. Practically speaking, technicians in many institutes cannot and would not perform SEP examinations, although they routinely perform ABR and VEP examinations, and many neurologists are not confident when to order, or perform, SEPs. In this review, I would like to discuss the potential use of SEPs in diagnostic medicine, as well as what is necessary to promote their utilization. 2. The utility of SEPs: general consideration As SEPs, notably their subcortical components, can detect the activity of a number of structures along the somatosensory pathway, they are expected to be able to localize lesions, namely to determine where the somatosensory pathway is interrupted. Here, highly developed imaging tests, especially MRI, are potent rivals. When considering the utility of SEPs, we must consider what SEPs can provide that MRI cannot. First, some lesions are not detected by MRI, including degenerative, metabolic, and peripheral demyelinating diseases. SEPs must be primarily useful for such diseases. Second, MRI provides no information about

114 function; therefore, we cannot judge whether subtle to moderate changes frequently seen in aged subjects are pertinent to the present symptoms, such as sensory disturbances. Spondylotic changes in the cervical spine (Boden et al., 1990) and subcortical ischemic changes are such examples. Lastly, SEPs can serve as the first test to determine where to use MRI. This role is especially useful when the patient shows obvious sensory loss. In this sense, SEPs are comparable to motor nerve conduction study (MCS) in the presence of muscular weakness, and possibly superior because not only the peripheral but central nervous system can be investigated, and the signal generated by the pathway itself, even if it is interrupted halfway, can be detected, which is not possible for MCS. As mentioned above, the potential power of SEPs has not yet been fully demonstrated for clinical applications. Fundamental requirements to improve the clinical applications of SEPs are as follows: first, we should know the precise generator of each SEP component. It is not until complete identification of the origin of every SEP component that accurate localization of the lesion becomes possible. Second, we should know the normal variations of SEP waveforms and the normal limit values of each parameter. As the sensory pathway evaluated by SEPs extends over a long distance, the values of latency or interval parameters would be greatly influenced by the height or age of the subject. We established normal values while fully considering subject factors such as height and age for both median and tibial SEPs (Sonoo et al., 1996a; Miura et al., 2003). Intersubject variation of the SEP waveform is substantial and loss of a particular component, e.g. P11 in median SEPs, should not be immediately regarded as pathological (Sonoo et al., 1996a). The last and most practical point is that we should control the recording technique. Here, the greatest problem is the control of stimulus intensity. In our experience, the most frequent cause of erroneous results in SEP recording is inadequate stimulus intensity, which is especially problematic for tibial SEPs. We propose new standard methods to ensure sufficient stimulus intensity for both median and tibial SEPs (Miura et al., 2003; Fukuda et al., in press). The following section concentrates on the first requirement and presents examples of how accurate

identification of the origin of an SEP component contributes to the progression of its clinical applications. 3. Origin of P15 in tibial nerve SEPs and its clinical applications Through important experimental and simulation studies, the generation mechanism of the far-field potential has been correctly understood. According to Dumitru and Jewett (1993), the far-field potential (junctional potential) is generated by one of the following four mechanisms (Fig. 1): (1) change in the volume conductor size; (2) change in volume conductor impedance; (3) nerve course change; (4) beginning or termination of the nerve action potential. Kimura et al. (1986) described the characteristics that a junctional potential should retain: (1) The volume conductor is divided into two compartments regarding the junctional potential. (2) Two electrodes within the same compartment are isopotential. (3) The junctional potential is recorded between two electrodes each belonging to a different compartment. Since establishing the concept of the junctional potential, many SEP components have been attributed to the junctional potential. However, most such opinions remained mere speculation and did not properly consider the above characteristics of the junctional potential. In my opinion, the origin of no SEP components had been completely elucidated in terms of a junctional potential before our documentation of the P15 potential in tibial SEPs (Sonoo et al., 1992a). P15 (or P17) in tibial nerve SEP was first described as the earliest “far-field potential” recorded in the scalpcontralateral knee lead (Kc) (Yamada et al., 1982). However, it was a small potential and was difficult to record because of artifact contamination. We investigated the characteristics of the P15 potential and demonstrated that it is a typical example of the junctional potential for the following reasons (Sonoo et al., 1992a): first, we investigated the distribution of the P15 potential using common reference at the greater trochanter contralateral to the stimulation (GTc), and

115

A

B

C

D

E

F

Fig. 1. Generating mechanism of far-field potentials (junctional potentials). (Reproduced with minor revision by permission of Medical Review Co. Ltd from Sonoo, M. (1998) Somatosensory evoked potentials. Brain Medical (Tokyo), 10: 131–136.) (A) Postsynaptic potentials can produce an electrical dipole, hence a far-field potential, if they are generated on one end of a long-shaped neuron. (B) Conducting action potentials are equated to two equivalent electrical dipoles, a leading dipole and a trailing dipole. Although a near-field potential is recorded, no far-field potential is registered because the two dipoles cancel each other out on a distant recording electrode. Far-field potentials (junctional potentials) are registered when the symmetry of the two dipoles are broken. This occurs when the size of the volume conductor changes (C), when the impedance of the volume conductor changes (D), when the nerve course changes (E), or at the beginning or termination of the nerve action potential (F).

found that P15 is distributed over the contralateral and rostral portion of the buttock, whereas N15 with the same latency is distributed over the ipsilateral and caudal buttock (Fig. 2). The Cz′ electrode over the scalp also registered a P15, and thus the body was divided into two compartments of P15 and N15. The Cz′-ICC (contralateral iliac crest) lead registered no potential despite the long interelectrode distance, indicating that two electrodes belonging to the same compartment have isopotential. In contrast, when the ICC electrode was connected to the GTi (ipsilateral greater trochanter) electrode, a large P15 potential was recorded, i.e. the lead connecting two electrodes belonging to different compartments registered a definite potential. In this way, the P15 potential perfectly fulfilled the three characteristics of the junctional potential. Furthermore, there is no synapse around the suspected location of the P15 generator, indicating that it must be a junctional potential. By investigating sequential bipolar leads and comparing with simulation studies, we concluded that P15 is a junctional potential generated when the sciatic nerve enters the bone at the greater sciatic foramen (Sonoo et al., 1992a).

The ICC-GTi lead was best suited to recording P15 because it has the largest amplitude for this derivation and nonetheless, the artifact contamination was relatively small, reflecting the short interelectrode distance. We demonstrated that P15 in this lead was clearly and easily recorded in every control subject, including elderly subjects (Miura et al., 2003). This component is expected to be a useful tool to evaluate the most proximal segment of the tibial nerve, just distal to the dorsal root ganglion, i.e. it was useful to diagnose diseases affecting the cauda equina and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) (Sonoo, 1996; Sonoo et al., 1993). Here, I would like to demonstrate an example of how P15 and the following components in tibial SEPs contributed to the determination of the primary lesion site of tabes dorsalis (Sonoo et al., 2005). There is controversy regarding the primary lesion site of tabes dorsalis (Gray and Alonso, 2002), and the proximal and distal ends of the dorsal roots are two potential candidates. The former is known as the theory of Obersteiner and Redlich who emphasized the lesion where the cauda equina pierces the pia mater, whereas the latter is the

116 the first peak was completely lost. By applying stimuli at two different intensities, we demonstrated that the first peak, which was lost in the tabetic patient, is the sensory component, whereas the second peak is the antidromic volley of the motor nerve. P15 preservation and selective loss of the sensory component of cauda equina potentials from the most caudal lumbar bipolar lead (L5SL4S) in this early tabetic patient support the hypothesis that the primary lesion site of tabes dorsalis is at the distal end of the dorsal root, namely the “radicular nerve” theory of Nageotte and Richter.

Fig. 2. P15 in tibial nerve SEPs as a junctional potential. (Reproduced with minor revision by permission of Seiwa Shoten Co. Ltd from Sonoo, M. (2003) Somatosensory evoked potentials (SEPs) (1): SEP montage and origin of each component. Brain Science (Tokyo), 25: 991–998.) The waveform at the side of each electrode is recorded with a common reference at the contralateral greater trochanter (GTc). P15 is distributed over rostral and contralateral regions, including the scalp, while N15 is over caudal and ipsilateral regions. The contralateral knee (Kc) belongs to a slightly negative compartment and therefore small negativity is recorded there. Therefore, the Cz′-Kc lead registers P15, with significant artifact contamination because of the long interelectrode distance. The Cz′-ICC lead registers no potential because both electrodes belong to the same P15 compartment. The ICC-GTi lead registers a large P15 with low artifact contamination, respectively because they belong to different compartments. These characteristics coincide with those of a junctional potential, and therefore we concluded that P15 is generated when the sciatic nerve enters the bone at the greater sciatic foramen (arrow).

theory of Nageotte and Richter who gave importance to the lesion at the so-called “radicular nerve” (Greenfield, 1963). We examined tibial SEPs in a patient showing the typical clinical profile of tabes dorsalis with a relatively short duration of symptoms, who improved after penicillin treatment. P15 was normally preserved, whereas N21 and P38 were lost. The lumbar bipolar leads registered a traveling peak 5–6 ms after the P15 peak (Fig. 3), the latency of which corresponded to the second peak in some control subjects after the first and main traveling peak 2–3 ms after the P15 peak. In the tabetic patient,

4. Origin of P13/P14 in median nerve SEPs and its clinical applications The origin of the P13/P14 complex in median nerve SEPs recorded over the scalp using a noncephalic (NC) reference is one of the most controversial issues regarding SEP generators. The P13/P14 complex often appeared bilobed (Cracco and Cracco, 1976; Desmedt and Cheron, 1980; Anziska and Cracco, 1981) or even trilobed (Sonoo et al., 1997) with substantial intersubject variability, which further confused the controversy. We investigated the intersubject variability and found that the P13/P14 complex is composed of three latent subcomponents, named P13, P14a, and P14b (Sonoo et al., 1997). Various proposed origins of the P13/P14 complex include the medial lemniscus, cuneate nucleus, thalamus or thalamocortical tract, junctional potential at the foramen magnum, presynaptic dorsal column fibers, and high cervical cord. Among them, the most influential opinion ascribes the main origin of the P13/P14 complex to the medial lemniscus (Nakanishi et al., 1978; Anziska and Cracco, 1981; Desmedt and Cheron, 1981). However, no satisfying explanation as to why activity at the medial lemniscus produces a far-field potential over the scalp has been provided. We demonstrated that only the beginning of the medial lemniscus was sufficient to generate the principal part of P13/P14 (Sonoo et al., 1996b), and proposed that the earlier two subcomponents of the P13/P14 complex, P13 and P14a, are junctional potentials generated by the onset of two repetitive bursts of the ascending lemniscal volley (Sonoo et al., 1997).

117

Fig. 3. Tibial nerve SEPs in a tabetic patient at two different intensities. (Reproduced with minor revision by permission of Lippincott Williams & Wilkins from Sonoo, M., Katayama, A., Miura, T., Shimizu, T. and Inoue, K. (2005) Tibial nerve SEPs localized the lesion site in a patient with early tabes dorsalis. Neurology, 64: 1452–1454.) (A) Motor threshold (stimulus intensity 5.6 mA). A ring electrode at the big toe registers only an antidromic sensory nerve action potential (SNAP). A prominent P15 is observed but no later components are identified except for a small and delayed negative cortical component in the Cz′-ICC lead (open square in the lowermost compressed trace). A small negative peak in the L5S-L4S lead (asterisk) is too early for N17 (cauda equina potential) at this lead (0.9 ms after P15; N17 at this lead is normally 2.1 ± 0.6 ms after P15) and may be related to the P15 far-field potential or to the termination of the ascending potential. A small positive peak in the L1S-ICC lead (asterisk) must have similar significance. (B) Well over the motor threshold (stimulus intensity 7.6 mA). The ring electrode registers prominent contamination of the compound muscle action potential (CMAP) after the SNAP. Late traveling peaks at 5.1–5.9 ms after the P15 peak are now observed in the lumbar bipolar leads, as well as the notch after the P15 peak, which are interpreted as the antidromic volleys of motor axons. The L1S-ICC lead registers small positive–negative components (open circles) that are not observed in (A), and which must be related to the approach and termination of the antidromic motor volleys. The late cortical component is essentially the same as in (A) (open square).

We underscored that the distribution of the P13/P14 complex also indicates its nature as a junctional potential (Sonoo, 2003), i.e. scalp electrodes are positive, whereas neck and body electrodes are negative: the earlobe must therefore be near the boundary (Fig. 4).

Therefore, the scalp–NC lead registers a large P13/P14 far-field potential, whereas both scalp–ear and ear–NC leads register smaller P13/P14 potentials. Of course, this simply indicates that there is a dipole at the boundary and does not always imply that the dipole is a junctional

118

Fig. 4. P13/P14 in median nerve SEPs as a junctional potential. (Reproduced by permission of Seiwa Shoten Co. Ltd from Sonoo, M. (2003) Somatosensory evoked potentials (SEPs) (1): SEP montage and origin of each component. Brain Science (Tokyo), 25: 991–998.) Regarding P13/P14, it is supposed that the scalp forms a positive compartment while the neck and trunk forms a negative compartment. Therefore the scalp–NC lead registers a positive potential, P13/P14. The earlobe electrode (Ai) is near or slightly caudal to the boundary. Therefore, both the scalp–Ai and Ai–NC leads register positive potentials, the former being slightly larger. These results indicate the presence of dipolar static potential, which can be a junctional potential or a postsynaptic potential. We considered that the origin must be a junctional potential generated at the beginning of the lemniscal volley, considering the other conditions.

potential; it can be a postsynaptic potential generated at the cuneate nucleus. However, the sharp onset and rapid notches are unlikely to be a postsynaptic potential. The observations of preserved upper cervical N13, which we attributed to the postsynaptic potential at the cuneate nucleus (Sonoo et al., 1990a), and absent scalp P13/P14 in patients with medullary lesions (Mauguière et al., 1983; Sonoo et al., 1996b) also do not support a major contribution of the postsynaptic potential at the cuneate nucleus to scalp P13/P14. Another influential opinion attributed P13/P14 to a junctional potential, explaining that P13/P14 is generated at the cranio-vertebral junction between the narrow spinal canal and wide scalp (Lueders et al., 1983). We have previously refuted this hypothesis on the following grounds (Sonoo, 1999, 2000): first, a junctional potential makes sense only when the leading and trailing

dipole are generated on the same axon at the same time. At the cranio-vertebral junction, the two dipoles are generated on different neurons, separated in time intercalated by a synapse. Even if we ignore these factors, the wide scalp should become negative according to the junctional potential theory (Stegeman et al., 1987; Dumitru and Jewett, 1993), which is opposite to the actual polarity of P13/P14. In contrast, if we postulate that P13/P14 is a junctional potential generated at the beginning of the lemniscal volley, the observed polarity coincides with the simulation (Dumitru and Jewett, 1993). The last opinion to be mentioned on P13/P14 origin, which is relevant to clinical application, is to ascribe the origin of the earlier part of the P13/P14 complex, usually labeled as P13, to presynaptic dorsal column fibers or a high cervical cord. This hypothesis has been supported by the observation that P13 was preserved in patients with cervicomedullary junction (Mavroudakis et al., 1993; Restuccia et al., 1995; Valeriani et al., 1998) or in brain death (Anziska and Cracco, 1980; Wagner, 1991, 1996). We reported that such a P13-like potential is observed even in patients with a lesion at the C3/4 vertebral level, corresponding to the C5 spinal level. Therefore, this was named lcP13 (lower cervical P13) and its origin ascribed to the beginning of the second spinal ascending volley (Sonoo et al., 2001). We also stressed that lcP13 is not a direct counterpart of normal P13 as the first subcomponent of the P13/P14 complex, but a potential that becomes visible only when the whole P13/P14 complex is lost. It is especially important for the diagnosis of brain death to recognize lcP13 and to understand that it has a lower cervical origin. To use, ideally a nasopharyngeal electrode, but practically an earlobe electrode, as a reference electrode would cancel the lcP13 component and enable us to evaluate the P13/P14 of pure brainstem origin (Wagner, 1991, 1996; Restuccia et al., 1995; Sonoo et al., 2001). The diagnosis of brain death using SEPs has been often confusing because of the presence of lcP13 (Anziska and Cracco, 1980; Belsh and Chokroverty, 1987; Facco et al., 1990). We demonstrated that to use P13/P14 with an ear reference, together with the N18 ascribed to the cuneate nucleus (Sonoo et al., 1990b, 1991, 1992b, 1996b), would definitely

119 establish the usefulness of SEPs in the diagnosis of brain death (Sonoo et al., 1999; Hatanaka et al., 2003). Failure to recognize that lcP13 is of lower cervical origin would cause erroneous localization of the lesion site in patients with a high cervical lesion. True P13/P14 and the following cortical components may be lost or attenuated and delayed due to a high cervical lesion, but if lcP13 were erroneously interpreted as true P13/P14, the lesion would be localized intracranially. Thus, enhanced knowledge of SEP generators will increase the localization accuracy using SEPs and will broaden the possible clinical applications of SEPs. 5. Acknowledgements I would like to thank all the collaborators who contributed to the content of this review: Yuuki Hatanaka, Takaaki Miura, Yasunobu Tsai-Shozawa, Hiroshi Tsukamoto, Shingo Kawakami, Atsuko Mochizuki, and Professor Teruo Shimizu, Department of Neurology, Teikyo University School of Medicine; Hiroyuki Fukuda, Department of Medicine, University Hospital, Mizonokuchi, Teikyo University School of Medicine; and Akira Katayama and Professor Kiyoharu Inoue, The Jikei University School of Medicine. References Anziska, B.J. and Cracco, R.Q. (1980) Short latency somatosensory evoked potentials in brain dead patients. Arch. Neurol., 37: 222–225. Anziska, B.J. and Cracco, R.Q. (1981) Short latency SEPs to median nerve stimulation: comparison of recording methods and origin of components. Electroencephalogr. Clin. Neurophysiol., 52: 531–539. Belsh, J.M. and Chokroverty, S. (1987) Short-latency somatosensory evoked potentials in brain-dead patients. Electroencephalogr. Clin. Neurophysiol., 68: 75–78. Boden, S.D., McCowin, P.R., Davis, D.O., Dina, T.S., Mark, A.S. and Wiesel, S. (1990) Abnormal magnetic-resonance scans of the cervical spine in asymptomatic subjects. J. Bone Joint Surg. Am., 72: 1178–1184. Cracco, R.Q. and Cracco, J.B. (1976) Somatosensory evoked potential in man: far field potentials. Electroencephalogr. Clin. Neurophysiol., 41: 460–466. Desmedt, J.E. and Cheron, G. (1980) Central somatosensory conduction in man: neural generators and interpeak latencies of the far-field components recorded from neck and right or left scalp and earlobes. Electroencephalogr. Clin. Neurophysiol., 50: 382–403.

Desmedt, J.E. and Cheron, G. (1981) Prevertebral (oesophageal) recording of subcortical somatosensory evoked potentials in man: the spinal P13 component and the dual nature of the spinal generators. Electroencephalogr. Clin. Neurophysiol., 52: 257–275. Dumitru, D. and Jewett, D.L. (1993) Far-field potentials. Muscle Nerve, 16: 237–254. Facco, E., Casartelli Liviero, M., Munari, M., Toffoletto, F., Baratto, F. and Giron, G.P. (1990) Short latency evoked potentials: new criteria for brain death? J. Neurol. Neurosurg. Psychiatr., 53: 351–353. Fukuda, H., Sonoo, M., Kako, M. and Shimizu, T. (2006) The optimal method to determine the stimulus intensity for median nerve somatosensory evoked potentials. J. Clin. Neurophysiol., in press. Gray, F. and Alonso, J.M. (2002) Bacterial infections of the central nervous system. In: D.I. Graham and P.L. Lantos (Eds.), Greenfield’s Neuropathology, Vol. 2, 7th Ed. Arnold, London, pp. 151–193. Greenfield, J.G. (1963) Infectious diseases of the central nervous system. In: W. Boackwood, W.H. McMenemey, A. Meyer, R.M. Norman et al. (Eds.), Greenfield’s Neuropathology, 2nd Ed. Williams and Wilkins, Baltimore, pp. 164–181. Hatanaka, Y., Sonoo, M., Shimizu, T., Nagashima, H., Nakatani, T. and Kobayashi, K. (2003) Comparison of specificity and sensitivity for the diagnosis of brain death between evoked potentials and clinical brainstem reflexes. Clin. Electroencephalogr., 45: 717–724. Kimura, J., Ishida, T., Suzuki, S., Kudo, Y., Matsuoka, H. and Yamada, T. (1986) Far-field recording of the junctional potential generated by median nerve volleys at the wrist. Neurology, 36: 1451–1457. Lueders, H., Lesser, R., Hahn, J., Little, J. and Klem, G. (1983) Subcortical somatosensory evoked potentials to median nerve stimulation. Brain, 106: 341–372. Mauguière, F., Courjon, J. and Schott, B. (1983) Dissociation of early SEP components in unilateral traumatic section of the lower medulla. Ann. Neurol., 13: 309–313. Mavroudakis, N., Brunko, E., Delberghe, X. and Zegers de Beyl, D. (1993) Dissociation of P13-P14 far-field potentials: clinical and MRI correlation. Electroencephalogr. Clin. Neurophysiol., 88: 240–242. Miura, T., Sonoo, M. and Shimizu, T. (2003) Establishment of standard values for the latency, interval and amplitude parameters of tibial nerve somatosensory evoked potentials (SEPs). Clin. Neurophysiol., 114: 1367–1378. Nakanishi, T., Shimada, Y., Sakuta, M. and Toyokura, Y. (1978) The initial positive component of the scalp-recorded somatosensory evoked potential in normal subjects and in patients with neurological disorders. Electroencephalogr. Clin. Neurophysiol., 45: 26–34. Restuccia, D., Di Lazzaro, V.M., Valeriani, M., Conti, G., Tonali, P. and Mauguière, F. (1995) Origin and distribution of P13 and P14 far-field potentials after median nerve stimulation. Scalp, nasopharyngeal and neck recording in healthy subjects and in patients with cervical and cervico-medullary lesions. Electroencephalogr. Clin. Neurophysiol., 96: 371–384. Sonoo, M. (1996) P15 in tibial nerve SEP as a simple example of the junctional potential. In: J. Kimura and H. Shibasaki (Eds.), Recent Advances in Clinical Neurophysiology. Elsevier, Amsterdam, pp. 260–265.

120 Sonoo, M. (1999) How much has been solved regarding SEP generators? In: C. Barber, G.C. Celesia, I. Hashimoto and R. Kakigi (Eds.), Functional Neuroscience: Evoked Potentials and Magnetic Fields (EEG Suppl. 49). Elsevier Sciences BV, Amsterdam, pp. 47–51. Sonoo, M. (2000) Anatomic origin and clinical application of the widespread N18 potential in median nerve somatosensory evoked potentials. J. Clin. Neurophysiol., 17: 258–268. Sonoo, M. (2003) Somatosensory evoked potentials (SEPs) (1): SEP montage and origin of each component. Brain Science (Tokyo), 25: 991–998. Sonoo, M., Shimpo, T., Genba, K., Kunimoto, M. and Mannen, T. (1990a) Posterior cervical N13 in median nerve SEP has two components. Electroencephalogr. Clin. Neurophysiol., 77: 28–38. Sonoo, M., Shimpo, T., Genba, K., Masaki, T., Sakuta, M. and Mannen, T. (1990b) Origin of upper cervical N13 (ucN13) and widespread N18 in the median nerve SEP. Electroencephalogr. Clin. Neurophysiol., 76: 106P. Sonoo, M., Sakuta, M., Shimpo, T., Genba, K. and Mannen, T. (1991) Widespread N18 in median nerve SEP is preserved in a pontine lesion. Electroencephalogr. Clin. Neurophysiol., 80: 238–240. Sonoo, M., Genba, K., Iwatsubo, T. and Mannen, T. (1992a) P15 in tibial nerve SEPs as an example of the junctional potential. Electroencephalogr. Clin. Neurophysiol., 84: 486–491. Sonoo, M., Genba, K., Zai, W., Iwata, M., Mannen, T. and Kanazawa, I. (1992b) Origin of the widespread N18 in median nerve SEP. Electroencephalogr. Clin. Neurophysiol., 84: 418–425. Sonoo, M., Shimizu, T., Ugawa, Y. and Uesaka, Y. (1993) Clinical application to diseases including the immune-mediated neuropathy of the P15 potential in tibial nerve SEP. In: Annual Report of the Neuroimmunological Disorders. The Ministry of Health and Welfare of Japan, 1992, pp. 251–253. Sonoo, M., Kobayashi, M., Genba-Shimizu, K., Mannen, T. and Shimizu, T. (1996a) Detailed analysis of the latencies of median nerve SEP components: 1. selection of the best standard parameters and the establishment of the normal value. Electroencephalogr. Clin. Neurophysiol., 100: 319–331. Sonoo, M., Hagiwara, H., Motoyoshi, Y. and Shimizu, T. (1996b) Preserved widespread N18 and progressive loss of P13/14 of

median nerve SEPs in a patient with unilateral medial medullary syndrome. Electroencephalogr. Clin. Neurophysiol., 100: 488–492. Sonoo, M., Genba-Shimizu, K., Mannen, T. and Shimizu, T. (1997) Detailed analysis of the latencies of median nerve somatosensory evoked potential components, 2: analysis of subcomponents of the P13/14 and N20 potentials. Electroencephalogr. Clin. Neurophysiol., 104: 296–311. Sonoo, M., Tsai-Shozawa, Y., Aoki, M., Nakatani, T., Hatanaka, Y., Mochizuki, A., Sawada, M., Kobayashi, K. and Shimizu, T. (1999) N18 in median nerve SEPs: a new indicator of the medullary function useful for the diagnosis of brain death. J. Neurol. Neurosurg. Psychiatr., 67: 374–378. Sonoo, M., Mochizuki, A., Fukuda, H., Oosawa, Y., Iwata, M., Hatanaka, Y., Tsai-Shozawa, Y., Okano, M. and Shimizu, T. (2001) Lower cervical origin of the P13-like potential in median SEPs. J. Clin. Neurophysiol., 18: 185–190. Sonoo, M., Katayama, A., Miura, T., Shimizu, T. and Inoue, K. (2005) Tibial nerve SEPs localized the lesion site in a patient with early tabes dorsalis. Neurology, 64:1452–1454. Stegeman, D.F., Van Oosterom, A. and Colon, E.J. (1987) Far-field evoked potential components by a propagating generator: computational evidence. Electroencephalogr. Clin. Neurophysiol., 67: 176–187. Valeriani, M., Restuccia, D., Di Lazzaro, V., Le Pera, D., Barba, C. and Tonali, P. (1998) The scalp to earlobe montage as standard in routine SEP recording: comparison with the non-cephalic reference in patients with lesions of the upper cervical cord. Electroencephalogr. Clin. Neurophysiol., 108: 414–421. Wagner, W. (1991) SEP testing in deeply comatose and brain dead patients: the role of nasopharyngeal, scalp and earlobe derivations in recording the P14 potential. Electroencephalogr. Clin. Neurophysiol., 80: 352–363. Wagner, W. (1996) Scalp, earlobe and nasopharyngeal recordings of the median nerve somatosensory evoked P14 potential in coma and brain death: detailed latency and amplitude analysis in 181 patients. Brain, 119: 1507–1521. Yamada, T., Machida, M. and Kimura, J. (1982) Far-field somatosensory evoked potentials after stimulation of the tibial nerve. Neurology, 32: 1151–1158.


121

Chapter 17

Intraoperative recording of the very fast oscillatory activities evoked by median nerve stimulation in the human thalamus R. Hanajimaa,b,*, R. Chena, Peter Ashbya, A.M. Lozanoa,d, W.D. Hutchisona,c, K.D. Davisa,d and J.O. Dostrovskya,c a Toronto Western Research Institute, University Health Network, Toronto (Canada) Division of Neurology, Department of Medicine, University of Tokyo Hospital, Tokyo (Japan) c Department of Physiology, and d Division of Neurosurgery, Department of Surgery, University of Toronto, Toronto, ON M5T 2S8 (Canada) b

1. Introduction Recently, very high-frequency oscillatory activity (above 500 Hz), associated with somatosensory evoked potentials (SEPs) (very fast oscillations (VFOs)), has been revisited. VFOs at a frequency of 500–700 Hz are superimposed on scalp-recorded cortical potentials from the human sensory cortex (S1) (Curio et al., 1994; Hashimoto et al., 1996; Curio, 2000; Hashimoto, 2000). VFOs have also been observed as small notches superimposed on positive SEPs recorded with electrodes in the human thalamus (Katayama and Tsubokawa, 1987; Morioka et al., 1989; Klostermann et al., 1999). Although it has been hypothesized that thalamic VFOs originate within or near the sensory thalamus and are generated by different generators from those for the VFOs of S1 (Klostermann et al., 2000, 2002), this remains to be confirmed. The aim of this study was to

*Correspondence to: R. Hanajima, Department of Neurology, University of Tokyo Hospital, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Tel: +81-3-5800-8672; Fax: +81-3-5800-6548; E-mail: [email protected]

characterize thalamic VFOs and their relationship, if any, with thalamic neuronal firing. 2. Methods 2.1. Subjects We studied eight patients undergoing surgery for chronic deep brain stimulation (DBS) of the thalamus. In five patients, DBS macroelectrodes were implanted in the nucleus ventrointermedialis of the thalamus (Vim) (cerebellar thalamus) for the treatment of tremor, and three patients suffering chronic pain had DBS electrodes implanted in the nucleus ventrocaudalis of the thalamus (Vc) (sensory thalamus). All patients gave free and informed consent to the procedures approved by the University Health Network Research Ethics Board. 2.2. Surgical procedures The targets were identified by MRI and located stereotactically. The use of microelectrode recordings to confirm DBS electrode placement in the thalamus has

122 previously been described in detail (Lenz et al., 1988; Tasker and Kiss, 1995; Davis et al., 1998). For thalamic DBS placements, a tentative initial target was first selected at the hand sensory area of the ventrocaudal nucleus (Vc) about 15 mm lateral to the midline. Single and multiple unit neuronal recordings were made in both spontaneous and sensory evoked conditions in order to detect tactile neurons, kinesthetic or tremorrelated cells. Microstimulation was also performed at regular intervals and perceptual and/or motor effects were determined. This strategy allowed a physiological map to be constructed relative to the imaged coordinates. After the final target was determined by stimulating and recording with microelectrodes, a quadripolar DBS electrode (Model 3387, Medtronic, USA) was implanted. 2.3. Median nerve somatosensory evoked potentials (SEPs) from microelectrodes We recorded the field potentials and responses of single units elicited by median nerve stimulation from microelectrodes during the operation. The parylene-C-coated tungsten microelectrodes had an exposed tip size of 15–25 μm and the tips were plated with gold and platinum to reduce the impedance to about 0.1–0.5 MΩ at 1 kHz. Stimuli (a square-wave pulse with a duration of 0.2 ms and intensity 1.1–1.2 times the motor threshold) were administered to the median nerve at the wrist at 2 Hz. Responses were amplified and filtered between 10 Hz–5 kHz for the field potentials and 500 Hz–5 kHz for single unit responses (GS3000 system, Axon Instruments, Foster City, CA). Slow wave averages and post-stimulus time histograms (PSTHs) were calculated for 140–200 repetitions of the stimulus (at 2 Hz), using Spike2 software. In order to compare field potentials recorded with microelectrodes to the results of VFOs from macroelectrodes, we digitally changed the bandpass 500 Hz–2.5 kHz (Spike 2, Cambridge, UK) and then obtained the averaged field responses. The sampling rate was 12–15 kHz. In two patients, the averaged potentials were recorded at several successive depths 1 mm apart. We also recorded neuronal activity responses to median nerve stimulation. In Vc, we studied tactile cells,

which responded to tactile stimuli to the contralateral hand. Data were stored and analyzed off-line (Spike 2, Cambridge, UK). PSTHs triggered by median nerve stimulation were made from single unit data. We also analyzed the interspike intervals. 3. Results 3.1. Field potentials Twenty-four monopolar microelectrode recordings referenced to the ground were obtained in 7 thalamic patients. Figure 1 shows an example of the averaged field potentials recorded from a microelectrode at successive depths 1 mm apart in response to median nerve stimulation in one patient (0 mm is the level of the anterior commisure (AC)–posterior commisure (PC) line calculated with preoperative MRI). Single unit recordings revealed responses to tactile stimuli to the contralateral hand from 0 to 3 mm above the AC–PC line in this patient. At 0 mm, the cells responded to brushing of the second digit suggesting that the electrode was in the Vc. VFOs can be seen superimposed on the slow waves and became more obvious when the bandpass filter was changed from 500 to 5000 Hz. The VFOs of the evoked field potentials recorded from the microelectrode had the largest amplitude between −1 and +1 mm (Fig. 1A). When differences between SEPs recorded at successive sites were computed (Fig. 1B), phase reversal of VFOs occurred at the same position as the phase reversal of the slow components of SEPs (0 mm). The amplitudes of these potentials were the largest from sites close to the AC–PC line. We obtained a similar positive slow component of the SEP with VFOs around the AC–PC line from the other 23 recordings. 3.2. Unit recordings We recorded the activity of 7 single Vc neurons responding to median nerve stimulation in 6 patients. Figure 2 shows the PSTH of a single unit firing and the averaged field potential recorded by the microelectrode. Four peaks were detected in the PSTH and they tended to occur 1.0–1.6 ms apart. The VFO peaks shown in the

123

Fig. 1. (A) Monopolar recordings of field potentials from Vc at successive depths 1 mm apart in response to median nerve stimulation in one patient. The potential was most positive at 0 mm on the AC–PC line. When the bandpass filter was changed from 500 to 5000 Hz (A, right), VFOs similar to those observed from DBS electrodes were observed. Vertical lines indicate the latency of the positive wave of the VFOs at 0 mm. (B) When we computed the difference between adjacent traces, the slow component showed phase reversal at 0 mm (*). VFOs shown with the 500–5000 Hz bandpass also reversed their phases at this point (B right). The vertical lines indicate the latency of the negative wave peaks of the VFOs at −1 – 0 and asterisks indicate phase reversal. (Reproduced from Fig. 4 in Hanajima et al. (2004) J. Neurophysiol., 92: 3171–3182.)

averaged field potential appear to be closely related to single unit firing as can be seen in the PSTH. These PSTH peaks were not caused by repetitive firing of a single cell as can be seen in the raster plot (not shown), which shows that one cell usually fired only once per trial. The spikes in response to median nerve stimulation usually corresponded with one of the VFO peaks. Figure 3 shows an example of unit responses to median nerve stimulation with a brief burst of spikes (Fig. 3A). The PSTH of the unit discharge using all spikes aligned to median nerve stimulation indicates

multiple peaks with shorter intervals (0.4–1.2 ms) than intervals within the bursts, similar to those of peaks in the averaged field potential waveform. By contrast, the interspike interval histogram (Fig. 3B) showed that the shortest interval was around 2.6 ms. This interval was longer than those of VFO or PSTH peaks, indicating that VFOs cannot be accounted for by spike timing within the bursts. Similar VFOs were obtained from the 5 other single units. PSTH to median nerve stimulation also revealed multiple peaks with intervals between 0.4 and 1.6 ms (0.9 ± 0.25 ms: mean ± SD).

124

Fig. 2. The post-stimulus time histogram (PSTH) of the discharges of a thalamic (Vc) single unit and averaged wave following median nerve stimulation. The timing of unit firing after stimulus is shown. Several peaks were detected in the PSTH. The solid lines indicate the peak latencies. The peaks tended to occur at preferred times (1.0–1.6 ms apart: 16.0, 17.4, 18.4, 20.0 ms) related to the VFO peaks of the averaged wave (upper trace). (Reproduced from figures in Hanajima et al. (2004) J. Neurophysiol., 92: 3171.)

4. Discussion The recorded VFOs showed phase reversal at the AC–PC line and their amplitudes were largest near the AC–PC line. Klostermann et al. (2002) observed that the VFO amplitude with monopolar recordings decreased with the distance from the thalamus and there was no phase reversal 10–20 mm above the target position for Vc. They suggested that VFOs were generated in the thalamus, probably by thalamocortical projection neurons: our data provide direct support for this hypothesis. In addition, the microelectrode recordings showed that the VFOs were fairly constant in amplitude over a distance of about several micrometers, which corresponded to the thickness of Vc, but their amplitude dropped off rapidly above and below this region.

The most novel and instructive part of this study was the observation that tactile neuron firing was closely correlated with VFOs. Although the neurons do not fire at each successive oscillation, when they do fire, their action potentials are highly correlated with the oscillation although they may only fire once or twice for each stimulus. This observation suggests that VFOs may either determine the exact time of neuron firing in Vc or that Vc neuron firing is somehow time-locked in a repetitive pattern, giving rise to oscillatory VFO. One possible explanation is that VFOs are generated by action potentials comprising the stimulation response. However, this requires that neurons fire in a high-frequency burst following each stimulus and that these bursts are very similar in timing among different neurons so that summation occurs to give rise to widespread oscillatory LFPs. This mechanism is unlikely

125

Fig. 3. Intraoperative microelectrode recordings of a thalamic single unit that had a discharge burst following median nerve stimulation. In this case, a single stimulus generated a brief burst of spikes in this single unit. Two sweeps are shown in (A). (B) shows interspike interval time histograms of the spikes. The histogram indicates that the interval between the burst discharges was 2.6 ms or more. This is longer than the intervals of VFOs peaks. (Reproduced from Fig. 6 in Hanajima et al. (2004) J. Neurophysiol., 92: 3171–3182.)

since thalamic sensory neurons do not display such highfrequency bursts. In our recordings, the burst discharge frequency of neurons in Vc that responded in a burst was much lower. A second possibility is that some other and unknown mechanism generates the VFO and these LFP oscillations tend to synchronize neuron firing. This hypothesis then predicts that at least part of the oscillatory LFP results from membrane oscillations of the neurons in Vc. Perhaps the VFOs result from a high-frequency resonance phenomenon in the Vc region but it remains

a mystery how and why they are generated and synchronized. It has been hypothesized that gap junctions may play a role in high-frequency synchronization between neurons in the cortex and in the dorsal lateral geniculate nucleus of the thalamus (Hughes et al., 2002). Hughes et al. (2002) reported that thalamocortical neurons in the cat dorsal lateral geniculate nucleus were electrotonically coupled via gap junctions with spikelets representing attenuated action potentials from adjoining cells. However, even electrotonic conduction via gap junctions involves some delay and may not be fast enough (Traub et al., 1999) to mediate the very highfrequency oscillatory activity observed in the thalamus. Our findings are similar to those recently observed in animal studies in the somatosensory cortex. For example, Barth (2003) reported that neuronal responses to vibrissa stimulation are time-locked to VFO field potentials elicited by the same stimuli. The blockade of cortical inhibitory circuits by application of the GABA antagonist, bicuculline, fails to block these cortical VFOs, indicating that inhibitory interneurons are not involved (Jones and Barth, 2002). In the hippocampus there are also VFOs termed ripples, although their frequency is much lower (200–400 Hz) than those observed in the thalamus and they may be generated by a similar mechanism to those generated in the thalamus and cortex. It has been suggested that population ripples in the hippocampus may be generated by a mechanism involving axonal gap junctions (Traub et al., 1999). VFOs have been observed at several sites in the somatosensory system such as the cuneate nucleus (Rasmusson and Northgrave, 1997; Canedo et al., 1998), the thalamus and S1 (Curio et al., 1994; Hashimoto et al., 1996; Curio, 2000; Hashimoto, 2000) as well as in other systems. Although the oscillation frequency differs by region, the underlying mechanisms and function may be similar. VFOs have also been observed with natural stimuli (Baker et al., 2003; Barth, 2003) and thus appear to be a normal manifestation of brain function. The function and consequence of such VFOs are unknown, but may be to synchronize and perhaps prolong important input, thus increasing the signal to noise ratio.

126 5. Acknowledgements This work was supported by the Canadian Institutes of Health Research, grant nos. 85102 (PA), MOP15128 (RC), MOP-42505 (JOD), National Institutes of Health (NINDS) RO1 NS 40872 (JOD) and the Japan Society for the Promotion of Science (RH). I am grateful to Dr Ugawa, Department of Neurology, University of Tokyo for his helpful comments on this chapter. This chapter is a summarized part of our previously published article (Hanajima et al., 2004). References Barth, D.S. (2003) Submillisecond synchronization of fast electrical oscillations in neocortex. J. Neurosci., 23: 2502–2510. Baker, S.N., Curio, G. and Lemon, R.N. (2003) EEG oscillation at 600 Hz are macroscopic markers for cortical spike bursts. J. Physiol., 550: 529–534. Canedo, A., Martinez, L. and Marino, J. (1998) Tonic and bursting activity in the cuneate of the chloralose-anesthetized cat. Neuroscience, 84: 603–607. Curio, G. (2000) Linking 600 Hz “spike-like” EEG/MEG wavelet (‘sigmabursts’) to cellular substrates: concepts and caveats. J. Clin. Neurophysiol., 17: 377–396. Curio, G., Mackert, B.-M., Burghoff, M., Koeitz, R., AbrahamFuchs, K. and Härer, K.W. (1994) 600 Hz activity in the cerebral comatosensory system. Electroencephalogr. Clin. Neurophysiol., 91: 483–487. Davis, K.D., Lozano, A.M., Tasker, R.R. and Dostrovsky, J.O. (1998) Brain target for pain control. Stereotact. Funct. Neurosurg., 71: 173–179. Hanajima, R., Chen, R., Ashby, P., Lozano, A.M., Hutchinson, W.D., Davis, K.D. and Dostrovsky, J.O. (2004) Very fast oscillation evoked by median nerve stimulation in the human thalamus and subthalamus nucleus. J. Neurophysiol., 92: 3171–3182. Hashimoto, I. (2000) High-frequency oscillations of somatosensory evoked potentials and fields. J. Clin. Neurophysiol., 17: 309–320.

Hashimoto, I., Mashiko, T. and Imada, T. (1996) Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex. Electroencephalogr. Clin. Neurophysiol., 100: 189–203. Hughes, S.W., Blethyn, K.L., Cope, D.W. and Crunelli, V. (2002) Properties and origin of spikelets in thalamocortical neurones in vitro. Neuroscience, 110: 395–401. Jones, M.S. and Barth, D.S. (2002) Effects of bicuculline methiodide on fast (> 200 Hz) electrical oscillations in rat somatosensory cortex. J. Neurophysiol., 88(2): 1016–1025. Katayama, Y. and Tsubokawa, T. (1987) Somatosensory evoked potentials from the thalamic sensory relay nucleus (VPL) in humans: correlations with short latency somatosensory evoked potentials recorded at the scalp. Electroencephalogr. Clin. Neurophysiol., 68: 187–201. Klostermann, F., Funk, T., Vesper, J. and Curio, G. (1999) Spatiotemporal characteristics of human intrathalamic high frequency (> 400 Hz) SEP components. Neuroreport, 10: 3627–3631. Klostermann, F., Funk, T., Vesper, J., Siedenberg, R. and Curio, G., (2000) Double-pulse stimulation dissociates intrathalamic and cortical high-frequency (> 400 Hz) SEP components in man. Neuroreport, 11: 1295–1299. Klostermann, F., Gobbele, R., Buchner, H. and Curio, G. (2002) Intrathalamic non-propagating generators of high-frequency (1000 Hz) somatosensory evoked potential (SEP) bursts recorded subcortically in man. Clin. Neurophysiol., 113: 1001–1005. Lenz, F.A., Dostrovsky, J.O., Kwan, H.C. and Tasker, R.R., (1988) Methods for microstimulation and recording of single neurons and evoked potentials in the human ventral nervous system. J. Neurosurg., 68: 630–634. Morioka, T., Shima, F., Kato, M. and Fukui, M. (1989) Origin and distribution of thalamic somatosensory evoked potentials in humans. Electroencephalogr. Clin. Neurophysiol., 74: 86–193. Rasmusson, D.D. and Northgrave, S.A. (1997) Reorganization of the racoon cuneate nucleus after peripheral denervation. J. Neurophysiol., 78: 2924–2934. Tasker, R.R. and Kiss, Z.H.T. (1995) The role of the thalamus in functional neurosurgery. Neurosurg. Clin. N. Am., 6: 73–104. Traub, R.D., Schmitz, D., Jefferys, J.G. and Draguhn, A. (1999) High-frequency population oscillations are predicted to occur in hippocampal pyramidal neuronal networks interconnected by axoaxonal gap junctions. Neuroscience, 92: 407–426.


127

Chapter 18

Cortical processing of noxious information in humans: a magnetoencephalographic study Koji Inui*, Xiaohong Wang, Yunhai Qiu, Takeshi Tsuji, Hiroki Nakata and Ryusuke Kakigi Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585 (Japan)

1. Introduction Recent functional imaging studies in humans have provided evidence that multiple regions of the brain are involved in pain perception, including the primary (SI) and secondary (SII) somatosensory cortices, insula and anterior cingulate cortex (for review, see Bushnell et al., 1999; Schnitzler and Ploner, 2000). In support of the involvement of these regions in pain perception, neurophysiological studies in monkeys have demonstrated nociceptive neurons in SI (Biedenbach et al., 1979; Kenshalo and Isensee, 1983), SII (Robinson and Burton, 1980; Dong et al., 1994) and the insula (Robinson and Burton, 1980; Dostrovsky and Craig, 1996; Zhang et al., 1999). However, the precise temporal sequence of cortical activation is not well understood, especially in humans. In this study, we sought to clarify the timing of multiple cortical activities following noxious stimulation using magnetoencephalography (MEG). MEG has

*Correspondence to: Koji Inui, M.D., Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585, Japan. Tel: +81 564 55 7814; Fax: +81 564 52 7913; E-mail: [email protected]

an advantage over imaging modalities in that it can provide temporal information about an activity in addition to its location. 2. Methods The experiment was performed on 12 healthy male volunteers, aged 28–42 (mean 33.8 ± 4.3) years. The study was approved in advance by the Ethical Committee of the National Institute for Physiological Sciences and written consent was obtained from all subjects. 2.1. Noxious stimulation For noxious stimuli, we used intra-epidermal electrical stimulation, a method that we recently developed (Inui et al., 2002a). A pushpin-type needle electrode with a needle tip 0.2 mm in length was used. By pressing the electrode plate gently against the skin, the needle tip was inserted adjacent to the nerve endings of the thin myelinated fibers in the epidermis and superficial part of the dermis. A surface electrode, 1.0 cm in diameter, was placed on the skin at a distance of 4 cm from the needle electrode as the anode. The electric stimulus was a current constant square wave

128 pulse delivered at random intervals of 0.1–0.3 Hz. The stimulus duration was 0.5 ms. The current intensity (0.19 ± 0.07 mA) was the level producing a definite pain sensation in each subject, which was determined prior to the experiment. The stimulated site was the dorsum of the left hand between the first and second metacarpal bones. By using this method, we could selectively stimulate cutaneous Aδ fibers (Inui et al., 2002a, b).

A

B

2.2. SEF recording and analysis The somatosensory evoked magnetic fields (SEFs) were measured using dual 37-channel axial-type firstorder biomagnetometers (Magnes, Biomagnetic Technologies, San Diego, CA), as described elsewhere (Kakigi et al., 2000). The magnetic fields were recorded with a 0.1–200 Hz filter at a sampling rate of 2083 Hz and then filtered at low pass, 100 Hz. The analysis window was 100 ms pre- to 400 ms post-stimulus. A hundred responses were collected and averaged. Since several cortical activities overlapped temporally following painful stimulation (Watanabe et al., 1998), we used a multiple source model instead of a single equivalent current dipole (ECD) model. We used a brain electric source analysis (BESA) software package (NeuroScan, Inc, McLean, VA, USA) to analyze theoretical multiple source generators. The goodness of fit (GOF) indicated the percentage of data that can be explained by the model. We used the GOF value to determine whether the model was appropriate. To identify the best location of a source, movements were made in steps of 0.5–2.0 mm and the GOF was calculated at each location. We repeated this procedure until the largest GOF was obtained. 3. Results A consistent magnetic field (termed 1M) was identified following epidermal stimulation (ES) in the hemisphere contralateral to the stimulated side (contralateral hemisphere) in all 12 subjects and in the hemisphere ipsilateral to the stimulated side (ipsilateral hemisphere) in 11 subjects (Fig. 1). Its peak latency was 149.0 ± 11.1 ms for the contralateral hemisphere

Fig. 1. Evoked magnetic fields following painful stimulation of the dorsum of the left hand. (A) Superimposed waveforms recorded from 37 channels in the hemisphere contralateral to the stimulation following epidermal stimulation in subject 1. (B) Isocontour map at the peak latency (shown by an arrow in (A)) of 1M. Note the complicated topography indicating multiple sources.

and 166.4 ± 13.1 ms for the ipsilateral hemisphere (1M (I)). The latency difference between the hemispheres was 17.4 ms and significant (p < 0.0001). The topography at the peak of 1M (Fig. 1B) often indicated the presence of multiple source activities at this latency point. 3.1. Procedure of BESA analysis First, a latency range of about 10–160 ms was chosen as the period of ES evoked magnetic fields because one or two later activities emerged around the peak latency of 1M, as described below. We started the analysis with one source (source 0) placed around the Sylvian fissure since this area has been reported as a major source responsible for 1M evoked by noxious stimulation (Kakigi et al., 1995). As shown in Fig. 2A, the most effective source was usually the upper bank

129 A

D

E

B

C

Fig. 2. Procedures for multiple source analysis. Brain electric source analysis (BESA) for epidermal stimulation (ES) evoked magnetic fields recorded from the hemisphere contralateral to the stimulation in subject 1. Traces show temporal profiles of source strength in each step of analysis. Bars in schematic drawings of the source location indicate directions of upward deflection of the waveform. (A)–(D) Results for the one-, two-, three-, and four-source model, respectively. (E) Location of source generators overlaid on MRI scans. Sources 1–4 correspond to the insular cortex, secondary somatosensory cortex (SII), primary somatosensory cortex (SI) and amygdala/hippocampal formations(medial temporal (MT) area), respectively. GOF – goodness of fit. (Adopted from Inui et al., 2003.)

or bottom of the Sylvian fissure, supporting previous reports including our own. However, this source could not explain the magnetic fields during the analysis period (mean GOF = 65.2%) as expected from the complicated topography of 1M in Fig. 1. We therefore then tested a pair of sources placed around the Sylvian fissure that explained the magnetic field most effectively. In most cases, one source (source 2) was located in the upper bank of the Sylvian fissure and the other (source 1) near the insular circular sulcus (Fig. 2B). When the GOF did not exceed 90% with the twosource model, we placed one more source (source 3) around the central sulcus (Fig. 2C) since recent MEG studies have revealed that 1M contains an activity

originating from SI (Ploner et al., 1999; Kanda et al., 2000; Inui et al., 2002b). The two- or three-source model provided a GOF value of more than 90% in all subjects, and the location and orientation of the sources were fixed. The period of analysis was then expanded to 10–400 ms (whole recording), and one or two sources were added, if necessary, to obtain a GOF larger than 90% (Fig. 2D). We placed sources in the medial temporal (MT) area, around the amygdalar nuclei or hippocampal formation and at the cingulate cortex because both areas were considered to contribute to pain perception (Talbot et al., 1991; Watanabe et al., 1998). After this process, a three- to five-source model was obtained in each subject and

130 TABLE 1 PEAK LATENCIES OF CORTICAL RESPONSES (in ms) SI

Contralateral Ipsilateral

Insula

Early

Late

93.9

160.8

147.8 164.5

the results were used for further analysis. The actual location of the sources was confirmed by overlaying on MR images (Fig. 2E). In all subjects, SII and insular sources were identified in the contralateral hemisphere. The peak latency was 152.2 and 147.8 ms, respectively (Table 1), and corresponded approximately to that of 1M (149.0 ms). These sources were therefore considered major components of 1M. Activity in the contralateral SI was identified in all subjects. The time course of SI activity was complicated compared with the insula and SII activities, i.e. SI activity consisted of 1–3 brief components followed by a relatively long component (Fig. 3). However, the first early SI component was very small in amplitude and was identified in only three subjects. We therefore used the second brief component (early SI, n = 9) and the later component (late SI, n = 12) for analysis. The peak latency of the early SI component was 93.9 ms, which slightly preceded the onset of SII activity (98.3 ms) but was slightly delayed compared to the onset of insula activity (90.9 ms). The peak latency of the late SI component was 160.8 ms; therefore, the late SI activity also contributed to 1M production. From the recordings obtained from the ipsilateral hemisphere, insula and SII activity was identified in ten subjects. The peak latency of SII (170.5 ms) and the insula (164.5 ms) approximately corresponded to that of 1M (I) (166.4 ms), indicating that these activities were major components of 1M (I). The peak latency of the ipsilateral SII activity was 18.7 ms later than the contralateral SII activity ( p = 0.001). Similarly, the ipsilateral insular activity peaked 16.5 ms later than the contralateral response ( p = 0.0016).

SII

152.2 170.5

MT

Cingulate

Early

Late

184.0 196.5

289.7 313.3

208.5 206.5

For magnetic fields later than 1M, a source in the MT area was necessary in eleven out of twelve subjects. Activity in the ipsilateral hemisphere was identified in eleven subjects, but in only three subjects in the contralateral hemisphere. The activity in this area always consisted of two peaks in opposite directions. The onset latency of this activity (ipsilateral hemisphere) was 155.6 ± 16.7 ms, near the peak latency of insula and SII activity. As another source of magnetic fields later than 1M, the cingulate cortex was identified in five subjects, the contralateral hemisphere in three, the ipsilateral hemisphere in one, and both hemispheres in one. This ECD always pointed laterally and the vector of superior–inferior orientation was very small (Fig. 3). The onset and peak latencies were almost identical for the first component of the MT source (Table 1). ECDs for this activity were localized to the anterior part of the cingulate cortex (Fig. 3). 4. Discussion Results in experimental animals (Biedenbach et al., 1979; Kenshalo and Isensee, 1983) and human imaging studies (Bushnell et al., 1999) clearly indicate that SI is involved in nociception. In this study, early SI activity clearly showed a shorter response latency than other source activities. Only a few MEG studies have shown SI activation following painful stimulation. Ploner et al. (1999) demonstrated the simultaneous activation of SI and SII by painful laser stimulation for the first time, indicating a parallel thalamocortical distribution of nociceptive information, which was confirmed later by Kanda et al. (2000) and Inui et al. (2002b). Anatomical data from experimental animals

131

Fig. 3. Temporal profile of cortical activities following painful epidermal stimulation. Cortical responses to epidermal stimulation in subject 2. The upper three traces are superimposed waveforms recorded from 37 channels in both hemispheres. The lower seven traces are temporal profiles of each source strength. Filled circles indicate a group of early SI activities. Arrowheads indicate the peak latency of early and late SI activity analyzed in this study. Right: locations of source generators overlaid on MRI scans. (Adopted from Inui et al., 2003.)

132 also supported the parallel thalamocortical distribution of nociceptive information (Kenshalo et al., 1980; Friedman and Murray, 1986). However, the SI activity mentioned above corresponded to our late SI activity, i.e. MEG studies observed late SI activity but did not identify early components. Previous studies probably failed to identify early SI activities because laser stimulation takes a relatively long time to activate nociceptors because of temperature conduction and this causes latency jittering of the response. Since early SI activities consisted of several brief components oriented in opposite directions, 10–20 ms of latency jittering easily canceled out these activities after averaging the trials. Our method, in contrast, uses electrical stimulation and therefore provides constant responses in terms of latency. ES activated the anterior/mid-part of the insular cortex. The anterior location of pain-related activation in the insula was consistent with most functional imaging studies (for review, see Schnitzler and Ploner, 2000). Unitary recordings in monkeys have provided evidence of nociceptive neurons in the insula (Robinson and Burton, 1980; Dostrovsky and Craig, 1996). The failure to detect insular activities in previous MEG studies is probably due to the similar time course and the proximity of insular and SII activity. As the insula is located deeper than SII and recorded magnetic fields from the insula are weaker than those from SII, insula activity may be buried in those from SII when we analyze data using a single dipole model. Our results showed that noxious cutaneous stimuli produced SII activation as consistently demonstrated by brain imaging studies. The latency of the SII response coincided with the results of previous MEG studies (Kakigi et al., 1995; Ploner et al., 1999). While this study could not clarify whether there is hierarchical activation in SI and SII in pain processing, our data showing sequential activation in these areas favor a serial mode rather than a parallel mode of pain processing in SI and SII. This notion is consistent with anatomical findings in monkeys that SI receives projections from the lateral thalamic nuclei and, in turn, sends fibers to SII. For middle–late components of ES evoked magnetic responses, we found anterior cingulate cortex activity

in five subjects, in which recent imaging studies have consistently found pain-related activity (for review, see Schnitzler and Ploner, 2000). We identified the MT, including the amygdala nuclei and hippocampus as another source of activity for middle–late ES evoked responses. Since limbic structure activity emerged later than SI, SII and insular activity, and since corresponding pain-related vertex potentials were modulated by arousal and attentional levels in previous studies, these limbic structures are considered involved in the attentional and emotional aspect of nociception. 4.1. Temporal sequence of cortical activity Our findings that neither onset nor peak latencies differed between SII and the insula indicated parallel processing. Our data showing that ES evoked SII activity appeared 7.4 ms later than insular activity also indicated an origin other than SII (Friedman et al., 1986) for insula activation, including the thalamus (Friedman and Murray, 1986) and SI (Mufson and Mesulam, 1982). As activation in the anterior/mid-insula is related to noxious stimuli in imaging studies, it seems possible that the insula receives input from modalityspecific neurons in the thalamus. For example, Craig et al. (1994) demonstrated a very high concentration (97%) of pain- and thermo-specific neurons in the posterior part of the ventral medial thalamic nucleus (VMpo), which has dense lamina I spinothalamic tract terminations. VMpo projects to the insula (Craig et al., 1994) and the dorsal anterior region of the insula in monkeys contains a concentrated number of nociceptive-specific neurons (Dostrovsky and Craig, 1996). The insula has been shown to project to limbic structures including the amygdaloid complex (Mesulam and Mufson, 1982; Friedman et al., 1986) and cingulate cortex (Vogt and Pandya, 1987). Our results revealed that onset latencies of MT and the cingulate cortex approximately corresponded to the peak latencies of insular activity, so it is possible that both MT and the cingulate cortex were driven by the insula. As the anterior insula sends fibers directly to the amygdaloid complex and the hippocampal formation indirectly via the perirhinal cortex (Friedman et al., 1986), this corticolimbic pathway may play a role in pain recognition or

133 emotional reactions to noxious events (LeDoux, 1995). Our results therefore suggested two parallel pathways of pain processing, one through the lateral thalamic nuclei, SI, and SII serially and another through the medial thalamic nuclei, insula, and limbic structures serially. These two distinct pathways seem to correspond to the classic dichotomy of pain processing, the lateral and medial systems, which are involved in discriminative and emotional aspects of pain, respectively. References Biedenbach, M.A., Van Hassel, H.J. and Brown, A.C. (1979) Tooth pulp-driven neurons in somatosensory cortex of primates: role in pain mechanisms including a review of the literature. Pain, 7: 31–50. Bushnell, M.C., Duncan, G.H., Hofbauer, R.K., Ha, B., Chen, J.I. and Carrier, B. (1999) Pain perception: is there a role for primary somatosensory cortex? Proc. Natl. Acad. Sci. USA, 96: 7705–7709. Craig, A.D., Bushnell, M.C., Zhang, E.T. and Blomqvist, A. (1994) A thalamic nucleus specific for pain and temperature sensation. Nature, 372: 770–773. Dong, W.K., Chudler, E.H., Sugiyama, K., Roberts, V.J. and Hayashi, T. (1994) Somatosensory, multisensory, and task-related neurons in cortical area 7b (PF) of unanesthetized monkeys. J. Neurophysiol., 72: 542–564. Dostrovsky, J.O. and Craig, A.D. (1996) Nociceptive neurons in primate insular cortex. Soc. Neurosci. Abstr., 22: 111. Friedman, D.P. and Murray, E.A. (1986) Thalamic connectivity of the second somatosensory area and neighboring somatosensory fields of the lateral sulcus of the macaque. J. Comp. Neurol., 252: 348–373. Friedman, D.P., Murray, E.A., O’Neill, B. and Mishkin, M. (1986) Cortical connections of the somatosensory fields of the lateral sulcus of macaques: evidence for a corticolimbic pathway for touch. J. Comp. Neurol., 252: 323–347. Inui, K., Tran, D.T., Hoshiyama, M. and Kakigi, R. (2002a) Preferential stimulation of Aδ fibers by intra-epidermal needle electrode in humans. Pain, 96: 247–252. Inui, K., Tran, D.T., Qiu, Y., Wang, X., Hoshiyama, M. and Kakigi, R. (2002b) Pain-related magnetic fields evoked by intra-epidermal electrical stimulation in humans. Clin. Neurophysiol., 113: 298–304. Inui, K., Tran, T.D., Qiu, Y., Wang, X., Hoshiyama, M. and Kakigi, R. (2003) A comparative magnetoencephalographic study of

cortical activations evoked by noxious and innocuous somatosensory stimulations. Neuroscience, 120: 235–248. Kakigi, R., Koyama, S., Hoshiyama, M., Kitamura,Y., Shimojo, M. and Watanabe, S. (1995) Pain-related magnetic fields following CO2 laser stimulation in man. Neurosci. Lett., 192: 45–48. Kakigi, R., Hoshiyama, M., Shimojo, M., Naka, D., Yamasaki, H., Watanabe, S., Xiang, J., Maeda, K., Lam, K., Itomi, K. and Nakamura, A. (2000) The somatosensory evoked magnetic fields. Prog. Neurobiol., 61: 495–523. Kanda, M., Nagamine, T., Ikeda, A., Ohara, S., Kunieda, T., Fujiwara, N., Yazawa, S., Sawamoto, N., Matsumoto, R., Taki, W. and Shibasaki, H. (2000) Primary somatosensory cortex is actively involved in pain processing in human. Brain Res., 853: 282–289. Kenshalo, Jr. D.R. and Isensee, O. (1983) Responses of primate SI cortical neurons to noxious stimuli. J. Neurophysiol., 50: 1479–1496. Kenshalo, Jr. D.R., Giesler, Jr. G. J., Leonard, R. B. and Willis, W. D. (1980) Responses of neurons in primate ventral posterior lateral nucleus to noxious stimuli. J. Neurophysiol., 43: 1594–1614. LeDoux, J.E. (1995) Emotion: clues from the brain. Annu. Rev. Psychol., 46: 209–235. Mesulam, M.M. and Mufson, E.J. (1982) Insula of the old world monkey. III: Efferent cortical output and comments on function. J. Comp. Neurol., 212: 38–52. Mufson, E.J. and Mesulam, M.M. (1982) Insula of old world monkey. II: Afferent cortical input and comments on the claustrum. J. Comp. Neurol., 212: 23–37. Ploner, M., Schmitz, F., Freund, H.J. and Schnitzler, A. (1999) Parallel activation of primary and secondary somatosensory cortices in human pain processing. J. Neurophysiol., 81: 3100–3104. Robinson, C.J. and Burton, H. (1980) Organization of somatosensory receptive fields in cortical areas 7b, retroinsula, postauditory and granular insula of M. fascicularis. J. Comp. Neurol., 192: 93–108. Schnitzler, A. and Ploner, M. (2000) Neurophysiology and functional neuroamatomy of pain perception. J. Clin. Neurophysiol., 17: 592–603. Talbot, J.D., Marrett, S., Evans, A.C., Meyer, E., Bushnell, M.C. and Duncan, G.H. (1991) Multiple representations of pain in human cerebral cortex. Science, 251: 1355–1358. Vogt, B.A. and Pandya, D.N. (1987) Cingulate cortex of the rhesus monkey: II. cortical afferents. J. Comp. Neurol., 262: 271–289. Watanabe, S., Kakigi, R., Koyama, S., Hoshiyama, M. and Kaneoke, Y. (1998) Pain processing traced by magnetoencephalography in the human brain. Brain Topogr., 10: 255–264. Zhang, Z.H., Dougherty, P.M. and Oppenheimer, S.M. (1999) Monkey insular cortex neurons respond to baroreceptive and somatosensory convergent inputs. Neuroscience, 94: 351–360.


135

Chapter 19

Mechanism of voluntary and involuntary movements in humans Sadatoshi Tsuji*, Takenori Uozumi, Naoki Akamatsu, Akira Tamagawa, Kaoru Matsunaga, Hiroshi Ishiguchi, Tomoko Hashimoto and Yuki Kojima Department of Neurology, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555 (Japan)

1. Introduction Transcranial magnetic stimulation (TMS) is a noninvasive technique for activating the brain through the skull. It has been widely used for diagnostic and scientific purposes to evaluate the neurophysiological characteristics of sensorimotor cortex, cerebellum, upper and lower motor neurons, and higher brain functions. We report here two investigations. One is to elucidate the neurophysiological functions of primary negative motor area (area 44), in motor control such as voluntary movements, studied by single-pulse TMS and subdural electrical stimulation. The other is the relationship between cortical oscillatory activity and voluntary and involuntary hand movements in humans. 2. How are Broca’s area, area 44, negative motor area, and mirror neuron system related to hand motor control in humans? In order to elucidate whether area 44 is essential for the organization of voluntary hand movements, we *Correspondence to: Prof. Sadatoshi Tsuji, Department of Neurology, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka, Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555, Japan. Tel: 81-93-691-7438; Fax: 81-93-693-9842; E-mail: [email protected]

examined the effects of single-pulse TMS of area 44 on voluntary hand movements and electromyographic (EMG) activity in hand muscles (Uozumi et al., 2004). The Broca’s area in humans corresponds to the left 44 and 45 Brodmann areas, which are located in the caudal part of the inferior frontal gyrus. Many neurologists consider only Broca’s area to be related to the speech function, because it has long been known that stimulation of Broca’s area induces speech disturbances such as slow speech. In addition, destructive lesions in this area have revealed an oral expressive type of aphasia. We found that magnetic and electrical stimulation of area 44 produced motor evoked potentials (MEPs) from the hand muscles, but did not produce MEPs from the lower extremities. It is concluded that area 44 has direct fast-conducting corticospinal projections (Uozumi et al., 2004). It is not surprising that electrical stimulation of primary motor area (M1) with a stimulus intensity of 8 mA induced MEPs, followed by a silent period from the thenar muscles during tonic contraction. Furthermore, clear MEPs followed by silent periods were also produced by electrical stimulation of the primary negative motor areas (PNMAs) through subdural electrodes (Table 1). The PNMA in humans is located in the lateral frontal area just rostral to the primary tongue motor area which was stimulated

136 TABLE 1 NEUROPHYSIOLOGICAL CHARACTERISTICS OF PRIMARY MOTOR AREA (PMA) AND PRIMARY NEGATIVE MOTOR AREA (PNMA) IN RESPONSE TO STIMULATION BY SUBDURAL ELECTRODES PMA

PNMA (area 44)

Repetitive electrical stimulation (50 Hz)

MEP long silent period Clonic convulsion

Motor distribution

Homunculus

Small MEP short silent period 1. Disturbance of voluntary movements 2. Tonic muscle contraction 3. Muscle weakness Distal muscle dominant

Single electrical stimulation

through the subdural electrode. These subdural electrodes were implanted for clinical evaluation prior to the surgical treatment of intractable epilepsy. The Broca’s area overlaps the PNMA in the dominant hemisphere. These areas may play an important role in the organization and integration of fine hand motor movements. We noted that repetitive 50 Hz electrical stimulation of area 44 in the patients with intractable epilepsy induced disturbance of voluntary movements, tonic muscle contraction, and muscle weakness, which were similar to motor apraxia (Table 1). 2.1. Neurophysiological functions of the mirror neuron system Recent studies suggest the presence of a mirror neuron system in the human brains. In monkey, mirror neurons are found in the ventral premotor area (F5). This area responds to both observed and executed actions (Di Pellegrino et al., 1992; Gallese et al., 1996). Some indirect physiological evidence from PET (Rizzolatti et al., 1996), functional MRI (Iacoboni et al., 1999), and event-related MEG (Nishitani and Hari, 2000) studies indicated the existence of a mirror neuron system in the human brain. In addition, some recent studies (Rizzolatti et al., 1996; Iacoboni et al., 1999; Nishitani and Hari, 2000) have suggested an homology between the PNMA and the mirror neuron system. The activity of mirror neurons preceded those of M1 by ∼200–300 ms. Furthermore, the activation

of mirror neurons were noted during the imitation, for inter-individual communication, of oral-facial, and hand gestures (Nishitani and Hari, 2000). From these findings, the mirror neuron system has been thought to match observation and execution of the actions, and could possibly be located in area 44 of the human brain. The amplitudes of MEPs by TMS of the left area 44 were facilitated by simultaneous oral counting and imitation by a gesture of finger count, in comparison with those amplitudes elicited by finger tapping, observation, or only imitation of gesture. 2.2. Neurophysiological characteristics and MEPs in the negative motor areas PNMA probably plays important roles in the organization and integration of fine motor movements. It is impossible to continue muscle contraction and serial voluntary movements in hands during stimulation of PNMA (Lüders et al., 1987). There are two negative motor areas in humans, namely, the PNMA and the supplementary negative motor area. We focus on the PNMA in the present communication. The PNMA is identified in the lateral frontal area just rostral to the primary tongue motor area in humans where it is called area 44, and located in the anterior ascending ramus of sylvian fissure. We studied the effects of TMS on area 44. The stimulating site of area 44 is approximately 11 cm lateral

137 and 3 cm anterior to the vertex in human subjects. Otherwise, area 44 is located 6 cm lateral from primary hand motor area and 2–3 cm anterior to the primary tongue motor area. There were clear motor evoked responses of the right thenar muscles followed by long silent periods after TMS of the left primary hand motor areas (M1) with an onset latency of 22.8 ms. On the other hand, MEPs followed by short silent periods after TMS of area 44 were also recorded clearly from the thenar muscles during voluntary contraction, and this onset latency was 22.8 ms. It is interesting that the onset latency of MEPs after TMS of area 44 was the same as that after TMS of M1. However, the amplitudes of MEPs after TMS of area 44 were distinctly smaller than those after TMS of M1. There were no responses after TMS of the midpoint between M1 and area 44. TMS of Broca’s area (left area 44) induced MEPs followed by short silent periods from the right thenar

muscles (Fig. 1). The motor threshold of left area 44 was higher than that of M1 area. The silent period after TMS of left area 44 was distinctly shorter than that by M1 stimulation, and the amplitude was also significantly smaller than elicited by M1 stimulation. However, the onset latencies were similar in both MEPs. These data support the view that human area 44 has direct fast-conducting corticospinal projections (Uozumi et al., 2004). Ipsilateral motor evoked responses with an onset latency of 30.0 ms were recorded from the left thenar muscles after TMS of the left area 44. On the other hand, the onset latency of the contralateral MEP was 22.8 ms. The onset latency of the ipsilateral MEP is 7.2 ms longer than that of the contralateral MEP. In contrast, such a large ipsilateral response could never be elicited by TMS of M1 in adult human subjects. These results suggest the presence of trans-callosal projections between left and right area 44.

Fig. 1. Motor evoked potentials (MEPs) by transcranial magnetic stimulation (TMS) of Broca’s area (left area 44) and primary motor area (M1).

138 2.3. Negative effects after single-pulse TMS over area 44 during drawing a swirl After TMS over area 44, subjects immediately had slowness and clumsiness of fine finger movements (Uozumi et al., 2004). However, there were no clinical deficits when rotating the wrist or moving the elbow. TMS of the left area 44 interfered with the accuracy of fine movements of the right hand while drawing about half a circle after stimulation. In contrast, TMS over M1 showed a short-lived interference with drawing a swirl due to muscle contraction. Right area 44 magnetic stimulation did not show these effects. 2.4. Influences of single-pulse TMS of area 44 on EMG activity of bilateral thenar muscles during voluntary tonic and phasic contraction To evaluate the electrophysiological effects of TMS over area 44 on voluntary finger movements, changes in EMG activity were recorded from the bilateral thenar muscles. There were two stages in changes of EMG activity in the contralateral hand after the stimulation of area 44 during tonic contraction, namely, early inhibition and late excitation lasting for more than 200 ms. Late excitation is also observed in the ipsilateral thenar muscles. TMS of area 44 immediately before the execution of finger actions during voluntary phasic contraction such as finger tapping brought EMG bursts of low amplitude and long duration, and interruption of rhythmic finger tapping. From these results, it appears that human area 44 has inhibitory and facilitatory influence over tonic and phasic finger movements (Uozumi et al., 2004). 2.5. Clinical application of TMS of area 44 in patients with neurological disease TMS of area 44 was applied in patients with corticobasal degeneration who showed disturbance of their right voluntary hand movements. No responses were elicited from the affected thenar muscles after TMS of the left area 44. On the other hand, normal MEPs were recorded from the unaffected left thenar muscles after stimulation of right area 44. This technique is useful in

differentiating between lesions in area 44 and the M1 area. MEPs by TMS of area 44 in eight patients with poor voluntary movements of hands and in six patients with dystonia showed no response and/or absent inhibitory and facilitatory effects on the EMG activity in 10 out of 14 patients. In contrast, three patients with choreic movements and three patients with myoclonus had normal MEPs after TMS of area 44. 2.6. Tongue MEPs after TMS of tongue M1 area We recorded clear tongue MEPs by TMS of tongue M1 area, which is located 11 cm lateral from Cz. The onset latency of tongue MEP was approximately 8.4 ms and its amplitude was 1.0 mV. An absence of MEPs after TMS of left tongue M1, and normal MEPs after TMS of left area 44 and right tongue M1, were noted in the patient with corticobasal degeneration, who showed oral-facial apraxia. MRI revealed focal atrophy of the left Broca’s area. 2.7. Neurophysiological function of supplementary motor areas (SMAs) We evaluated the neurophysiological function of the SMAs in relation to voluntary movements. TMS was applied to pre-SMA and SMA proper areas in normal subjects. No MEP was produced when the pre-SMA area was stimulated. On the other hand, TMS over SMA proper could produce clear MEPs from the proximal muscles of the upper extremities. TMS over SMA proper could not interrupt voluntary movements of the upper extremities, such as drawing a circle. 2.8. Summary of the neurophysiological characteristics of human area 44 (1) Area 44 plays important roles in the organization and integration of fine finger movements; (2) Area 44 has inhibitory and facilitatory effects lasting for more than 200 ms over both tonic and phasic finger movements;

139 (3) Our results are the first direct evidence that area 44 in humans is strongly involved in voluntary hand movements and that it has direct fastconducting corticospinal projections; (4) Finally, it is well known that area 44 in humans participates in speech. 3. The relationship between voluntary and involuntary movements and sensorimotor oscillatory activity 15–50 Hz oscillatory activity has been identified as local field potentials in monkey motor cortex during motor preparation and attention. MEG and EMG recordings in humans have shown that 20–30 Hz activity was induced during weak sustained muscle contraction and 35–60 Hz activity was noted during strong muscle contraction (Brown et al., 1998). 3.1. High-frequency oscillations (HFOs) of somatosensory evoked potentials (SEPs) recorded from subdural electrodes To investigate the relationship between the primary somatosensory cortex and HFOs, we examined HFOs of somatosensory evoked potentials (SEPs), recorded directly from subdural electrodes, to median and ulnar nerve stimulation. Subdural electrodes were implanted for clinical evaluation prior to surgical treatment of intractable epilepsy. The HFOs were clearly recorded from primary sensory cortex (area 3b) of the thumb. The somatosensory HFOs showed a strictly somatotopic source arrangement. There was phase inversion between the sensory and motor cortices in the prophase components of HFOs. However, the phase was synchronized in the latter part of the HFOs. These results indicate that the origins of the early and late components of HFOs are different, and that there is a clear somatotopy (Kojima et al., 2001) 3.2. Cortical oscillatory activity during voluntary and involuntary movements in humans Movement-related cortical potentials (MRCPS) are well known and have been studied extensively by

Prof. H. Shibasaki. In the EEG, event-related desynchronization (ERD) (Pfurtscheller, 1977) of alpha and beta activity was noted on both primary sensorimotor areas just before and during voluntary movements. Event-related synchronization (ERS) (Pfurtscheller, 1992) of 40–50 Hz (the so-called gamma ERS) was noted on the sensorimotor areas just before and during voluntary hand movements. These ERD and gamma ERS could be related to idling and preparation of voluntary movements (Kuhlman, 1978). We found that post-movement synchronization (PMS) of 15–25 Hz (so-called beta ERS) was induced on the PNMA and SMA at approximately 1s after voluntary hand movements. This activity could be related to the readiness of subsequent voluntary movements. 3.2.1. The relationship between PNMA (area 44) and voluntary hand movements by electrocorticogram (ECoG) in patients with intractable epilepsy ERD and gamma ERS were recorded from subdural electrodes placed over the thumb and hand motor areas just before and during tapping (phasic voluntary movements). High-amplitude PMS (so-called beta ERS) was observed in the negative motor area at the end of voluntary movements (Fig. 2). The onset of PMS in the negative motor area was distinctly earlier than that recorded from primary motor areas (PMAs). PMS was also noticed in the ipsilateral PNMA after the phasic voluntary movements. It is interesting there were apparently different frequencies of beta ERS between the PNMA and PMAs. Frequency analysis of beta ERS showed that 13–18 Hz beta activity arose at the PMA. On the other hand, 18–22 Hz activity recorded from the PNMA was faster than that from the PMAs. Furthermore, beta ERS recorded from the PNMA appeared 200–300 ms earlier than that noted at the PMAs. Furthermore interesting, high-amplitude beta ERS recorded from the PNMA appears at the beginning of tonic contraction and continues at the end of movements. These findings differ distinctly from the appearance of beta ERS during the phasic contractions. Beta ERS recorded from hand motor areas could be related to the idling of voluntary hand movements.

140

Fig. 2. Contralateral electrocorticogram (ECoG) during phasic voluntary movements of left hand. Lt. APB – left abductor polis brevis, Right ECoG – right electrocorticogram.

(Kuhlman, 1978). On the other hand, beta ERS recorded from PNMA could possibly be related to inhibitory effects to the motor area, (Salmelin et al., 1995), and the function of PNMA is related to the readiness of subsequent voluntary movements in humans. 3.2.2. Effects of phasic voluntary movements on epileptiform activities Phasic voluntary movements suppress epileptiform activity and this epileptiform activity is replaced by gamma ERS during and after the movements. 3.2.3. Cortical oscillatory activity at 15–25 Hz in patients with cortical reflex positive myoclonus It is well known that electrical stimulation of the median nerve induces double C reflexes at 23 Hz in patients with cortical reflex myoclonus. Jerklocked back averaging also shows rhythmic cortical activity at 23 Hz in association with myoclonus.

An EEG–EMG polygram after left tibial nerve stimulation reveals rhythmic cortical oscillatory activity and muscular activity at 20 Hz in patients with cortical myoclonus. The onset of oscillatory activity in the EEG is approximately 50 ms earlier than that noted in the EMG recorded from the abductor hallucis muscles. TMS over motor and sensory areas produces rhythmic muscular activity at 20 Hz (Fig. 3). The time lag between sensory and motor cortices is 7 ms, which is the conduction time between these two areas. In paired TMS studies, facilitatory effects of MEP appear when the inter-stimulus interval is 45 ms. It indicates that the rhythm of cortical excitability is 22 Hz. In jerk-locked MEPs, the facilitatory effects of MEP appear at 44 ms after the myoclonus. It indicates that the rhythm of cortical excitability is 23 Hz. From these findings, patients with cortical reflex positive myoclonus have rhythmic cortical excitability at approximately 23 Hz.

141

Fig. 3. Single transcranial magnetic stimulation (TMS) of motor and sensory foot area. MEP Recording muscles: left abductor hallucis muscles

3.2.4. Cortical oscillatory activity at 40–50 Hz in patients with action myoclonus 40–50 Hz oscillations of EMG and EEG activity were recorded in the patient with action myoclonus. The onset of cortical oscillatory activity was 23 ms earlier than that of muscular activity recorded from the thenar muscles. After the medication, this oscillatory activity disappeared in association with improved symptoms. In paired TMS, the facilitatory effect of MEP appeared when the inter-stimulus interval was 24 ms in the patient with action myoclonus due to cerebral hypoxia. In jerk-locked MEPs, the facilitatory effect of MEP also appeared at 24 ms after the myoclonus. It indicates that the rhythm of cortical excitability is 42 Hz. These cortical oscillations could possibly be related to ERS which is seen during and after the voluntary movements. Cortical oscillatory activity at 40–50 Hz was noted in the patients with action myoclonus, and

could be related to gamma ERS. On the other hand, cortical oscillations at 20 Hz was noted during rest in the patients with cortical reflex myoclonus and could be related to beta ERS. We speculate that 40–50 Hz hyperoscillations in sensorimotor cortex could be related to action myoclonus. On the other hand, 15–25 Hz hyperoscillations could be related to cortical reflex myoclonus. In contrast, patients with negative myoclonus have no hyperoscillatory activity and have significantly longer silent periods after magnetic stimulation of the motor cortex (Matsunaga et al., 2000). Furthermore, study on the recovery function of somatosensory evoked potentials (SEPs) suggests a decrease in excitability of the somatosensory cortex. The long-lasting decrease in excitability of the sensorimotor cortices after stimulation could be related to the occurrence of negative myoclonus (Matsunaga et al., 2000).

142 3.2.5. Very fast oscillatory activity in the human motor system We recorded very fast oscillatory activity of EEG and EMG at 300 Hz in a patient with cortical myoclonus. Polyphasic MEPs with very fast oscillations were also recorded in cortical positive myoclonus. 4. Conclusion In conclusion, it has been shown that: (1) the motor and association areas including area 44 (mirror neuron system) are closely related to voluntary and involuntary movements in humans; (2) area 44 has direct fast-conducting corticospinal projections and also participates in speech; (3) cortical oscillatory activity in human brains has a close relationship with voluntary hand movements and modalities of cortical reflex myoclonus. References Brown, P., Salenius, S., Rothwell, J.C. and Hari, R. (1998) Cortical correlate of the piper rhythm in humans. J. Neurophysiol., 80(6): 2911–2917. Di Pellegrino, G., Fadiga, L., Fogassi, L. et al. (1992) Understanding motor events: a neurophysiological study. Exp. Brain Res., 91: 176–180.

Gallese, V., Fadiga, L., Fogassi, L. and Rizollati, G. (1996) Action recognition in the premotor cortex. Brain, 119: 593–609. Iacoboni, M., Woods, R.P., Brass, M. et al. (1999) Cortical mechanisms of human imitation. Science, 286: 2526–2528. Kojima, Y., Uozumi, T., Akamatsu, N., Matsunaga, K., Urasaki, E. and Tsuji, S. (2001) Somatosensory evoked high frequency oscillations recorded from subdural electrodes. Clin. Neurophysiol., 112: 2261–2264. Kuhlman, W.N. (1978) Functional topography of the human mu rhythm. Electroencephalogr. Clin. Neurophysiol., 44: 83–93. Lüders, H., Lesser, R.P., Morris, H.H., Dinner, D.S. and Hahn, J. (1987) Negative motor responses elicited by stimulation of the human cortex. In: P. Wolf et al. (Eds.), Advances in Epileptology, Vol. 16. Raven Press, New York, pp. 229–231. Matsunaga, K., Uozumi, T., Akamatsu, N., Nagashio, Y., Qingrui, L., Hashimoto, T. and Tsuji, S. (2000) Negative myoclonus in Creutzfeldt-Jakob disease. Clin. Neurophysiol., 111: 471–476. Nishitani, N. and Hari, R. (2000) Temporal dynamics of cortical representation for action. Proc. Natl. Acad. Sci. USA, 97: 913–918. Pfurtscheller, G. (1977) Graphical display and statistical evaluation of event-related desynchronization (ERD). Electroencephalogr. Clin. Neurophysiol., 43: 757–760. Pfurtscheller, G. (1992) Event-related synchronization (ERS): an electrophysiological correlate of cortical areas at rest. Electroencephalogr. Clin. Neurophysiol., 83: 62–69. Rizzolatti, G., Fadiga, L., Matelli, M. et al. (1996) Localization of grasp representations in humans by PET: 1. observation versus execution. Exp. Brain Res., 111: 246–252. Salmelin, R., Hämäläinen, M., Kajola, M. et al. (1995) Functional segregation of movement-related rhythmic activity in the human brain. Neuroimaging, 2: 237–243. Uozumi, T., Tamagawa, A., Hashimoto, T. and Tsuji, S. (2004) Motor hand representation in cortical area 44. Neurology, 62: 757–761.


143

Chapter 20

High-frequency oscillations in the human motor system Takenori Uozumi*, Akira Tamagawa, Tomoko Hashimoto and Sadatoshi Tsuji Department of Neurology, University of Occupational and Environmental Health, School of Medicine, Kitakyushu City, Fukuoka 807-8555 (Japan)

1. Introduction Many researchers have reported that high-frequency oscillations (HFOs) are recorded overlapping with the cortical primary response (N20) of somatosensory evoked potentials (SEPs) caused by stimulation of the median nerve (Emori et al., 1991; Curio et al., 1994, 1997; Hashimoto et al., 1996; Gobbele et al., 1998; Ozaki et al., 1998). The HFOs are 300–900 Hz elements and are markedly inhibited by sleep. Curio et al. (1994, 1997) measured somatosensory evoked field potentials and reported that the source of HFOs is in the primary sensory cortex (area 3b). Hashimoto et al. (1996) speculated that HFOs are caused by the activity of GABAergic inhibitory interneurons, because the amplitudes of HFOs and N20 show opposite movements during the sleep cycle. Little is known about HFOs in the human motor system or whether such HFOs occur. To begin addressing this issue, we should look at the periodicity of indirect waves in the pyramidal tract. Day et al. (1989) outlined the direct (D) and indirect (I) wave hypothesis of transcranial magnetic stimulation (TMS) on the basis

*Correspondence to: Takenori Uozumi, Department of Neurology, University of Occupational and Environmental Health, School of Medicine, Kitakyushu City, Fukuoka 807-8555, Japan. E-mail: [email protected]

of single motor unit data. TMS elicits, in the pyramidal tract and corticospinal tract, an unrelayed D wave followed by multiple I waves at frequencies as high as 500–700 Hz. It is thought that activity of similar cortical interneurons in the motor area is also associated with control of movement, but it has been difficult to record HFOs in the human motor area. However, when the C reflex of patients with myoclonus is recorded, polyphasic EMG activity, which is similar to HFOs, is observed in some patients. The current study examined the possibility that the polyphasic C reflex expresses HFOs in the human motor system. 2. Materials and methods The study involved 6 patients (2 males, 4 females) with cortical myoclonus who demonstrated polyphasic C reflexes and large HFOs of SEPs. The mean age of patients was 50 years (range 11–76 years). The causal disease was progressive myoclonus epilepsy (PME) in 3 patients, metabolic encephalopathy in 2 patients, and paraneoplastic syndrome in 1 patient. Action myoclonus was the main feature in all patients, and negative myoclonus was observed in two patients. The C reflex was recorded from the abductor pollicis brevis muscle using a surface electrode. It was recorded at rest, during voluntary contraction, or after electrical stimulation of the median nerve at the wrist.

144 The stimulus strength was set to just below the M-wave threshold, a strength at which the C reflex was most easily recorded. Simultaneous EEG and EMG recordings were performed during voluntary motor actions or after median nerve stimulation. EMG and EEG signals were amplified using a Neuropack λ computer system (Nihon Kohden) with a 1–3000 Hz filter setting. HFOs were observed by 200–900 Hz digital filtering. The motor cortex was stimulated with a magnetic stimulator (Type AAA-81077, Nihon Kohden, Tokyo, Japan). We used a plane figure-eight coil with an inside diameter of 5 cm in each loop. A maximum magnetic field of 0.9 T was obtained with an output voltage of 1800 V about 67 μs after the onset of discharge.

in a PME patient. At rest, the second component of the C reflex (C2) was evoked by weak stimulation of the median nerve. During voluntary contraction, both C1 and C2 were induced even if the stimulation was weak, and HFOs were observed according to the timing of C2. The interpeak intervals of HFO’s of C reflexes in the six cases were 2.5–4 ms (250–400 Hz), and the mean interpeak interval was about 3.2 ms (about 300 Hz). Premyoclonic spikes were seen in all patients on examination by the jerk-locked averaging method at rest. However, there were no HFOs overlapping the premyoclonic spikes. 3.2. Relationship between negative myoclonus and HFOs

3. Results

Figure 1 shows an original waveform of a C reflex during a slight voluntary contraction, and a waveform after digital filtering in a patient with metabolic encephalopathy. About ten HFO components were clearly seen and the HFO’s interpeak interval was 3–3.6 ms. All patients showed motor HFOs only during voluntary contraction. Figure 2 shows double C reflexes

The HFOs of a PME patient who showed negative myoclonus were examined (Fig. 3). The waveform of the C reflex at rest was simple and showed no clear HFOs. A period of inhibited background EMG activity (negative myoclonus) was observed when the voluntary contraction was maintained. The period when the excitability of the motor cortex was reduced could be estimated using the period of inhibited background EMG activity. The waveform of the C reflex during voluntary contraction was similar to the waveform

Fig. 1. The original waveform of the C reflex and the waveform after digital filtering in a patient with metabolic encephalopathy. HFOs consisting of about 10 peaks were clearly seen and the HFO interpeak interval was 3–3.6 ms.

Fig. 2. HFOs in a PME patient with a double C reflex at rest. During voluntary contraction, HFOs were observed according to the timing of the second C reflex (C2).

3.1. HFOs of the C reflex

145

Fig. 3. HFOs in a PME patient with negative myoclonus. The waveform of the C reflex at rest was simple and showed no clear HFOs. The waveform of the C reflex was similar to that at rest, when the impulse that generates the C reflex reached the motor cortex during suppression of the motor cortex. However, when the impulse reached the motor cortex during excitation of the motor cortex, clear HFOs were seen in the C reflex.

at rest, when the impulse that generates the C reflex reaches the motor cortex during the suppression of the motor cortex. However, the onset latency of the C wave during voluntary contraction was about 4 ms later than that at rest. When the impulse reached the motor cortex while background EMG activity was maintained, clear HFOs were seen in the C reflex. Moreover, the onset latency was about 2 ms shorter than at rest. In another patient with positive–negative myoclonus due to metabolic encephalopathy, magnetic stimulation was applied to the motor area of the contralateral hand during voluntary contraction. During this procedure, rhythmic EMG activity appeared 50–80 ms later than the onset latency of motor evoked potential (MEP; Fig. 4). This EMG activity was similar to HFOs of the C reflex, and its interpeak interval was 2.4–3 ms. Moreover, such rhythmic EMG activity did not occur when magnetic stimulation was applied during the period when the motor cortex was not activated, which was estimated using the inhibition of background EMG activity.

Fig. 4. EMG activity after magnetic stimulation to the hand motor area in a patient with transient myoclonus. The EMG activity was similar to HFOs of the C reflex and its interpeak interval was 2.4–3 ms. There was no rhythmic EMG activity even when magnetic stimulation was performed during the period when the motor cortex was not activated.

3.3. Simultaneous EEG and EMG recordings during voluntary contraction Polygraphic analysis showed short-term oscillatory activity over the motor area coupled to EMG oscillations after median nerve stimulation. EMG oscillations lagged EEG oscillations by about 20 ms, and frequencies of both oscillations were about 300 Hz (Fig. 5A). In the same patient, oscillatory activity over the motor area coupled to EMG oscillations was seen in spontaneous myoclonus during voluntary contraction (Fig. 5B). The EEG oscillations preceded the EMG oscillations by 20 ms, and frequencies of both oscillations were about 300 Hz. 3.4. Discussion There is considerable uncertainty about the origin of HFOs of SEP. A recent study conducted with magnetoencephalography (MEG) has shown that HFOs come from the vicinity of region 3b in the primary sensory

146 A

B

Fig. 5. (A) 300 Hz EEG/EMG oscillations after median nerve stimulation in a patient with cortical myoclonus. Polygraphic analysis showed short-term oscillatory activities over the motor area coupled to EMG oscillations after median nerve stimulation. (B) 300 Hz EEG/EMG oscillations and spontaneous myoclonus in the same patient. There was oscillatory activity over the motor area coupled to EMG oscillations in spontaneous myoclonus.

area (Curio et al., 1997; Ozaki et al., 1998). It is known that in animal studies, inhibitory interneurons (fast spiking cells) of the primary sensory area generate the altofrequent spike of 500 Hz (Jones et al., 2000). The existence of a similar neural mechanism is also postulated in the motor cortex. However, HFOs in the human motor cortex have never been reported. The most probable reason that HFOs have not been reported in the human motor cortex is that recording of HFOs of SEP would require averaging of 1000–2000 responses, which is difficult in the motor system. The C reflex used in the present study is the EMG activity that occurred via the sensory cortex, motor cortex, spinal motoneurons, and peripheral nerve. Therefore, it cannot be said that the C reflex reflects all activity of the motor system. However, the characteristics of HFOs observed in the C reflex in the present study are that they are not seen in all patients at rest,

and that they are not seen in patients with negative myoclonus unless the motor cortex is activated. From these results, we conclude that it is a prerequisite that the motor cortex be activated for manifestation of HFOs in the C reflex. Polygraphic analysis showed oscillatory activity of about 300 Hz over the sensorimotor area coupled to EMG oscillations. We speculate that patients with cortical myoclonus have a pathological, synchronous discharge of large populations of cortical neurons and many motor units. Voluntary contraction increases the size and number of I waves evoked by TMS. When the subject contracts maximally, a fourth I-wave becomes visible, and the EMG response is much larger and more complex (Di Lazzaro et al., 1998). This finding suggests that this phenomenon is the result of increased excitability of elements post-synaptic to the site of stimulation in the motor cortex. Additionally, short-interval, pairedpulse, TMS has been used to understand the mechanisms of periodicity of the I-waves. The facilitation (ISI: 1.0–1.5 ms, 2.5–3.0 ms, or 4.5 ms) most likely occurred within the cerebral motor cortex and reflects interactions between circuits normally responsible for production of I-waves (Tokimura et al., 1996). More recent study indicates that two subthreshold conditioning stimuli facilitate inhibitory interactions, but it lacks the rapid periodicity typical of I-wave interaction (Bestmann et al., 2004). This finding suggests that inhibitory interaction involves a different cellular connectivity than that in the excitatory network that elicits I waves. We believe that it is useful to analyze oscillations in the motor system of patients with myoclonus. Highfrequency EEG/EMG oscillations were first investigated by Brown et al. (1999). They performed spectral analysis of surface EEG and EMG activity in eight patients with cortical myoclonus. They reported that 3 patients had significant coherence at higher frequencies (up to 175 Hz), that large populations of cortical neurons must be synchronized to give a surface-recorded EEG wave, and that many motor units must be synchronously activated to elicit a myoclonic jerk. Another work by Mochizuki et al. (1999) indicates that patients with myoclonus epilepsy have large, high-frequency oscillations at about the latency of enlarged P25 and

147 N33 components. In order to address the high-frequency oscillations in the human motor system, we suggest it may be useful to consider the oscillatory motor activity in patients with myoclonus. Axonal gap junctions may be essential for generating very fast oscillations and can induce oscillatory activity in large neuronal networks (Traub et al., 2002). Antidromic propagation in principal cell somata leads to high-frequency small population spikes. Orthodromic propagation to interneurons can lead to very fast firing of interneurons, generating rhythmic, very fast inhibitory postsynaptic potentials (IPSPs) in principal neurons. The latter mechanism may produce HFOs in SEPs. The role of gap junctions in the pathology of diseases of the central nervous system is still unclear. However, we propose that hypersynchronization of neural networks due to pathological gap junctions may contribute to the production of seizure and myoclonus. For example, one can speculate that neurons in a dysplastic region exhibit abnormally extensive axonal branching. If gap junctions are important for seizure and myoclonus in humans, there could be practical consequences for therapy. 4. Conclusion The polyphasic C reflex in patients with cortical myoclonus may induce HFOs in the motor system. Polygraphic analysis revealed oscillatory activity of about 300 Hz occurring over the sensorimotor area that was coupled to EMG oscillations. Our findings also indicate that HFOs in the motor system are abnormally enhanced in patients with cortical myoclonus. Determination of the precise function of HFOs in the human motor system awaits future studies. References Bestmann, S., Siebner, H.R., Modugno, N., Amassian, V.E. and Rothwell, J.C. (2004) Inhibitory interactions between pairs of subthreshold conditioning stimuli in the human motor cortex. Clin. Neurophysiol., 115: 755–764.

Brown, P., Farmer, S.F., Halliday, D.M., Marsden, J. and Rosenberg, J.R. (1999) Coherent cortical and muscle discharge in cortical myoclonus. Brain, 122: 461–472. Curio, G., Mackert, B.M., Burghoff, M., Koetitz, R., Abraham-Fuchs, K. and Hårer, W. (1994) Localization of evoked neuromagnetic 600 Hz activity in the cerebral somatosensory system. Electroencephalogr. Clin. Neurophysiol., 91: 483–487. Curio, G., Mackert, B.M., Burghoff, M., Neumann, J., Nolte, G., Scherg, M. and Marx, P. (1997) Somatotopic source arrangement of 600 Hz oscillatory magnetic fields at the human primary somatosensory hand cortex. Neurosci. Lett., 234: 131–134. Day, B., Dressler, D., De Noordh, C., Marsden, C., Nakashima, K., Rothwell, J. and Thompson, P. (1989) Electrical and magnetic stimulation of the human motor cortex surface EMQ and single motor unit responses. J. Physiol. Lond., 412: 449–473. Di Lazzaro, V., Restuccia, D., Oliviero, A., Profice, P., Ferrara, L., Insola, A., Mazzone, P., Tonali, P. and Rothwell, J.C. (1998) Effects of voluntary contraction on descending volleys evoked by transcranial stimulation in conscious humans. J. Physiol., 508: 625–633. Emori, T., Yamada, T., Seki, Y. et al. (1991) Recovery functions of fast frequency potentials in the initial negative wave of median SEP. Electroencephalogr. Clin. Neurophysiol., 78: 116–123. Gobbele, R., Buchner, H. and Curio, G. (1998) High-frequency (600 Hz) SEP activities originating in the subcortical and cortical human somatosensory system. Electroencephalogr. Clin. Neurophysiol., 108: 182–189. Hashimoto, I., Mashiko, T. and Imada, T. (1996) Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex. Electroencephalogr. Clin. Neurophysiol., 100: 189–203. Jones, M.S., MacDonald, K.D., Choi, B., Dudek, F.E. and Barth, D.S. (2000) Intracellular correlates of fast (>200 Hz) electrical oscillations in rat somatosensory cortex. J. Neurophysiol., 841: 1505–1518. Mochizuki, H., Ugawa, Y., Machii, K., Terao, Y., Hanajima, R., Furubayashi, T., Uesugi, H. and Kanazawa, I. (1999) Somatosensory evoked high-frequency oscillation in Parkinson’s disease and myoclonus epilepsy. Clin. Neurophysiol., 110: 185–191. Ozaki, I., Suzuki, C., Yaegashi, Y., Baba, M., Matsunaga, M. and Hashimoto, I. (1998) High frequency oscillations in early cortical somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 108: 536–542. Plenz, D. and Kitai, S.T. (1996) Generation of high-frequency oscillations in local circuits of rat somatosensory cortex cultures. J. Neurophysiol., 76: 4180–4184. Tokimura, H., Ridding, M.C., Tokimura, Y., Amassian, V.E. and Rothwell, J.C. (1996) Short latency facilitation between pairs of threshold magnetic stimuli applied to human motor cortex. Electroencephalogr. Clin. Neurophysiol., 101: 263–272. Traub, R.D., Draguhn, A., Whittington, M.A., Baldeweg, T., Bibbig, A., Buhl, E.H. and Schmitz, D. (2002) Axonal gap junctions between principal neurons: a novel source of network oscillations, and perhaps epileptogenesis. Rev. Neurosci., 13: 1–30.


149

Chapter 21

Functional changes in cortical components of somatosensory evoked responses by stimulus repetition Minoru Hoshiyamaa,b,* and Ryusuke Kakigia a

Department of Integrative Physiology, National Institute for Physiological Sciences, Myodaiji, Okazaki 444-8585 (Japan) b Department of Health Sciences, Faculty of Medicine, Nagoya University, Nagoya 461-8673 (Japan)

1. Introduction The recovery function (RF) indicates how a brain responds to the second of two stimuli delivered at a short interstimulus interval (ISI). For sub-cortical somatosensory evoked potential (SEP) components, the RF of the component is simply determined by the number of synapses interposed between the stimulus site and the source of the response (Rosner et al., 1960; Allison, 1962; Meyer-Hardting et al., 1983). Up to the initial cortical response, N20 (N19 in previous articles), the RF of short-latency SEP components complied with the theory (Emori et al., 1991). However, for the cortical components of SEP/somatosensory evoked magnetic field (SEF), RF was not simple and did not depend on the peak latency (Hoshiyama and Kakigi, 2001). In the first experiment of this study, we studied the RF of SEP and SEF using very short ISIs to investigate the possible functional relationship between components of evoked responses, and the functional mechanism of

*Correspondence to: Minoru Hoshiyama, M.D., Ph.D., Department of Health Sciences, Faculty of Medicine, Nagoya University, Nagoya 461-8673, Japan. Tel: +81-(0)52-719-3159; E-mail: [email protected]

the RF was discussed in the second experiment of repetitive stimulation evoked SEF. 2. Subjects and methods 2.1. Subjects Ten volunteers from our department (5 males and 5 females, aged 26–39 years) were examined. Informed consent to participate in the study, which was first approved by the Ethical Committee of the National Institute for Physiological Sciences, Okazaki, Japan, was obtained from all participants prior to the study. The subjects were not told about the prospective results of the study prior to the recording. 2.2. Recovery function of SEP and SEF 2.2.1. SEP recording We focused on two major cortical SEP components, N20 and P25, which can be recorded in the post-central region with a cephalic reference at the mid-frontal area with a high signal to noise ratio. The electroencephalogram (EEG) was recorded, using a pair of 7 mm silver– silver chloride disk electrodes. The exploring electrode was placed in the parietal area contralateral to the stimulated side (5 cm posterior to Cz and 7 cm left of

150 the mid-line) with an Fz reference of the International 10–20 system (Mauguière et al., 1999). Another exploring electrode was placed on the right Erb point with a reference electrode at the left Erb point to record nerve action potentials (NAPs) (Mauguière et al., 1999). The impedance between electrodes was less than 3 kΩ. Τhe signals were amplified with a bandpass filter from 1 to 500 Hz for EEG and 5–1000 Hz for the NAP at the right Erb point. 2.2.2. SEF recording The SEFs were measured with a 37-channel magnetoencephalography system (MEG, Magnes, Biomagnetic Technologies Inc., San Diego, CA). The detection coils of the MEG were arranged in a uniformly distributed array in concentric circles over a spherically concave surface. The device was 144 mm in diameter and its radius of curvature was 122 mm. Each coil was 20 mm in diameter and the distance between the coil centers was 22 mm. The measurement matrix was centered in the right hemisphere at around C4 of the International 10–20 system in each subject. The magnetic responses were filtered with a 1–200 Hz bandpass filter and digitized at a sampling rate of 2048 Hz. The analysis time was 50 ms before and 400 ms after the application of each stimulus. The DC offset of magnetic signals was achieved using the pre-stimulus period as the baseline. Trials in which MEG deflection exceeded 3 fT were excluded from the average. Two hundred artifact-free MEG signal periods of single (control) and double stimulation were averaged separately. 2.3. Experiment 1: RF study using very short ISI We studied the RF with a paired stimulation technique similar to previous methods (Meyer-Hardting et al., 1983; Emori et al., 1991). Single (control) and double stimulation at an ISI of 0.5–100 ms was applied to the right median nerve at the wrist using an electrical stimulator (SEM-4201, Nihon-Kohden). The electrical stimulus was a constant current square-wave pulse of 0.2 ms duration. The intensity of each stimulus was adjusted to produce a slight thumb twitch, which was sufficient to produce an SEP of saturated amplitude in response to a single stimulus. The mean stimulus intensity was 4.7 ± 0.9 mA (mean ± SD).

An advantage of this study was that the stimulation and data collection were fully controlled by a signal processor (1401-plus, CED) with a personal computer system. We set up the recording equipment so as to collect all data in one session to minimize the intraindividual and intra-trial variation of EEG and MEG signals and to shorten the recording time. One SEP recording session comprised all recording periods; one control recording with a single stimulus and 17 doublestimuli recordings with different ISIs, i.e. 0.5, 0.8, 1, 1.2, 2, 3, 5, 7, 9, 12, 15, 20, 25, 30, 40, 60, and 100 ms. For SEF recording, ISIs between 1 and 100 ms were used because of system capacity limitations. The control and double-stimuli conditions were applied in a pseudo random order. The interval between the single or second stimulation of the double stimuli and the initial stimulus of the following period was 0.7 s. Three hundred periods were collected for each condition with digital marking in each. One session took approximately 60 min and the sessions were repeated on another day to test the reproducibility of the responses. 2.4. Experiment 2: repetitive stimulation The SEF was recorded in the second experiment. Stimulation was a single- to 6-train variety with three different ISIs, i.e. 10, 20, and 30 ms. We used 16 sets of stimuli, which comprised single stimulation, doubleto 6-train stimulation with a 10 ms ISI, double- to 6-train stimulation with a 20 ms ISI and double- to 6-train stimulation with a 30 ms ISI. The interval between the offset of the last train stimulation and the onset of the following train stimulation was 1 s. Each single or train stimulus was randomly delivered to the right median nerve at the wrist using a saddle-type electrode. The electrical stimulus was a constant current square-wave pulse of 0.2 ms duration. The intensity of each stimulus was adjusted to produce a slight thumb twitch, which was sufficient to produce an SEF of saturated amplitude to a single stimulus. The average intensity was 4.6 mA (range 3.7–5.3 mA). The duration of the recording period was 300 ms after the onset of train stimulation. The bandpass filter was 1–200 Hz with 2048 Hz sampling. Two hundred periods were collected separately for each condition with digital marking in each stimulus

151 train condition. One session took approximately 50 min with a short rest every 10 min during recording. 3. Results 3.1. RF of the cortical components of SEP/SEF After a single stimulation under control conditions, an SEP waveform consisting of N20 and P25 was recorded in all subjects. For the SEF, 1M and 2M components were obtained 20 ms and 30 ms after stimulation, respectively. The N9 potential was obtained at the Erb point. By subtracting the single-stimulation waveform from the waveform following double stimulation, the waveform evoked by the second stimulation was obtained at each ISI (Fig. 1). The recovery curve of the N9 response showed minimal response at an ISI of 0.8 ms, the response at 0.5 ms being questionable (Fig. 1). The N9 was consistently recognized at ISI of 1 ms or longer, and the amplitude recovered progressively as the ISI increased. Full recovery of the N9 amplitude was obtained at an ISI of 5 ms or longer. The N20 and 1M components were markedly attenuated at a shorter than 10 ms ISI. Full recovery of the N20 amplitude needed an ISI of 100 ms, while the latency recovered at 15 ms. Both the amplitude and latency changed as a function of ISI in P25 and 2M, as compared to N20 (P < 0.01, ANOVA). The P25 and 2M were clearly recognized at an ISI of 1.2 ms in all subjects, and 6 subjects showed the component even at 1 ms. The amplitude recovered progressively; however, a relatively large standard deviation (SD) was recognized from 1 to 5 ms and from 12 to 30 ms ISI, while a large SD of the N20 amplitude was seen from 20 to 60 ms (Fig. 2). In seven subjects, the SEP waveform showed an additional positive component after the P25 at an ISI from 1.2 to 9 ms, shown with an asterisk in Fig. 1. In two of seven subjects, the additional component was not identified at ISIs of 30 and 40 ms. In the SEF waveform, a similar additional component was seen after the 2M component at short ISI in the subject in Fig. 1, but it was not consistently recognized in other subjects. The additional SEP component fused with the P25 component or became very small at an ISI of 20 ms or longer, and its peak latency became shorter

Fig. 1. Somatosensory evoked potentials (SEPs) (left column), nerve action potentials (NAPs) (middle column), and somatosensory evoked magnetic fields (SEFs) (right column) in a representative subject. The top traces (S) were obtained by single stimulation following stimulation of the median nerve. The waveforms were obtained by subtracting the S waveform from the waveform obtained following double stimulation with various interstimulus intervals (ISIs). The N20 and 1M components markedly attenuated when the ISI was less than 10 ms, while the P25 component remained even at an ISI of 1.2 ms. A sub-component of SEP appeared at between 1.2 and 9 ms (asterisk). The NAP recorded at the ipsilateral Erb point was recognized at an ISI of 0.8 ms or longer.

152

Fig. 2. Functional recovery curve (RF) for the amplitude and peak latency of SEP components and NAP (N9) (×) in all subjects. The N20 (䊊) component was attenuated when the ISI was shorter than 40 ms, and diminished in amplitude when the ISI was less than 10 ms. P25 (●) remained even when the ISI was 3 ms. The recovery curve of N20 was significantly different from that of P25 (P < 0.01). All values are standardized to the control. Similar to the amplitude, the RF for peak latency differed between N20 and P25.

153 TABLE 1 THE PEAK LATENCY AND RMS VALUES OF SEF COMPONENTS FOLLOWING SINGLE STIMULATION SEF components

Peak latency (mean ± SD) (ms) RMS (mean ± SD) (fT)

1M

2M

3M

4M

18.9 ± 0.9 85.5 ± 30.6

28.9 ± 1.7 68.1 ± 38.7

45.5 ± 3.2 49.0 ± 19.7

72.6 ± 7.2 104.4 ± 32.6

as ISI increased, from 59.7 (ISI = 1.2 ms) to 39.5 ms (ISI = 40 ms). There was no corresponding component in the control waveform in any of the 7 subjects.

or among the number of stimulus repetitions. The latency ratio of the 3M component was significantly greater than that of 1M in the 20 and 30 ms sequences (P < 0.02, ANOVA).

3.2. SEF changes during repetitive stimulation 4. Discussion In the waveforms following the single stimulation, four SEF components were identified. The components were termed 1M, 2M, 3M, and 4M, and the peak latency and RMS value are shown in Table 1. The SEF waveform evoked by the last train stimulation was obtained by subtracting the waveform of a less numerous train stimulation from the waveform, e.g. the SEF waveform evoked by the fifth stimulation was obtained by subtracting the 4-train stimulation from the 5-train stimulation (Fig. 3). The RMS values of 1M and 4M components attenuated with stimulation repetition in all ISI sequences (P < 0.01), and the attenuation was greater in the 10 ms ISI sequence than in the 20 and 30 ms ISI sequences (P < 0.02) (Fig. 4); however, there was no significant change among the number of repetitions. For the 2M and 3M components, there was evidently a different change in the RMS values. The RMS values of the 2M component did not decline, and for the 3M component, the RMS values increased in the 20 and 30 ms ISI sessions with the number of repetitions (P < 0.02, ANOVA). The latency was generally prolonged and the prolongation of the 1M component was significantly greater in the 10 ms ISI sequence than in the 20 and 30 ms ISI sequences (P < 0.02, ANOVA); however, for the 2M, 3M, and 4M components, there was no significant difference in the prolongation among the sequences

4.1. Recovery function of the cortical evoked responses at very short ISI This RF study has provided the functional characteristics of cortical components of the median nerves, SEP and SEF: first, the RF of the cortical components of the median nerve SEP did not depend on their peak latency; second, the SEP sub-components, which were not recognized by conventional SEP following singlepulse stimulation, appeared following double stimulation with a very short ISI. In previous studies of experimental animals, the RF of a SEP component was determined by the number of synapses interposed between the stimulus site and the potential generator (Wiederholt, 1978). In other words, an SEP component with longer peak latency needed a longer ISI to recover than a component with shorter peak latency. In a human study (Meyer-Hardting et al., 1983), there was less attenuation of the SEP components at short ISIs, compared with the animal study results, although the result was compatible with the theory of the relationship between peak latency and the RF. A subsequent RF study of the sub-cortical components and the N20 of the median nerve SEP (Emori et al., 1991) did not contradict the theory that SEP components with longer peak latency showed more delayed recovery, although the same group reported that

154

Fig. 3. The SEF waveform evoked by the last train stimulation of a train was obtained by subtracting the waveform of a less numerous train stimulation from the waveform, e.g. the SEF waveform evoked by the fifth stimulation (v) was obtained by subtracting the 4-train stimulation from the 5-train stimulation. The small roman numerals (i–vi) indicate the number of the last stimulus in the train.

maximum depression occurred at a longer ISI for the early components than for the late components in a study of lower limb SEP (Saito et al., 1992). In this study, it was clearly shown that N20 and 1M needed a longer ISI to recover than P25 and 2M. These SEP results were consistent with the SEF study, although the SEF study results do not apply simply to SEP, since MEG records mainly tangential cortical responses to the recording coil surface. The mechanisms underlying different RFs among SEP components may be complex. Attenuation in the RF of SEP components at short ISIs has been thought to be functional refractoriness of neurons in the sensory cortex. Since the attenuation of N20 and P25 was clearly different, and since there was, as far as we could ascertain, no evidence of ISI-specific inhibitory

processes in the central nervous system, we speculated that each SEP component has a specific refractory period. The RF study showed at least one additional positive sub-component recognized at a very short ISI, which was not identified in the control waveform. The sub-component seemed to fuse with the P25 component or to diminish at ISI longer than 20 ms. The existence of a sub-component indicated that some SEP components might have multiple generators. If one component had an inhibitory effect on another component or sub-component, the balance inhibitory and excitatory factors might determine the waveform under control conditions. We confirmed that, in the recovery course, a set of SEP components found under control conditions was not preserved at a short ISI, but sub-components

155

Fig. 4. The RMS changes of SEF components during stimulus repetition. Each value was expressed as a ratio to that of the SEF waveform following single stimulation. The small roman numerals (ii–vi) indicate the stimulus number in the train. The 1M and 4M components were attenuated during train stimulation, while 2M and 3M were not. There was no significant decrease in the RMS values in the stimulus train, but for the 3M component, the RMS values rose in the 20 and 30 ms ISI sessions as the number of repetitions increased (P < 0.02, ANOVA), and the RMS values of 3M at the fourth (iv) to sixth (vi) stimulations with ISIs of 20 and 30 ms were greater than those of the second (ii) and third (iii) stimulations (*P < 0.02, ANOVA, Bonferroni–Dunn’s correction).

156 appeared or a combination of components could be modulated by ISI. 4.2. SEF changes during repetitive stimulation We recognized attenuation of the SEF components in terms of amplitude following the second stimuli train but no waning or waxing of the response was recognized between the third and sixth stimuli, and the pattern of attenuation differed among the SEF components. Previous studies reported the attenuation of SEP components evoked by stimulation with a relatively high stimulus rate up to 30 Hz (Pratt et al., 1980; Abbruzzese et al., 1990; Delberghe et al., 1990; Fujii et al., 1994), and that most of the loss occurred after the second stimulation in a repetitive series with an ISI of between 500 and 800 ms (Angel, 1985). Saito et al. (1992) and Fujii et al. (1994) reported that the attenuation mechanism in the RF of SEF was partially due to interference between the electrical stimulation and secondary afferent input from the muscle contraction. Since the muscle also contracted during train stimulation in this study and we could not know whether the afferent signal from the muscle was stable during train stimulation, we could not exclude the effects of muscle contraction on SEF changes, especially on SEF change following the second stimulation. However, since the decrement of 1M and 2M components was not affected by stimulus repetition during the second to sixth stimulation, we considered that we could focus on another point. We considered that the functional refractoriness of the neuron might partially relate to the synaptic ability to respond to high-frequency stimulation, and that SEP change during repetitive stimulation with a very short ISI would give information on the synaptic factor of the central nervous system. If the synaptic factor mainly contributed to the attenuation of an SEF component during repetitive stimulation, the response might change in amplitude between the second and later components. The decrement of the 1M component was approximately 50% following second stimulation at an ISI between 10 and 30 ms, and there was no significant change in the value with the following stimulations. Therefore, the synaptic factor was unlikely to be a major factor in SEF

attenuation during double and repetitive stimulation, at least at an ISI of 10–30 ms in healthy subjects, although we could not exclude the partial contribution of the synaptic factor. In our previous study, we speculated that tapering axons might contribute to SEF attenuation in the RF (Hoshiyama and Kakigi, 2001, 2002). We considered that the morphological structure of axons might be responsible for the 1M attenuation change in this study. The pattern of 1M attenuation was evidently different from that for 2M. The 1M components decreased during repetitive stimulation, but the 2M component showed no significant attenuation. These results were consistent with the first RF experiment. With regard to later responses, the 3M component showed no significant attenuation during repetitive stimulation, but it was strengthened at the fourth to sixth stimuli. For the 1M and 2M components, the major generator has been estimated to exist around the primary somatosensory cortex (SI) (Hari et al., 1984; Wood et al., 1985; Kakigi, 1994; Hoshiyama and Kakigi, 2001), but at the 3M latency, activity in the secondary somatosensory cortex (SII) (Karhu and Tesche, 1999; Wegner et al., 2000) and posterior parietal cortex (PPC) started to overlap (Forss et al., 1994). The SII response was sustained during repetitive stimulation, while the SI response was sharp and transient (Forss et al., 2001). The facilitation of 3M by stimulus repetition might be due to differences in the timing and strength of the inhibition followed by excitation (Forss et al., 2001) and in this study, sustained muscle contraction during train stimulation might contribute to component change. On the other hand, the 4M component showed amplitude attenuation during stimulus repetition. The 3M and 4M components could contain SII activity, but these results suggested that the function of the 4M component was different from that of the 3M component. Concerning the similarity of change between the 1M and 4M components during train stimulation, we could not confirm the mechanism from these results, although one possibility is that the 4M response was the secondary response from the 1M signal. In conclusion, we reported the RF of cortical SEF components at very short ISI and the change in components during repetitive stimulation. The RF of N20 and

157 P25 components did not depend on the peak latency of those components. A sub-component was recognized at a very short ISI, which was not identified following a single stimulation under control conditions. Amplitude attenuation of the SEF components, 1M and 4M, recognized after the second stimulation, did not change with the third and later stimulations, but one component, 3M, was facilitated by stimulus repetition. We considered that the SEP waveform at a stimulus rate was determined by excitatory and inhibitory balance among the components, which might be changed by the temporal factor of stimulation. References Abbruzzese, G., Dall’Agata, D., Morena, M., Reni, L., Trivelli, G. and Favale, E. (1990) Selective effects of repetition rate on frontal and parietal somatosensory evoked potentials (SEPs). Electroencephalogr. Clin. Neurophysiol. Suppl., 41: 145–148. Allison, T. (1962) Recovery functions of somatosensory evoked responses in man. Electroencephalogr. Clin. Neurophysiol., 14: 331–343. Angel, R.W., Quick, W.M., Boylls, C.C., Weinrich, M. and Rodnitzky, R.L. (1985) Decrement of somatosensory evoked potentials during repetitive stimulation. Electroencephalogr. Clin. Neurophysiol., 60: 335–342. Delberghe, X., Mavroudakis, N., Zegers de Beyl, D. and Brunko, E. (1990) The effect of stimulus frequency on post- and pre-central short-latency somatosensory evoked potentials (SEPs). Electroencephalogr. Clin. Neurophysiol., 77: 86–92. Emori, T., Yamada, T., Seki, Y., Yasuhara, A., Ando, K., Honda, Y., Leis, A.A. and Vachatimanont, P. (1991) Recovery functions of fast frequency potentials in the initial negative wave of median SEP. Electroencephalogr. Clin. Neurophysiol., 78: 116–123. Forss, N., Hari, R., Salmelin, R., Ahonen, A., Hämäläinen, M., Kajola, M., Knuutila, J. and Simola, J. (1994) Activation of the human posterior parietal cortex by median nerve stimulation. Exp. Brain Res., 99: 309–315. Forss, N., Narici, L. and Hari, R. (2001) Sustained activation of the human SII cortices by stimulus trains. Neuroimage, 13: 497–501. Fujii, M., Yamada, T., Aihara, M., Kokubun, Y., Noguchi, Y., Matsubara, M. and Yeh, M.H. (1994) The effects of stimulus rates upon median, ulnar and radial nerve somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 92: 518–526.

Hari, R., Reinikainen, K., Kaukoranta, E., Hämäläinen, M., Ilmoniemi, R., Penttinen, A., Salminen, J. and Teszner, D. (1984) Somatosensory evoked cerebral magnetic fields from SI and SII in man. Electroencephalogr. Clin. Neurophysiol., 57: 254–263. Hoshiyama, M. and Kakigi, R. (2001) Two evoked responses with different recovery functions in the primary somatosensory cortex in humans. Clin. Neurophysiol., 112: 1334–1342. Hoshyima, M. and Kakigi, R. (2002) New concept for the recovery function of short-latency somatosensory evoked cortical potentials following median nerve stimulation. Clin. Neurosphysiol., 113: 535–541. Kakigi, R. (1994) Somatosensory evoked magnetic fields following median nerve stimulation. Neurosci. Res., 20: 165–174. Karhu, J. and Tesche, C.D. (1999) Simultaneous early processing of sensory input in human primary (SI) and secondary (SII) somatosensory cortices. J. Neurophysiol., 81: 2017–2025. Mauguière, F., Allison, T., Babiloni, C., Buchner, H., Eisen, A.A., Goodin, D.S., Jones, S.J., Kakigi, R., Matsuoka, S., Nuwer, M., Rossini, P.M. and Shibasaki, H. (1999) Somatosensory evoked potentials. In: G. Deuschl and A. Eisen (Eds.), Recommendations for the Practice of Clinical Neurophysiology: Guidelines of the International Federation of Clinical Physiology, EEG Suppl. 52. Elsevier, Amsterdam, pp. 79–90. Meyer-Hardting, E., Wiederholt, W.C. and Budnick, B. (1983) Recovery function of short-latency components of the human somatosensory evoked potential. Arch. Neurol., 40: 290–293. Pratt, H., Politoske, D. and Starr, A. (1980) Mechanically and electrically evoked somatosensory potentials in humans: effects of stimulus presentation rate. Electroencephalogr. Clin. Neurophysiol., 49: 240–249. Rosner, B.S., Allison, T., Swanson, W. and Goff, W.R. (1960) A new instrument for the summation of evoked responses from the nervous system. Electroencephalogr. Clin. Neurophysiol., 12: 745–747. Saito, T., Yamada, T., Hasegawa, A., Matsue, Y., Emori, T., Onishi, H. and Fuchigami, T. (1992) Recovery functions of common peroneal, posterior tibial and sural nerve somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 85: 337–344. Wegner, K., Forss, N. and Salenius, S. (2000) Characteristics of the human contra- versus ipsilateral SII cortex. Clin. Neurophysiol., 111: 894–900. Wiederholt, W.C. (1978) Recovery function of short latency components of surface and depth recorded somatosensory evoked potentials in the cat. Electroencephalogr. Clin. Neurophysiol., 45: 259–267. Wood, C.C., Cohen, D., Cuffin, B.N., Yarita, M. and Allison, T. (1985) Electrical sources in human somatosensory cortex: identification by combined magnetic and potential recordings. Science, 227: 1051–1053.


159

Chapter 22

Origins and characteristics of high-frequency (>500 Hz) SEP components directly recorded from the cervical cord, thalamus, and cerebral cortex Eiichirou Urasakia,*, Rieko Maedaa, Naoki Akamatsub and Akira Yokotaa a

Department of Neurosurgery, and bDepartment of Neurology, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka,Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555 (Japan)

1. Introduction High-frequency (HF) activity along the somatosensory tract has been intensively investigated in recent years. The HF component (HFC) of cortically or scalprecorded somatosensory evoked potentials (SEPs) could be extracted from original low-frequency (LF) waveforms by digital filtering. The origins of cortically recorded HFC are postulated as thalamo-cortical radiation fibers, intracortical activities, or a combination of these activities (Curio, 2000). Burst firing with multiple spikes and HF firing rates are known to exist in the dorsal column nucleus, thalamus, and cerebral somatosensory cortex (Curio, 2000). Rhythmic HF waves appear to be conducted peripherally to centrally along the somatosensory tract, because such HFCs have been reported to exist even in more caudal regions, including the spinal cord (Prestor et al., 1997). It is pertinent to study the characteristics of HFCs and

*Correspondence to: Eiichirou Urasaki, Department of Neurosurgery, University of Occupational and Environmental Health, School of Medicine, 1-1 Iseigaoka,Yahatanishi-ku, Kitakyushu City, Fukuoka 807-8555, Japan. Tel: +81-93-603-1611; Fax: +81-93-691-8755; E-mail: [email protected]

corresponding LF SEPs at each level of the central nervous system to clarify how peripheral HFCs propagate centrally. We here report the different characteristics between HF and LF SEPs recorded from the cervical cord or cerebral cortex under different stimulus rates to the median nerve. Additionally, the dissociated loss of cortical HFC with preserved thalamic HFC in a patient with thalamic lesion is presented to provide evidence that thalamic HFO does not directly contribute to the formation of cortically recorded HFC. This study has not been published elsewhere except for the description of cortically recorded HFCs (Urasaki et al., 2002). 2. Patients and methods We recorded SEPs directly from the cervical dorsal epidural space of 8 patients, 7 males and one female, aged between 16 and 75 years (mean ± SD = 62.1 ± 20.1), who underwent laminoplasty for the treatment of narrow cervical canal (1 patient) or cervical spondylosis (7 patients) accompanied with radiculomyelopathy. Motor weakness was found in the bilateral upper and lower extremities in 6 patients, bilateral upper extremities in one, and unilateral upper limb in one. Sensory disturbance was detected below C4 in 4 patients,

160 below C5 in three, and below Th5 in one. Surgery was performed through C2–C6 in two patients, C2–C7 in three, C2–Th1 in one, C3–C7 in one, and C4–Th1 in one. Dome laminotomy was performed at C2 and laminoplasty using a ceramic spacer was performed at other levels. Direct SEP recordings were obtained from the C2 to C6 laminar levels in two patients, C2–C7 in two, C3–C7 in two, and C4–Th1 in two. After decompression of the cervical roots and spine, and just before reconstruction of the laminae, an epidural strip electrode was placed on the exposed areas. The diameter of the recording electrodes was either 1 or 4 mm, and the inter-electrode distance from center to center was either 5 or 10 mm. This study was approved by the institutional Review Board of the University of Occupational and Environmental Health. A reference needle electrode was placed on the dorsal neck skin near the skin incision. SEP was obtained using 0.2 ms square-wave electrical pulses delivered transcutaneously to the median nerve at the wrist. Saddle-type bipolar electrodes, a cathode 3 cm proximal to the anode, were used. The period of analysis was 50–100 ms. Stimulus intensity was adjusted to approximately 2–3 times the motor threshold during general anesthesia. To study the effects of different stimulus rates on HF and LF SEP waveforms, 3.3 and 12.3 Hz stimulus rates were employed. The filter setting was 10–1000 Hz, two averages of 500–1000 responses were obtained to confirm reproducibility, and two sets of responses were averaged using a computer for later analysis. HF SEPs were extracted by digital filtering of wideband-recorded LF responses through a bandpass of 500–1000 Hz, using a Viking IV computer system (Nicolet). One of the SEP waveforms with maximal amplitude was selected from SEPs recorded along the cervical dorsal epidural space to analyze the amplitude change between low and high stimulus rates. N11 and N13 components in the selected LF cervical SEPs were identified, and then two negative HFCs showing peaks respectively similar to N11 and N13 were identified among several HF peaks. N11 and N13 amplitudes were measured from the baseline, and those of corresponding negative HFCs were calculated from the preceding positive trough. Amplitude change ratios

were calculated by dividing SEP amplitudes at 12.3 Hz stimulus rate by those at 3.3 Hz, and expressed as a percentage. Anesthesia was induced with propofol combined with fentanyl. Following manual hyperventilation for several minutes, patients were intubated. Volatile anesthetic agents were used to maintain general anesthesia. Nitrous oxide was supplemented with isoflurane and/or sevoflurane. The concentration of volatile anesthesia was changed to maintain hemodynamic stability. Student’s two-tailed t-test was used for statistical analysis, and a p value less than 0.05 was considered significant. In one patient with thalamic glioma, direct recording from the thalamus was made, using a small ball electrode with a diameter of 1 mm, during surgery for partial removal of the tumor under general anesthesia. Cortical SEPs obtained from one out of 5 patients who underwent intracranial subdural electrode implantation for the pre-surgical evaluation of epilepsy focus were shown to emphasize the dissociated changes of cortically recorded LF and HF SEPs. This patient was included in a previous report, although the SEP waveform was not published. SEP was studied in an awake state, and has been described in greater detail elsewhere (Urasaki et al., 2002). 3. Results 3.1. Cervical SEPs Original cervical SEPs and the results of digital filtering are shown in Fig. 1. Appreciable HFCs with three or more negative peaks were clearly recorded (Fig. 1). During negative LF cervical SEP, from the end of the preceding positivity to N11 and N13 peaks followed by the descending slope of N13, there were three negative– positive HF deflections (Figs. 1–3). An additional negative–positive HFC was present during the duration of the preceding positivity but before the appearance of the ascending slope of N11 in LF SEPs (Figs. 1–3). Peak latencies of large two negative HFCs were close to those of N11 and N13, as shown in Figs. 2 and 3. When stimulus rates to the median nerve were increased from 3.3 to 12.3 Hz, the amplitude of N13

161

Fig. 1. Cervical SEPs recorded from the Erb point and C2–C7 laminar epidural space in a 64-year-old man with cervical spondylosis who underwent laminoplasty. On wide bandpass recording (10–1000 Hz), cervical SEP showed a positive wave followed by two negative potentials, designated as N11 (black circle) and N13 (white circle), respectively. The latency of the positive wave and N11 increased when the recording was made in rostral derivations. HFCs after digital filtering (500–1000 Hz) of wide bandpass recorded SEPs showing three or four negative peaks propagated rostrally. The Erb N9 peak coincided with the first negative peak of the HFC. Second, third, and fourth negative HFCs were present during the duration of N11 and N13, showing second negative HFC with maximal and fourth HFC with minimal amplitudes. When the stimulation rate to the median nerve was increased from 3.3 (dotted line) to 12.3 Hz (black line), the N13 amplitude decreased remarkably, while the HFC amplitudes remained unchanged.

was decreased while no HFCs showed amplitude reduction (Fig. 1). Figure 2 shows the superimposition of cervical SEPs recorded from C3 to C6. Dissociated amplitude changes between LF and HF components under different stimulus rates were clearly shown. N11 and N13 amplitudes decreased while corresponding HFCs showed no amplitude reduction. There was one exceptional case that showed a rather slight increase of cervical HFCs (Fig. 2). None of the other 7 cases showed any increase in cervical HFCs (Figs. 1 and 3). Table 1 and associated graphs show comparisons of amplitude change ratios between LH and HF cervical components. The amplitude change ratio of spinal N11 was 95.5%, indicating that the amplitude of N11 at 12.3 Hz stimulation was 95.5% of those at 3.3 Hz. The amplitude ratio of HFC corresponding to N11

Fig. 2. Cervical SEPs recorded from C3 to C6 in a 16-year-old man with a narrow cervical canal. LFCs and HFCs recorded from C3 to C6 were superimposed. The amplitudes of N11 and N13 were decreased under 12.3 Hz, but those of corresponding HFCs were not decreased. Black and white circles in the upper row indicate N11 and N13, and those in the lower row indicate HF negative peaks, those latencies being near N11 and N13.

Fig. 3. Cervical SEPs recorded from a 75-year-old man with cervical spondylosis, demonstrating ascending activities except for the cervical N13 component. Positivity before the peak of N11 and N11 conducted from C4 to C6 laminar levels in approximately 80 m/s. Initial HF positivity and subsequent HF negativity with three peaks, designated as HN1, HN2, and HN3, were propagated along the cervical cord at 108, 74, 99, and 94 m/s, respectively. Black and white circles on HN2 and HN3 are peaks with similar respective latency to those of N11 and N13, but there was no latency shift along the cervical cord in spinal N13, indicating characteristics different from NH3.

162 was 84.5%. There was no significant difference between changes in these two amplitude ratios. The amplitude of N13 components decreased by about 76% when the stimulus rate increased, but corresponding HFC showed little change as demonstrated by 94.3% of amplitude change ratio. The difference between these two amplitude change ratios was significant (p = 0.03), indicating the stable amplitude of

HFCs under a high stimulus rate to the median nerve (Table 1). HFCs contained several peaks and propagated in a caudal–rostral direction along the cervical cord with constant latency shift (Figs. 1 and 3). As shown in Fig. 3, three negative HFCs and the preceding positivity traveled along the dorsal cervical spine similar to the positive and subsequent N11 component under

TABLE 1 COMPARISONS OF AMPLITUDE CHANGE RATIOS (12.3 Hz amplitude/3.3 Hz amplitude between low- and high-frequency components (n = 8))

(%)

Spinal N11

Low frequency

Spinal N13

p value

frequency 84.5 ± 16.4%

0.07

N13 76.2 ± 7.4% 94.3 ± 23.5%

0.03*

(%)

N11 95.5 ± 6.5%

High

Comparisons of amplitude changes between spinal N11 and corresponding HFC, and spinal N13 and its corresponding HFC are shown. The amplitude of spinal N13 decreased significantly when compared to corresponding HFC under a higher stimulus rate. Squares and circles indicate LF N11, N13, and corres ponding HF components, respectively *Statistically significant (paired t-test, two-sided).

163 wide bandpass conditions. Spinal N13 demonstrated no apparent latency prolongation (Fig. 3). 3.2. Thalamic SEPs Figure 4 shows the dissociated loss of cortical LF and HF SEPs with preserved thalamic LF and HF SEPs in a patient with right thalamic glioma. During the removal of right thalamic tumor, cortical SEPs were monitored by placing the recording electrode on the right sensory cortex. LF and HF cortical SEPs were clearly recorded before tumor removal. During partial removal of the tumor, cortical SEPs, including LF and HF components, disappeared. Direct recording from the thalamic cavity showed residual thalamic SEPs of both LF and HF. From the configuration of thalamic SEPs, the recording site was assumed to be the VC nucleus (Urasaki et al., 1989; Hanajima et al., 2004). The onset of thalamic LF SEPs was slightly earlier than that of cortical N20 recorded before tumor removal. This tendency was also similar to HFCs. The duration of cortical and thalamic HFCs was within the duration of N20 and thalamic positivity, respectively (Fig. 4).

Fig. 4. SEPs directly recorded from the right-hand sensory area (SI) and thalamus during surgery for right thalamic glioma in an 18-year-old man. N20 and corresponding HFCs before tumor removal (upper row) disappeared after partial removal of the tumor (middle row), although large amplitude LF and HF components were clearly recorded from the thalamus (lower row). The recording site of the thalamus is shown on a post-operative CT scan.

3.3. Cortical SEPs The effect of a high stimulation rate on LFCs and HFCs is shown in Fig. 5. Direct recording from sensory and motor cortices disclosed primary cortical components of area 3b N20 and its phase reversal P20, respectively; corresponding HFCs were clearly recorded. When the stimulus rate to the median nerve was increased from 3 to 12 Hz, HFCs dramatically decreased in amplitude, while LFCs remained stable. Dissociated loss of HFCs with preserved LFCs was apparent under a high stimulus rate (Fig. 5). 4. Discussion 4.1. Cervical SEP The origins of cervical N11 and N13 were established as an ascending volley and cervical dorsal horn potentials, respectively (Austin and McCouch, 1955). Our previous study showed that the field distribution of spinal N13 in and around the cervical cord is analogous to that of post-synaptic activity in Rexed layers IV and V of the dorsal horn in animal study (Beall et al., 1977; Urasaki et al., 1990). This study, showing no amplitude reduction under a high (12.3 Hz) stimulus rate and latency shift along

Fig. 5. SEPs recorded from subdural electrodes implanted in a patient with intractable epilepsy. N20–P20 components and their corresponding HFCs are shown. HFCs were markedly decreased in amplitude despite the preserved N20–P20 at a higher stimulus rate.

164 the cervical cord, suggests that spinal HFCs reflect ascending activity without synaptic events. Another possibility is that HFCs may be activities interposed by a synaptic event, but a 12.3 Hz stimulus rate may not be sufficient to induce synaptic delay. This study at least indicated that cervical HFCs are more resistant to a change in stimulus rates than spinal LF N13. The origins of spinal HFCs have been suggested as population action potentials along different tracts (Maruyama et al., 1982), which might include the medial lemniscus, spinocerebellar tract, and others (Kuno et al., 1973; Jones et al., 1982; Beric et al., 1986; Halonen et al., 1989; Prestor et al., 1997), because spinal HFCs demonstrated multiple peaks with different conduction velocities (Figs. 1–3). Halonen et al. (1989) pointed out that spinocerebellar tract responses were elicited only when enough muscle afferents were activated, based on the presence of a probable spinocerebellar response when the tibial nerve was stimulated at the knee but the loss of that response when the nerve was stimulated at the ankle. Median nerve stimulation at the wrist in our study may, therefore, mainly activate the dorsal column pathway. A multiple firing pattern of the spinal dorsal horn has been reported (Holloway et al., 1976), but the dorsal horn spikes did not seem identical to spinal HFCs because spinal HFCs showed characteristics of an ascending volley and stronger resistance to a high stimulus rate than the N13 dorsal horn potential. However, doublet or triplet firing patterns in the dorsal horn may be linked to HFC multiplicity, when some HFCs are thought to reflect the same population action potentials. Further studies are required to determine whether cervical HFCs are correlated with a multiple firing pattern, as the burst firing mode in the dorsal horn was proposed as a unique coding method in the central nervous system (Holloway et al., 1976). 4.2. Thalamic SEP With regard to LFCs, we previously demonstrated large amplitude gradients of SEPs in the thalamus (Urasaki et al., 1989), that SEPs recorded below and above the thalamus have similar configurations and extrathalamic electrodes were not affected by the presence of

large-amplitude intrathalamic SEPs (Urasaki et al., 1993). These findings were reasonably interpreted as indicating that thalamic SEPs are present in the closed field and do not contribute to the formation of scalprecorded SEPs (Urasaki et al., 1989, 1993). The same appears to be true in HFCs of thalamic SEPs, as demonstrated in this study. Large residual thalamic HFCs did not project to cortical HFOs, similar to thalamic LFCs (Fig. 4). Our findings are in good agreement with a study of Klostermann et al. (2000a, b) who showed dissociated change between thalamic and cortical HFCs under propofol anesthesia or double-pulse stimulation. 4.3. Cortical SEP Selective loss of HFCs with preserved LF N20–P20 by a high stimulus rate (Fig. 5) suggests that major or some parts of cortical HFCs are activities interposed by a synaptic event after generation of a cortical response (Urasaki et al., 2002). In addition, the superficial terminal of thalamo-cortical fibers may be the origin of an early part of cortical HFCs, because we found a case showing that early HFCs before a peak of corresponding N20 were resistant to a high stimulus rate, while later HFCs after the N20 peak demonstrated an apparent amplitude reduction (Urasaki et al., 2002). Thalamo-cortical fiber activity and intracortical activity may be overlapped in cortical HFCs (Curio, 2000). References Austin, G.M. and McCouch, G.P. (1955) Presynaptic component of intermediary cord potential. J. Neurophysiol., 18: 441–451. Beall, J.E., Applebaum, A.E., Foreman, R.D. and Willis, W.D. (1977) Spinal cord potentials evoked by cutaneous afferents in monkey. J. Neurophysiol., 40: 199–211. Beric, A., Dimitrijevic, M.R., Prevec, T.S. and Sherwood, A.M. (1986) Epidurally recorded cervical somatosensory evoked potential in humans. Electroencephalogr. Clin. Neurophysiol., 65: 94–101. Curio, G. (2000) Linking 600-Hz “spike-like” EEG/MEG wavelets (“σ-bursts”) to cellular substrates. Concepts and caveats. J. Clin. Neurophysiol., 17: 377–396. Halonen, J.P., Jones, S.J., Edgar, M.A. and Ransford, A.O. (1989) Conduction properties of epidurally recorded spinal cord potentials following lower limb stimulation in man. Electroencephalogr. Clin. Neurophysiol., 74: 161–174.

165 Hanajima, R., Dostrovsky, J.O., Lozano, A.M., Hutchison, W.D., Davis, K.D., Chen, R. and Ashby, P. (2004) Somatosensory evoked potentials (SEPs) recorded from deep brain stimulation (DBS) electrodes in the thalamus and subthalamic nucleus. Clin. Neurophysiol., 115: 424–434. Holloway, J.A., Wright, L.E. and Trouth, C.O. (1976) Burst and doublet firing modes within spinal cord dorsal horn cells of the chicken (Gallus domesticus). Brain. Res., 117: 326–330. Jones, S.J., Edgar, M.A. and Ransford, A.O. (1982) Sensory nerve conduction in the human spinal cord: epidural recordings made during scoliosis surgery. J. Neurol. Neurosurg. Psychiatry, 45: 446–451. Klostermann, F., Funk, T., Vesper, J., Siedenberg, R. and Curio, G. (2000a) Double-pulse stimulation dissociates intrathalamic and cortical high-frequency (>400Hz) SEP components in man. Neuroreport, 11: 1295–1299. Klostermann, F., Funk, T., Vesper, J., Siedenberg, R. and Curio, G. (2000b) Propofol narcosis dissociates human intrathalamic and cortical high-frequency (>400Hz) SEP components. Neuroreport, 11: 2607–2610. Kuno, M., Munoz-Martinez, E.J. and Randic, M. (1973) Sensory inputs to neurons in clarke’s column from muscle, cutaneous and joint receptors. J. Physiol. Lond., 228: 327–342.

Maruyama, Y., Shimoji, K., Shimizu, H., Kuribayashi, H. and Fujioka, H. (1982) Human spinal cord potentials evoked by different sources of stimulation and conduction velocities along the cord. J. Neurophysiol., 48: 1098–1107. Prestor, B., Gnidovec, B. and Golob, P. (1997) Long sensory tracts (cuneate fascicle) in cervical somatosensory evoked potential after median nerve stimulation. Electroencephalogr. Clin. Neurophysiol., 104: 470–479. Urasaki, E., Wada, S., Kadoya, C., Yokota, A., Matsuoka, S. and Shima F. (1989) Origin of scalp far-field N18 of SSEPs in response to median nerve stimulation. Electroencephalogr. Clin. Neurophysiol., 77: 39–51. Urasaki, E., Wada, S., Kadoya, C., Yokota, A. and Matsuoka, S. (1990) Spinal intramedullary recording of human somatosensory evoked potentials. Electroencephalogr. Clin. Neurophysiol., 77: 233–236. Urasaki, E., Uematsu, S. and Lesser, R.P. (1993) Short latency somatosensory evoked potentials recorded around the human upper brain-stem. Electroencephalogr. Clin. Neurophysiol., 88: 92–104. Urasaki, E., Genmoto, T., Akamatsu, N., Wada, S. and Yokota, A. (2002) The effects of stimulus rates on high frequency oscillations of median nerve somatosensory-evoked potentials – direct recording study from the human cerebral cortex. Clin. Neurophysiol., 113: 1794–1797.


167

Chapter 23

The nature of facilitation of motor evoked potentials by heteronymous muscle contraction in the lower limb. Interactions between knee and ankle muscles Shin-Ichi Izumi* Department of Physical Medicine and Rehabilitation, Tohoku University Graduate School of Medicine, Sendai 980-8575 (Japan)

1. Introduction Facilitation is a keyword in neurological rehabilitation. Facilitation of motor evoked potentials (MEPs) following transcranial magnetic stimulation (TMS) is known as a phenomenon of amplitude increase or latency shortening induced by voluntary contraction or thinking about the movement of the target muscle (Izumi et al., 1995). Although both cortical and spinal mechanisms had been considered to produce MEP facilitation, the cortical mechanism could not be demonstrated until Di Lazzaro et al. (1998) showed that voluntary contraction increased I-waves recorded by epidural electrodes, and Abbruzzese et al. (1999) showed that imagery suppressed short latency intracortical inhibition using a paired stimulation paradigm. Planning of effective therapeutic exercise requires understanding of muscle interactions between joints,

*Correspondence to: Shin-Ichi Izumi, M.D., Ph.D., Department of Physical Medicine and Rehabilitation, Tohoku University Graduate School of Medicine, 2-1 Seiryo-cho, Aoba-ku, Sendai 980-8575, Japan. Tel: +81-22-717-7336; Fax: +8122-717-7340; E-mail: [email protected]

because even during efforts intended to represent isolated agonist contraction, a small amount of co-contraction of neighboring muscles may occur. Part of neural substrates for such interactions are overlaps of topographical representations in the motor cortex, and heteronymous Ia connections, such as those linking elbow muscles to motor neurons supplying wrist muscles (Cavallari et al., 1992). During co-contraction of antagonist muscles, various inhibitory mechanisms are regulated differently than during isolated agonist contraction (Nielsen and Kagamihara, 1992, 1993; Nielsen and Pierrot-Deseilligny, 1996). Information is incomplete regarding, which muscles are actually facilitated by contraction of a given muscle, and to what degree such facilitatory spread occurs in the lower extremity. To clarify these motor control issues, the H-reflex, MEP, and voluntary electromyographic (EMG) discharges all need to be examined, because neural substrates for these electrical activities differ from one another. Here, we present four experiments showing interactions of MEP, H-reflex, and voluntary discharges for the knee muscles and the ankle muscles of healthy volunteers. We investigated how isometric voluntary contraction of the ankle dorsiflexor and plantar flexor affect MEP amplitude in the target leg muscle

168 (experiment 1; Izumi et al., 1998), how voluntary isometric biceps femoris (BF) contraction affects MEP amplitude, background EMG amplitude, and H-reflex amplitude in ipsilateral leg muscles (experiment 2; Izumi et al., 2001), how voluntary isometric vastus medialis (VM) contraction affects MEP amplitude, background EMG amplitude in ipsilateral leg muscles (experiment 3; Furukawa et al., 2000), and how voluntary isometric contraction of either of tibialis anterior (TA) or soleus (SOL) muscle affects MEP amplitude, background EMG amplitude in the ipsilateral VM muscle (experiment 4; Furukawa et al., 1999). In each of the four experiments, subjects were seated on the edge of a bed with their hips and knees flexed at 90°, and the soles of their feet on the floor. MEP produced by TMS with an intensity 20% above the threshold and background EMG activities were recorded from the TA, SOL, or VM muscles at rest and

during voluntary contraction of the target muscle or an ipsilateral heteronymous muscle (Fig. 1). We chose sitting position without back support as the posture for the present study because this is a common starting position in therapeutic exercises for patients who cannot ambulate due to movement disorders. In experiments 1 and 2, H-reflexes were recorded from the SOL following electrical stimulation of the tibial nerve at the popliteal fossa. Stimulus intensity was adjusted to obtain an H-reflex with 50% maximum amplitude with the subject at rest without a preceding M response. The Friedman test was used to analyze differences in the MEP or H-reflex amplitudes under three conditions. The Wilcoxon signed-rank test was used to compare each pair. To standardize the degree of facilitation, we calculated the ratio of the MEP amplitude during voluntary contraction of the heteronymous or tested muscle

Experiment 1

KF/KE task

DF task stimulator coil

EMG biofeedback

Experiment 2 & 3

Background EMG

MEP

TA

TA SOL

SOL 100

0

TMS

Experiment 4

PF task VM BF SOL

VM-MEP

TA

DF task

SOL

PF task

TA

Fig. 1. Design of transcranial magnetic stimulation (TMS) studies in four experiments. Active electrodes (closed circles) and indifferent electrodes (open circles) were placed over the biceps femoris (BF), vastus medialis (VM), tibialis anterior (TA), and soleus (SOL) muscles. During voluntary contraction task, subjects were instructed to maintain tonic contraction of that muscle using electromyographic (EMG) biofeedback device. Motor evoked potentials (MEPs) and background EMG were recorded from the tested muscles. Data collection began 30 ms before TMS. DF – ankle dorsiflexion; PF – ankle plantar-flexion; KF – knee flexion; KE – knee extension.

169 to the amplitude at rest. The ratio of the background EMG amplitude during heteronymous muscle contraction to that measured during 10% MVC of the tested muscle was calculated to standardize the background EMG activity. 2. Experiment 1 There was a statistically significant difference in MEP amplitudes under the resting, ankle plantar-flexion (SOL contraction at 10% MVC) and dorsiflexion (TA contraction at 10% MVC) conditions. Voluntary contraction of the SOL muscle produced a strong facilitation of that muscle’s response to TMS (an 8.6-fold increase in the average) that was statistically significant. Voluntary contraction of the TA muscle produced a relatively weak MEP facilitation of the SOL muscle (a 4.9fold increase in the average) that was also statistically significant. There was also a statistically significant difference between the plantar flexion and the dorsiflexion conditions in MEP amplitude of the SOL muscle. The background EMG amplitudes of the SOL muscle recorded under the resting, plantar flexion, and dorsiflexion conditions were statistically significantly different. The maximal amplitude of the background EMG of the SOL muscle under the plantar-flexion condition was significantly greater than that recorded under the dorsiflexion and resting conditions, and that recorded under the dorsiflexion condition was significantly greater than that recorded under the resting condition. There was a statistically significant difference in MEP amplitudes under the resting, plantar-flexion, and dorsiflexion conditions. Voluntary contraction of the TA produced a strong facilitation of that muscle’s MEP amplitude (a 10.5-fold increase in the average) that was statistically significant. Voluntary contraction of the SOL muscle produced a relatively weak MEP facilitation of the TA muscle (a 3.7-fold increase in the average) and again, the facilitation was statistically significant. There was also a statistically significant difference between the dorsiflexion and the plantar-flexion conditions in TA MEP amplitude. There was a statistically significant difference in background EMG amplitude of the TA muscle under the resting, plantar-flexion, and dorsiflexion conditions.

The maximal amplitude of the background EMG of the TA muscle recorded under the dorsiflexion condition was significantly greater than that recorded in the plantar-flexion and resting conditions. The maximal amplitude of the background EMG of the TA muscle in the plantar-flexion condition was significantly greater than that in resting condition. There was no statistically significant difference between the two muscles in the degree of MEP facilitation produced during voluntary agonist contraction or voluntary antagonist contraction. The background EMG ratio of the SOL muscle under the dorsiflexion condition did not differ significantly from that of the TA muscle under the plantar-flexion condition. Voluntary contraction of the SOL muscle appeared to produce a facilitation of the H-reflex (a 2.1-fold increase in the average) that was statistically significant. Voluntary contraction of the TA muscle produced inhibition with an average decrease of 55%, which was also statistically significant. There was no statistically significant difference in the maximal amplitudes of the background EMG of the SOL muscle between TMS and H-reflex studies under dorsiflexion condition. The depression of the SOL H-reflex could be explained by spinal inhibitory mechanisms such as presynaptic inhibition exceeding facilitation by the background EMG. 3. Experiment 2 MEP produced by TMS with an intensity 20% above the threshold and background EMG activities were recorded from the TA and SOL muscles at rest, during BF contraction with 10% MVC, and during contraction of tested muscles with minimal, 5, 10, and 20% MVC. Simple regression analysis demonstrated that maximal background EMG amplitude and MEP amplitude ratio each correlated with %MVC for TA and SOL muscles. The regression intercept of background EMG amplitude vs. %MVC did not differ significantly from zero. No statistically significant difference in regression slope of the MEP facilitation was evident between the two muscles. Voluntary contraction of the BF muscle produced facilitation of the TA and SOL MEP, and co-contraction of the TA and SOL muscles. Significant differences

170 were observed in MEP amplitudes of the TA and SOL between at rest and during BF contraction. The maximal amplitude of the background EMG recorded during BF contraction was significantly greater than that seen during resting condition for both the TA and SOL muscles. Simple regression analysis demonstrated that maximal background EMG amplitude correlated with MEP amplitude for TA and SOL muscles both during BF contraction and during voluntary contraction of the tested agonist muscle. The regression equations for the relationship during agonist muscle contraction were MEP amplitude (mV) = 5.06 × background EMG (mV) + 1.85 for TA, and MEP amplitude = 7.22 × background EMG + 0.438 for SOL. Regression equations for the relationship during BF contraction were MEP amplitude (mV) = 13.5 × background EMG (mV) + 0.837 for the TA muscle, and MEP amplitude (mV) = 11.1 × background EMG (mV) + 0.179 for the SOL muscle. The slope of the regression line for the TA muscle during the knee flexion task was significantly steeper than during the ankle dorsiflexion task. In contrast, no statistically significant difference was noted in the slope for the SOL muscle between the knee flexion task and the ankle plantar-flexion task. Voluntary contraction of the BF muscle produced an increase of approximately 4.8-fold in TA MEP amplitude, and an increase of approximately 4.6-fold for SOL MEP amplitude. No statistically significant difference was apparent between the two muscles in the degree of facilitation produced during voluntary contraction of BF. The background EMG ratio of the SOL muscle during this condition was significantly greater than that of the TA muscle. Voluntary contraction of the BF muscle was associated with inhibition of the SOL H-reflex with a statistically significant mean decrease of 89%. The maximal amplitude of the SOL background EMG recorded during BF contraction was significantly greater than that seen at rest. The depression of the SOL H-reflex could be explained by spinal inhibitory mechanisms such as presynaptic inhibition originating from BF Ia-afferents (Nielsen and Petersen, 1994), exceeding facilitation by the background EMG.

4. Experiment 3 Significant differences were observed in MEP amplitudes of the TA and SOL between at rest and during VM contraction at 10% MVC. The maximal amplitude of the background EMG recorded during VM contraction at 10% MVC was significantly greater than that seen during resting condition for both the TA and SOL muscles. No statistically significant difference was apparent between the two muscles in the degree of MEP facilitation produced during voluntary contraction of VM. The background EMG ratio of the TA muscle during this condition was significantly greater than that of the SOL muscle. 5. Experiment 4 Voluntary contraction of each of TA or SOL muscle at 10% MVC enhances VM-MEP amplitude. MEP enhancement in the VM during isolated ankle dorsiflexion task was more prominent than that during an isolated ankle plantar-flexion task, while co-contraction of the VM during dorsiflexion did not differ from that during plantar flexion. 6. Functional significance Table 1 summarizes the results obtained from experiments 1 to 4. Both agonist and antagonist muscle contraction enhanced MEP amplitude of the TA and SOL muscles. There was no statistically significant difference between the two muscles in the degree of MEP facilitation during either of agonist or antagonist contraction. The degree of MEP facilitation during antagonistic muscle contraction was submaximal level, and the degree of background EMG was similar in the two muscles. Thus, we suggest that the effects of antagonist muscle contraction on the target muscle MEP are comparable between the two muscles. Voluntary knee extension and flexion produced MEP facilitation in TA and SOL muscles. No statistically significant difference was apparent between the two muscles in the degree of MEP facilitation for both tasks. In contrast, the co-contraction level of TA and

171 TABLE 1 INTERACTIONS AMONG KNEE AND ANKLE MUSCLES Task

Ankle-to-ankle Ankle dorsiflexion Ankle plantar-flexion Knee-to-ankle Knee extension Knee flexion

Ankle-to-knee Ankle dorsiflexion Ankle plantar-flexion

MEP facilitation

Co-contraction

H-reflex

TA

SOL

TA

SOL

++ +

+ ++

++ +

+ ++

Suppressed Enhanced

+ +

+ +

++ +

+ ++

Not examined Suppressed

VM-MEP facilitation

VM co-contraction

++ +

+ +

SOL

MEP – motor evoked potential; TA – tibialis anterior; SOL – soleus; VM – vastus medialis; + – present; ++ – more prominent than +.

SOL muscles differed and depended upon the task. Symmetry in net MEP facilitation during knee flexion/ extension task suggests that equilibrium between the TA and SOL would favor voluntary control of the ankle, making it less affected by a proximo-distal synergic pattern of movements. Patterns of co-contraction observed in this study appear to be appropriate for stand-up movement and ambulation. SOL co-contraction was greater than TA co-contraction during knee flexion, which is appropriate considering that voluntary knee flexion is an action that initiates rising from a sitting position, the BF acts as hip extensor with the sole touching the ground, and SOL is an antigravity muscle. In addition, TA co-contraction was greater than SOL co-contraction during knee extension, which is reasonable because intense co-contraction of the TA and quadriceps femoris muscles occurs at the heel strike of the gait cycle. We believe that these co-contraction patterns might have been acquired usedependently. Based upon the results of experiment 4, TA co-contraction would be useful for augmenting motor command for knee extension, because VM-MEP facilitation was greater during TA contraction than during

SOL contraction without difference in background EMG of the VM muscle. In conclusion, effects of heteronymous muscle contraction differ among MEP, background EMG, and H-reflex in the lower extremity. The relationship of MEP facilitation to background EMG varies task-dependently. Information on such heteronymous facilitation and inhibition should be useful in planning therapeutic exercises for patients with movement disorders. 7. Acknowledgements The author thanks Drs. Toshiaki Furukawa, Yuji Koyama, and Akira Ishida, Department of Rehabilitation Medicine, Tokai University School of Medicine for their collaboration in implementing the experiments shown in this study. References Abbruzzese, G., Assini, A., Buccolieri, A., Marchese, R. and Trompetto, C. (1999) Changes of intracortical inhibition during motor imagery in human. Neurosci. Lett., 263: 113–116. Cavallari, P., Katz, R. and Penicaud, A. (1992) Pattern of projections of group I afferents from elbow muscles to motoneurones

172 supplying wrist muscles in man. Exp. Brain Res., 91: 311–319. Di Lazzaro, V., Restuccia, D., Oliviero, A., Profice, P., Ferrara, L., Insola, A., Mazzone, P., Tonali, P. and Rothwell, J.C. (1998) Effects of voluntary contraction on descending volley evoked by transcranial stimulation in conscious humans. J. Physiol. (Lond.), 508: 625–633. Furukawa, T., Izumi, S.-I. and Ishida, A. (1999) Heteronymous facilitation: effects of voluntary ankle dorsal and plantar flexion on motor responses and myoelectric activities in the ipsilateral quadriceps femoris muscle. J. Clin. Rehabil., 8: 1218–1221 (Japanese). Furukawa, T., Izumi, S.-I. and Ishida, A. (2000) Facilitatory effects of ipsilateral quadriceps femoris voluntary contraction on motor responses and myoelectric activities in the tibialis anterior and soleus muscles. J. Clin. Rehabil., 9: 640–644 (Japanese). Izumi, S.-I., Findley, T.W., Ikai, T., Andrews, J., Daum, M. and Chino, N. (1995) Facilitatory effect of thinking about movement on motor evoked potentials to transcranial magnetic stimulation of the brain. Am. J. Phys. Med. Rehabil., 74: 207–213.

Izumi, S-I., Koyama, Y., Ishida, A., Arita, M. and Findley, T.W. (1998) Reciprocal inhibition in the leg during voluntary contraction. In: I. Hashimoto and R. Kakigi (Eds.), Recent Advances in Human Neurophysiology. Elsevier, Amsterdam, pp. 1005–1013. Izumi, S.-I., Furukawa, T., Koyama, Y. and Ishida, A. (2001) The nature of facilitation of leg muscle motor evoked potentials by knee flexion. Somatosens. Mot. Res., 18: 322–329. Nielsen, J. and Kagamihara, Y. (1992) The regulation of disynaptic reciprocal Ia inhibition during co-contraction of antagonistic muscles in man. J. Physiol. (Lond.), 456: 373–391. Nielsen, J. and Kagamihara, Y. (1993) The regulation of presynaptic inhibition during co-contraction of antagonistic muscles in man. J. Physiol. (Lond.), 464: 575–593. Nielsen, J. and Petersen, N. (1994) Is presynaptic inhibition distributed to corticospinal fibres in man? J. Physiol. (Lond.), 477: 47–58. Nielsen, J. and Pierrot-Deseilligny, E. (1996) Evidence of facilitation of soleus-coupled Renshaw cells during voluntary co-contraction of antagonistic ankle muscles in man. J. Physiol. (Lond.), 493: 603–611.


173

Chapter 24

Repetitive transcranial magnetic stimulation (rTMS) in monkeys Yoshikazu Ugawaa,*, Shingo Okabea, Takuya Hayashib, Takashi Ohnishic and Yukio Nonakad a

Department of Neurology, Division of Neuroscience, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655 (Japan) b Department of Investigative Radiology, Research Institute, National Cardiovascular Centre, 5-7-1 Fujishirodai, Suita, Osaka 565-8565 (Japan) c Department of Radiology, National Centre Hospital for Mental, Nervous, and Muscular Disorders, National Centre of Neurology and Psychiatry, 4-1-1 Ogawahigashi, Kodaira, Tokyo 187-8551 (Japan) d Nihon Kohden Corporation, Tokyo (Japan)

1. Introduction The combination of neuroimaging techniques with repetitive transcranial magnetic stimulation (rTMS) is a recently developed method for the functional analysis of brain function. To avoid epilepsy induction by rTMS or significant irradiation by PET tracers, we cannot perform these experiments repeatedly on the same human subject. However, to investigate the duration of long-term effects by rTMS or the best stimulation parameters for rTMS, we should perform many PET scans on the same subject or give several rTMSs to the same subject with different stimulation parameters. For such purposes, we sometimes use monkeys as a good human model, and we have therefore performed

*Correspondence to: Dr. Y. Ugawa, Department of Neurology, Division of Neuroscience, Graduate School of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8655, Japan. Tel: +81-3-5800-8672; Fax: +81-3-5800-6548; E-mail: [email protected]

several TMS experiments on monkeys. In this chapter, we will demonstrate how useful information is provided by experiments on monkey. 2. A coil for magnetic stimulation of the macaque monkey brain (Nonaka et al., 2003) To study brain function by TMS, it is necessary to focally stimulate a localized brain area. In humans, this focal stimulation has been partly accomplished using a figure-of-eight coil (Ueno et al., 1988). However, in a monkey brain within a small skull, a human brain coil cannot induce currents comparable to those elicited in the human brain. TMS is not able to induce sufficient current to activate neurons within small volume structures. For example, the spinal cord in the spinal canal can be activated by high voltage electrical stimulation (Ugawa et al., 1995) but cannot be activated by TMS, although the spinal roots are activated by TMS (Ugawa et al., 1989). In order to induce sufficient current to activate monkey brain neurons, we made a coil suitable for monkey brain stimulation. We compared the induced

174 currents in a model brain elicited by different coils and confirmed that focal stimulation is achieved in monkeys with our coil. 2.1. Methods We measured electric fields induced by single-pulsed TMS with three kinds of coil in a model brain: a coil specially developed for monkey brain stimulation, a small flat figure-of-eight coil and a figure-of-eight coil for the human brain. Magnetic stimulation was performed with a magnetic stimulator (AAA-15486, Nihon Kohden, Tokyo, Japan). We first made a plastic model of a macaque monkey skull based on magnetic resonance images (MRIs) of the skull and brain. We then made a small double cone coil (outer diameter of each coil, 62 mm) whose two coils were fixed to fit the curvature of the skull over the motor cortex (135⬚), a small flat figure-of-eight coil (outside diameter of each coil, 62 mm), similar to that often successfully used to activate the motor cortex in monkey experiments (Oliver et al., 2001), and a figureof-eight coil (outer diameter of each coil, 92 mm), which is usually used for focal cerebral stimulation in humans. These coils were placed over the motor cortex. The electric fields induced by TMS in the brain were measured using a probe made from a coaxial cable similar to those used in previous reports (Maccabee et al., 1991; Kobayashi et al., 1997). The coaxial cable was passed through an acrylic tube and connected to an amplifier. The distal 5 mm of the outside insulation and shield of the cable were stripped to record the voltage drop between the cable shield and the bared distal tip. The distal end of the probe was bent at right angles and submerged in the saline solution. By dividing the voltage drop by 5 mm, we calculated the induced electric field (mV/mm or V/m). The probe was placed in a skull model filled with isotonic saline. Measurement was performed at 54 sites, 1 cm apart. All sites were 5 mm deep from the inner surface of the skull at the level of the monkey cerebral cortex, judging from MRI images. All points were on a dome-shaped surface. At each site, we measured the voltage drop in two directions, antero-posterior and left–right. The amplitude and

direction of the vector by voltage drops measured in two directions were calculated at each point. The measured electric fields were depicted in a view of the dome-shaped inner surface of the skull. The vector amplitude was depicted in color. The center of the coil was placed over the left motor cortex. The intensity was fixed at 35% of the maximum stimulator output in all experiments. 2.2. Results Figure 1 shows the electric field maps of eddy currents in the model brain induced by a special coil for monkey brain stimulation over an MRI image of the monkey. High-level, moderately localized electric fields were evoked by our coil (Fig. 1) and were localized under the coil center. The highest amplitude was approximately 70 V/m just under the coil center. With a small flat figure-of-eight coil, electric fields were less localized and smaller compared with those evoked by the former coil. A maximum amplitude of 49 V/m was elicited under the coil center, which was about 65% of that evoked by the coil for the macaque monkey. With a human figure-of-eight coil, the electric field map was similar to that with a flat small figure-of-eight coil and the highest amplitude was 51 V/m (68% of that induced by the coil for macaque monkeys). 2.3. Discussion It is well known that TMS cannot induce sufficient current to activate neurons within a small volume structure. To identify the appropriate size and shape of the coil for focal stimulation of an animal brain, we investigated how induced currents are affected by different stimulating coils. The measurements of induced electric fields have shown that stronger, more localized fields were elicited with our monkey coil than with a flat, same-sized, figure-of-eight coil or with a larger coil for human brain stimulation. Focal stimulation was clearly achieved in the monkey brain using our coil, suggesting that focal activation can be optimally achieved using a coil that fits the animal skull tightly.

175

Fig. 1. Color map of induced currents in the monkey brain. The amplitude of induced currents by a coil for monkeys is shown in color in C of the monkey head constructed based on MRI images. The dotted lines indicate the magnetic coil for monkeys. Localized currents are induced under the coil center.

3. Long-term lasting effects of motor cortical rTMS (Hayashi et al., 2004) 3.1. Introduction rTMS has been regarded as having lasting effects on cortical neurons, such as transient enhancement or depression of cortical excitability (Chen et al., 1997). Several studies have used this to transiently modify regional cortical excitability and to localize cortical functions, and others have used it to treat several neuropsychiatric and neurological disorders. However, there is a lack of evidence demonstrating how long the effects continue. We have attempted to address the neurobiological effects of rTMS by examining temporal profiles in non-human primates. We performed rTMS on the precentral gyrus of the frontal lobe using a smallsized coil developed for monkey brains (Nonaka et al., 2003) and directly evaluated neuronal activity and functional connectivity changes using [18F]fluorodeoxyglucose (18F-FDG) and PET.

3.2. Methods Ten adult male cynomologous monkeys with a body weight of 4.9 ± 0.1 kg (mean ± S.E.M.) were used. Each animal received repeated 18F-FDG-PET scans at five time points: prior to the administration of rTMS (baseline), during rTMS (day 0), after rTMS at 1 day (day 1), 8 days (day 8), and 16 days (day 16) under anesthesia. The rTMS coil was placed over the primary motor cortex (M1). The anesthetic level and other physiological conditions were monitored with quantitative electroencephalography (EEG) or cardiac status over repeated experiments; and the spatial homogeneity of PET sensitivity was calibrated by a normalization scan obtained using an 18F-FDG-containing model once a week during the PET series. The study was performed within the guidelines for animal research on the Human Care and Use of Laboratory Animals (Rockville, National Institute of Health/Office for Protection from Research Risks, 1996) and approved by the ethical committee for animal research at the National Cardiovascular Center, Japan.

176 The PET scan was started after a waiting period of 2 hours to achieve a physiologically stable state after anesthesia. After each PET scan, anesthesia was withdrawn and the animal was housed in a thermo-controlled room. To enable PET data analysis in a standard space, we obtained high-resolution 3D T1-weighted MRIs (IR-FSPGR, TR = 9.4 ms, TE = 2.1 ms, TI = 600 ms) using a 3D MRI scanner (Signa LX VAH/I, GE, Milwaukee, USA) with the head of the animal fixed in a stereotaxic frame under anesthesia. The image was generated in a matrix of 256 × 256 × 60 (x, y, z) with a voxel size of 0.39 × 0.39 × 1.0 mm. MRI was performed 1–2 months before the PET scan series. For motor cortical stimulation, we used a smallsized double-cone coil (6 × 12 cm) developed for the brain of Macaca fascicularis, as described above. To fix the intensity of rTMS over the monkey brain, we tried to evaluate the lowest intensity that maintained sufficient electrophysiological action, and corresponded to the “active motor threshold (AT)” used in the human motor cortex (Nonaka et al., 2003). We adopted other stimulation parameters considered to be safely performed in humans: 20 trains of 5 Hz monophasic pulse for 20 s with an inter-train interval of 40 s (total of 2000 pulses). The coil was placed over M1, which was stereotactically determined using 3D-MRI imaging for each animal. PET scans were performed with an ECAT EXACT HR PET scanner (Siemens-CTI, Knoxville, USA). After a 15-min transmission scan for attenuation correction, we injected 187 MBq of 18F-FDG intravenously and acquired a 2D-mode PET scan with a midtime at 45 min postinjection in a dim quiet room. In the PET scan during rTMS, we started rTMS stimulation at the time of tracer injection, while at other time points, scans were performed in the same conditions without rTMS. To minimize inter-subject variability in brain shape and size and to allow voxel-based statistics across subjects, we transformed all PET data to a standard anterior commissure–posterior commissure (AC–PC) space for the Macaca fascicularis brain (Martin and Bowden, 2000). 3.3. Results Using voxel-wise analysis, we first estimated a linear contrast for the effect of conditions before (baseline)

and during rTMS (day 0) to specify rTMS-induced changes in regional activity. The result showed that during rTMS, activity increased in the contralateral anterior cingulate gyrus (ACC, Z = 4.83, corrected P < 0.01) and posterior cingulate gyrus (PCC, Z = 4.19, corrected P < 0.05) compared to the baseline. A subtle but significant decrease was seen at the motor/premotor cortex just beneath the coil (Z = −2.8, corrected P < 0.05), whereas the contralateral motor/premotor areas also showed decreases with larger and greater significance (Z = −3.26, corrected P < 0.01). The subcoil effect was clearly comparable to the localization of E measured in the model cranium. Next, we performed condition analysis at each time point after rTMS (days 1, 8, 16) to identify lasting effects of rTMS. At day 1 and day 8, activity increased in the contralateral orbitofrontal cortex (OFC) (Z = 4.22, corrected to P < 0.05 and Z = 4.78, corrected to P < 0.01 respectively) compared to the baseline, whereas no significant voxels were found at day 16 even at P = 0.3 (Fig. 2). When the significance level was increased to uncorrected P = 0.001, we also found additional activity increase or decrease in the ACC and M1 at day 1 and day 8 (Fig. 2). Thus, to identify any lasting effects, we performed value of information (VOI) analysis in four selected brain regions that showed significance in voxel-based analysis. We placed sphere VOIs in bilateral regions of OFC, M1, PCC, and ACC. In one-way analysis of variance (ANOVA), as a main effect of the subject with repeated measures of VOIs, a significant time effect was observed in the left OFC (F8, 4 = 9.08), left M1 (F8, 4 = 5.44), right ACC (F8, 4 = 4.89) and left and right PCC (F8, 4 = 4.44, 4.53) at corrected P < 0.05 (Bonferroni correction with the number of VOI). In these areas, post hoc comparison with Fisher’s PLSD showed lasting effects up to 8 days in the left OFC, left M1, and right ACC and up to 1 day in bilateral PCC. A lasting effect of rTMS for as long as 8 days was also supported by voxel-wise comparison of day 0, 1, and 8 time points to the baseline and day 16 time points, which showed significant increases in left OFC (x, y, z = −6, 16, 6.5, Z = 4.33, corrected to P < 0.05) and right ACC (x, y, z = 6, 6.5, 10, Z = 4.33, corrected to P < 0.05), and a decrease in left M1 (x, y, z = −16, 1, 12, Z = 3.01, P < 0.05 corrected by M1-VOI).

177

Fig. 2. Glucose metabolic changes on the 8th day after rTMS. Significant metabolic changes between the baseline before rTMS and the 8th day after rTMS are shown on MRIs of the monkey. Hot colors indicate metabolic increase induced by rTMS, and cold colors, metabolic decrease. Metabolic increase occurred at the orbitofrontal cortex (OFC) and anterior cingulate cortex and decreased at the bilateral primary sensorimotor cortices.

4.5 4 3.5 3 2.5

Ventral striatum

2 1.5 1 0.5 0

Fig. 3.

[11C]racroplide PET during rTMS. Areas showing [11C]racroplide binding capacity decrement are shown. The binding capacity was decreased by real rTMS compared with sham rTMS at the ventral striatum on both sides.

178 3.4. Discussion We have shown combined rTMS and repeated neuroimaging experiments in non-human primates. Such frequent assessments are not applicable to human studies because of the high dose of radiation exposure. Considering the localized electrical field induced with this coil, these findings of regional metabolic changes in multiple areas cannot be attributed to direct activation by rTMS. We have found that a single series of rTMS had a long-lasting effect up to 8 days in the primate brain and had a regional effect on local and distant areas with anatomical or functional connections with the area immediately under the coil. The effects of rTMS on regional activity surprisingly persisted for at least 8 days. It is well known that cerebral glucose metabolism correlates with neuronal activity, and more precisely, with excitatory glutamatergic synaptic activity (Chatton et al., 2003). In addition, macroscopic regional activity changes in neuroimaging techniques reflect the input and intracortical processing of a given area rather than its spiking output (Logothetis et al., 2001). Thus, it is reasonable to assume that lasting changes in glucose metabolism can be attributed to changes in neuronal excitability analogous to LTP or LTD. In rodents, long-term lasting changes in neuronal excitability have been found when direct electrical stimulations were delivered to the input pathway in the region of interest. The hippocampus is one of the most reactive regions in the brain, and an effect lasting as long as 3 weeks was found in the responsiveness of the hippocampus in awake rats when its input, i.e. perforant pathway was stimulated (Levkovitz et al., 1999). Even in the neocortex, which has been thought to be resistant to such modification, Trepel and Racine (1998) showed that LTP continued for 5 weeks in awake rats after electrical stimulation of the callosal pathway, provided that they were paced and repeated. In fact, rTMS in rodents also induced LTP-like, and more durable LTD-like changes in evoked spike rates in the neocortex (Wang et al., 1996). Although several issues remain to be clarified, the data presented here indicate that the neurobiological

effects of focal rTMS in the primate brain persistent in multiple distant brain regions. Motor rTMS induced activity changes not only in motor-related areas but also distant limbic-related areas via functional connections, suggesting that rTMS induces remodelling in a largescale functional network. 4. Endogenous dopamine release induced by rTMS over the primary motor cortex (Ohnishi et al., 2004) 4.1. Introduction rTMS has been applied in the treatment of Parkinson disease (PD). After many trials, at this time, it is unclear whether rTMS over the M1 has a beneficial effect on PD (Wassermann and Lisanby, 2001; Cantello et al., 2002; Okabe et al., 2003). The dopaminergic system is a candidate neurotransmitter system to explain therapeutic mechanisms. A human PET study and animal studies using microdialysis showed that rTMS over the frontal cortex has modulatory effects on the dopaminergic system (Strafella et al., 2001; Keck et al., 2002). The aim of this project was to study whether rTMS over M1 induces dopamine release in the striatum or other areas in the monkey. 4.2. Methods Young adult male cynomologous monkeys were used in this experiment. Animals were maintained and handled in accordance with guidelines for animal research on Human Care and the Use of Laboratory Animals (Rockville, National Institute of Health/Office for Protection from Research Risks, 1996). The study was approved by the ethical committee for animal research at the National Cardiovascular Center. Each animal underwent two [11C]raclopride PET scans under anesthesia, one after rTMS and the other after sham-sound stimulation as the control condition. The order of the two conditions was balanced among subjects. As in the previous experiment, rTMS consisted of 20 trains of 5 Hz rTMS for 20 s with an inter-train interval of 40 s. For the control condition, sham-sound stimulation consisted of sequential sounds of rTMS,

179 sampled from rTMS. The PET scan was started 5 min after the end of either stimulation. The same PET scanner and anesthesia were used as in the previous experiment. After waiting for 2 hours to achieve a physiologically stable state and to complete withdrawal from the effect of ketamine, animals were positioned in the PET scanner. A mass of 1–4 pmol/kg of [11C]raclopride (radioactivities of 166–370 MBq) was prepared based on the measured specific activity of the product. The mass of raclopride for each animal was adjusted to be as similar as possible between the two PET scans, and was not significantly different between conditions. Immediately after the end of rTMS, the tracer was injected intravenously. Data acquisition began at the onset of the tracer injection and continued for 60 min in 39 time frames of gradually increasing individual duration (10–300 s). The PET image of [11C]raclopride radioactivity was reconstructed by filtered back projection with a matrix of 128 × 128 × 47 and a voxel size of 1.1 × 1.1 × 3.13 mm. Voxel-wise images were analyzed using statistical parametric mapping (SPM99; Welcome Department of Cognitive Neurology, London, UK). PET images were summed and co-registered to the subject’s MRI using a mutual information algorithm (Ashburner and Friston, 1997). T1-weighted MRI images were then transformed to the standard brain space of Macaca fascicularis (Martin and Bowden, 2000). This transformation was applied to co-registered PET images and parametric images of binding potential (BP) of [11C]raclopride. The BP of [11C]raclopride was estimated in a voxel-wise parametric image method, based on a simple reference tissue model (Lammertsma and Hume, 1996). The change of BP in rTMS and sham stimulation was tested by voxel by voxel paired t-test. Statistical inferences were based on the theory of random gaussian field theory (Friston et al., 1995). 4.3. Results rTMS over the hand area of the right primary motor cortex (M1) decreased [11C]raclopride BP in the bilateral ventral striatum including the nucleus accumbens (NAc) compared with the sham condition (Fig. 3). Such a change is most likely due to an increase of

extracellular dopamine concentration in the ventral striatum after stimulation over M1. On the other hand, increased [11C]raclopride BP as compared with the sham condition was noted in the right putamen. The Wilcoxon test revealed a significant (P < 0.05) effect of stimulation for the bilateral NAc and the right posterior lateral part of the putamen, but not for other striatal regions. 4.4. Discussion We found that rTMS over the right M1 can induce dopamine release in the ventral striatum including the NAc in anesthetized monkeys. No significant reduction of BP was found in the dorsal striatum but a significant increase was observed in the right putamen. One advantage of this study was that we performed rTMS on anesthetized monkeys. Dynamic changes of the dopaminergic system, particularly the mesolimbic dopaminergic system, are susceptible to uncontrollable and/or subliminal mental status. Under anesthesia, such confounding effects should be negligible. As the anesthesia and physiological parameters were identical between real and sham rTMS in our experiments, we considered that observed changes in the dopaminergic system were not due to the generalized anesthesia. The ventral striatum, particularly NAc, is a target of the mesolimbic dopamine system, which arises in dopaminergic neurons in the ventral tegmental area (VTA) (Oades and Halliday, 1987). The NAc, and its dopaminergic inputs, play critical roles in rewards, reinforcement, and incentive motivation. Recently, alteration of the mesolimbic dopamine pathway has been suggested as involved in the pathogenesis of some major symptoms of depression, the pervasive absence of behavioral incentives, such as apathy, anhedonia, amotivation (Nestler et al., 2002). This suggests that activation of the ventral striatum by rTMS over M1 may increase non-specific motivation and have beneficial effects on depression and motor symptoms. The efficacy of rTMS for PD remains controversial. Modulatory effects on the VTA–NAc pathway (mesolimbic dopaminergic system) elicited by rTMS over M1 may improve the concomitant depressive status

180 in PD or increase motivation, and may lessen parkinsonian motor symptoms. Moreover, our data suggested that rTMS over M1 deactivated the mesostriatal dopaminergic pathway. This would possibly worsen the motor symptoms. The final outcome of rTMS on PD should be a summation of these two effects, non-specific increased motivation (the mesolimbic system) and possible decreased motor performance (mesostriatal system), which have significant inter-individual variability that may lead to controversial results. 5. Conclusion rTMS is a promising method for clinical neurology and brain research. Similar to other methodologies, increased knowledge about animals, especially monkeys, should help us to interpret phenomena occurring in humans. As shown above, the combination of rTMS and neuroimaging methods in monkeys provided important information in applying rTMS to humans although there are many issues remain to be solved before the results in monkeys can be applied to humans. References Ashburner, J. and Friston, K. (1997) Multimodal image coregistration and partitioning – a unified framework. Neuroimage, 6: 209–217. Cantello, R., Tarletti, R. and Civardi, C. (2002) Transcranial magnetic stimulation and Parkinson’s disease. Brain Res. Rev., 38: 309–327. Chatton, J.Y., Pellerin, L. and Magistretti, P.J. (2003) GABA uptake into astrocytes is not associated with significant metabolic cost: implications for brain imaging of inhibitory transmission. Proc. Natl. Acad. Sci. USA, 100: 12456–12461. Chen, R., Classen, J., Gerloff, C., Celnik P., Wassermann, E.M., Hallett, M. and Cohen, L.G. (1997) Depression of motor cortex excitability by low-frequency transcranial magnetic stimulation. Neurology, 48: 1398–1403. Friston, K.J., Holmes, A.P., Worsley, K.J., Poline, J.P., Frith, C.D. and Frackowiak, R.S.J. (1995) Statistical parametric maps in functional imaging: a general linear approach. Hum. Brain Mapp., 2: 189–210. Hayashi, T., Ohnishi, T., Okabe, S., Teramoto, N., Nonaka, Y., Watabe, H., Imabayashi, E., Ohta, Y., Jino, H., Ejima, N., Sawada, T., Iida, H., Matsuda, H. and Ugawa, Y. (2004) Motor cortical repetitive transcranial magnetic stimulation induces longterm lasting effects in brain regions with functional connections – a monkey PET study. Ann. Neurol., 56: 77–85. Keck, M.E., Welt, T., Muller, M.B., Erhardt, A., Ohl, F., Toschi, N., Holsboer, F. and Sillaber, I. (2002) Repetitive transcranial

magnetic stimulation increases the release of dopamine in the mesolimbic and mesostriatal system. Neuropharmacology, 43: 101–109. Kobayashi, M., Ueno, S. and Kurokawa, T. (1997) Importance of soft tissue inhomogeneity in magnetic peripheral nerve stimulation. Electroencephalogr. Clin. Neurophyiol., 105: 406–413. Lammertsma, A.A. and Hume, S.P. (1996) Simplified reference tissue model for PET receptor studies. Neuroimage, 4: 153–158. Levkovitz, Y., Marx, J., Grisaru, N. and Segal, M. (1999) Long-term effects of transcranial magnetic stimulation on hippocampal reactivity to afferent stimulation. J. Neurosci., 19: 3198–3203. Logothetis, N.K., Pauls, J., Augath, M., Trinath, T. and Oeltermann, A. (2001) Neurophysiological investigation of the basis of the fMRI signal. Nature, 412: 150–157. Maccabee, P.J., Amassian, V.E., Eberle, L.P., Rudell, A.P., Cracco, R.Q., Lai, K.S. and Somasundarum, M. (1991) Measurement of the electric field induced into inhomogeneous volume conductors by magnetic coils: application to human spinal neurogeometry. Electroencephalogr. Clin. Neurophyiol., 81: 224–237. Martin, R.F. and Bowden, D.M. (2000) Primate Brain Maps: Structure of the Macaque Brain. Elsevier Science, Amsterdam. Nestler, E.J., Barrot, M., DiLeone, R.J., Eisch, A.J., Gold, S.J. and Monteggia, L.M. (2002): Neurobiology of depression. Neuron, 34: 13–25. Nonaka, Y., Hayashi, T., Ohnishi, T., Okabe, S., Ueno, S., Mtsuda, H., Iida, H., Watabe, H. and Ugawa, Y. (2003) A coil for magnetic stimulation of the macaque monkey brain. In: W. Paulus, F. Tergau, M.A. Nische, J.C. Rothwell, U. Ziemann and M. Hallett (Eds.), Transcranial Magnetic Stimulation and Transcranial Direct Current Stimulation (Suppl. to Clin. Neurophysiol. Vol. 56). Elsevier, Amsterdam, pp. 75–80. Oades, R.D. and Halliday, G.M. (1987) Ventral tegmental (A10) system: neurobiology. 1. Anatomy and connectivity. Brain. Res., 434: 117–165. Ohnishi, T., Hayashi, T., Okabe, S., Nonaka, I., Matsuda, H., Iida, H., Imabayashi, E., Watabe, H., Miyake, Y., Ogawa, M., Teramoto, N., Ohta, Y., Ejima, N., Sawada, T. and Ugawa, Y. (2004) Endogenous dopamine release induced by repetitive transcranial magnetic stimulation over the primary motor cortex: an [11C]raclopride PET study in anesthetized macaque monkeys. Biol. Psychiatry, 55: 484–489. Okabe, S., Ugawa, Y. and Kanazawa, I. (2003) 0.2-Hz repetitive transcranial magnetic stimulation has no add-on effects as compared to a realistic sham stimulation in Parkinson’s disease. Mov. Disord., 18: 382–388. Oliver, E., Baker, S.N., Nakajima, K., Brocheir, T. and Lemon, R.N. (2001) Investigation into non-monosynaptic corticospinal excitation of macaque upper limb single motor units. J. Neurophysiol., 86: 1573–1586. Strafella, A.P., Paus, T., Barrett, J. and Dagher, A. (2001) Repetitive transcranial magnetic stimulation of the human prefrontal cortex induces dopamine release in the caudate nucleus. J. Neurosci., 21: RC157(1–4). Trepel, C. and Racine, R.J. (1998) Long-term potentiation in the neocortex of the adult, freely moving rat. Cereb. Cortex, 8: 719–729.

181 Ueno, S., Tashiro, T. and Harada, K. (1988) Localized stimulation of neural tissues in the brain by means of paired configuration of time-varying magnetic fields. J. Appl. Phys., 64: 5862–5864. Ugawa, Y., Rothwell, J.C., Day, B.L., Thompson, P.D. and Marsden, C.D. (1989) Magnetic stimulation over the spinal enlargements. J. Neurol. Neurosurg. Psychiatr., 52: 1025–1032. Ugawa, Y., Genba-Shimizu, K. and Kanazawa, I. (1995) Electrical stimulation of the human descending motor tracts at several levels. Can. J. Neurol. Sci., 22: 36–42.

Wang, H., Wang, X. and Scheich, H. (1996) LTD and LTP induced by transcranial magnetic stimulation in auditory cortex. Neuroreport, 7: 521–525. Wassermann, E.M. and Lisanby, S.H. (2001) Therapeutic application of repetitive transcranial magnetic stimulation: a review. Clin. Neurophysiol., 112: 1367–1377.


183

Chapter 25

The wavelet transformed EEG: a new method of trial-by-trial evaluation of saccade-related cortical activity Peter B. Forgacsa, Hans Von Gizyckib, Myroslav Harhulaa, Matt Avitableb, Ivan Selesnickc and Ivan Bodis-Wollnera,* a Department of Neurology, and Center for Scientific Computing, State University of New York, Downstate Medical Center, 450 Clarkson Avenue, Brooklyn, NY 11203-2098 (USA) c Department of Electrical and Computer Engineering, Polytechnic University, Brooklyn, NY 11201 (USA) b

1. Introduction To define temporal sequences in various sensory modalities prior to action is possible only with limitations even with the best event-related functional magnetic resonance imaging (fMRI) techniques. Traditional surface recorded electroencephalogram (EEG) signals have been used to discern pre-motor potentials and other slow wave phenomena (such as event-related potentials) in association with sensory decision and motor action. More recently, studies emerged which showed distinct functional role of synchronized oscillatory activity in the gamma range (30–100 Hz) of the human EEG. Gamma has been observed during a wide range of tasks since the original observations of (Gray and Singer, 1989). Although its functional significance is far from being fully understood, currently gamma oscillations around 40 Hz

*Correspondence to: Dr. Ivan Bodis-Wollner, Department of Neurology, SUNY Downstate Medical Center, 450 Clarkson Avenue, Box 1213, Brooklyn, NY 11203-2098, USA. Tel: (718) 270-1482; Fax: (718) 270-3840; E-mail: [email protected]

in humans are thought to represent synchronous activity of distributed neural networks dynamically linked for the same task. Gamma activity was demonstrated to be the underlying mechanism for feature “binding” and integration of visual information to a coherent object representation (Engel et al., 1991; Bertrand and Tallon-Baudry, 2000; Humphreys, 2003). Visual stimulus-induced gamma may be modulated by stimulation of the reticular activating system (Munk et al., 1996) or top-down processes (Tallon-Baudry et al., 1997). Recent studies have demonstrated the involvement of gamma activity in a variety of brain functions, such as attention (Fell et al., 2003), learning, memory (Gruber et al., 2001), motor control (Pfurtscheller et al., 1993; Salenius et al., 1996; Brown et al., 1998), and sensorimotor binding (Brown and Marsden, 1998), suggesting that gamma activity may reflect a general neural mechanism for synchronizing distributed cortical cell assemblies involved in specific types of information processing (Kaiser and Lutzenberger, 2003). Gamma activity can be demonstrated as a timelocked signal to a certain stimulus (so-called evoked gamma) or with a time jitter changing from trial to trial (induced gamma). The routine technique of revealing small signals is averaging. However, averaging the

184 response to a large number of trials can only reveal evoked gamma, but not induced gamma which may be linked to some internal processes without tight coupling to an external event. For detecting gamma linked to some internal decisions, single trial analysis offers a possible method. The topography of induced gamma is dependent on sensory modality and task, which supports the idea that induced gamma is a correlate of higher cognitive processes, as it reflects task-dependent synchronization of neural networks (Bertrand and Tallon-Baudry, 2000). The method we describe here allows the analysis of gamma changes during saccades in a trial-by-trial manner, providing fine temporal dissection of saccade-related quantification and detection of induced gamma. 2. Methods for analyzing short bursts of EEG gamma power Evaluation of the presence of high-frequency bands in addition to the classical beta and lower frequency bands of the EEG became possible with the age of fast computation and digital recording methods at high sampling rates. In particular, Fourier analysis allows the precise determination of the power and phase of different frequency bands. However, Fourier analysis requires an extended time period; therefore it is not suitable if the signal has time-varying frequency. It is possible to decrease the time period of the Fourier transform as we see in short-term Fourier transform (STFT), however due to the Heisenberg Uncertainty Principle applied to time-frequency information analysis, the smaller the time window the less the resolution of the frequency spectrum. Therefore Fourier transform is not applicable for the analysis of short bursts of high-frequency signals of the EEG. Wavelet transform (WT) (Addison, 2002) is able to decompose several seconds or only tens of milliseconds long EEG signals into orthogonal, equal bandwidth components. Discrete wavelet transform (DWT) uses a fix reference point in time; therefore it is applicable for descriptive quantification of evoked, “triggered” gamma frequency (Ademoglu et al., 1997; Bas¸ar et al., 1999). However, when we are looking for gamma activity in association with saccades,

we cannot a priori decide whether saccade onset, duration, or end will determine when gamma occurs. Continuous wavelet transform (CWT) uses a sliding window, without a fixed reference point. With CWT we can quantify and compare peak energy and peak energy timing of different frequency bands “detail functions.” CWT is the appropriate method for the analysis of induced gamma, not being biased to a discrete time event. In our experiments, the wavelet function we selected was the complex gaussian filter. For control purposes, to exclude the possible artifacts caused by particular wavelets the data were analyzed using other wavelets as well (simple gaussian, Morlet, and Coiflet II). The essential results were similar. To evaluate the power of the gamma frequency, the CWT coefficients at 38.4 Hz as a function of time were transformed using the Hilbert transform. The real part of the Hilbert transform creates an envelope function, giving better estimation of the gamma power. Within the appropriate perisaccadic time window an average of the real part of Hilbert transform was used to calculate the gamma power of each saccade for each subject. Advanced statistical analysis is needed to analyze the extent of the data obtained from a trial-by-trial analysis. Our results are based on detailed analysis of about 1200 individual saccades in 10 subjects. The statistical method we use is general linear mixed (GLM) model utilized for ANOVA analysis. It provides an estimation of the gamma power depending on every data point analyzed. 3. Experimental evidences for the existence of gamma power Is scalp-recorded gamma, an “artifact” of frequency analysis of the human EEG? What does it mean that a periodic signal “truly” exists? For instance, a squarewave function can be described by its amplitude vs. time course, or equivalently by its frequency spectrum, containing all the odd harmonics of the fundamental sine wave. Therefore higher frequencies truly exist in the signal if one describes a square wave or an edge in the frequency domain. Hence the question is rather, do the higher frequencies truly reflect brain processes or they are only artificial constructs of linear

185 decomposition of a periodic signal with rapid transitions? The question can be only answered if the frequency component identified has a biological, functional significance independent of other frequency components, e.g. linking different physiological functions to different frequency bands. 3.1. The dependence of gamma on parametric changes in simple visual stimuli Our group has shown (Tzelepi et al., 2000; BodisWollner et al., 2001) that the lower gamma band (18–28 Hz) in vision represents a foveally centered and spatially relatively broadly tuned mechanism reflecting paracentral retinotopy. The higher gamma band (around 37 Hz) is also foveally centered, but narrowly tuned and reflects purely foveal processing (Fig. 1). Therefore the argument that high-frequency components of the EEG indeed exist is based on the fact that different frequency bands show independent variance with visual input.

Fig. 1. A comparison of low gamma power (square symbols) and high gamma power (round symbols) recorded at the inion as the function of spatial frequency of stimulation. Both peak to 5.5 cpd but the bandwidth is narrower for high gamma power. Low gamma may represent two mechanisms, one tuned to 3.3 cpd and the other to 5.5 cpd. This physiologically relevant difference demonstrates between low and high gamma frequency in response to the same visual stimulus is not an artifact. (From Bodis-Wollner et al., 2001.)

3.2. The change of parieto-occipital gamma power dependent on the timing of voluntary saccades The existence of gamma power modulation in saccade-related cortical activity is plausible for several reasons. Given the average duration of a saccade (100–150 ms), lower frequencies, such as alpha or beta range, could not be well representing intrasaccadic changes of brain activity. The relative short duration of a saccade requires the involvement of higher frequencies. Gamma is a fast mechanism capable to mediate short-term functional neural connectivity changes during the saccade. Our results show gamma bursting starting 40–50 ms into the saccade, lasting until the new fixation is achieved. Figure 2

Fig. 2. Perisaccadic modulation of gamma power over the occipital and parietal cortices estimated by the statistical analysis for 10 subjects. Analysis was performed in 6 time windows. 2 time windows were used before the onset of the saccade and 2 time windows after saccade completion, each of them 150 ms long. Intrasaccadic time period was divided into 2 time windows at the middle of the saccade. Figure 3 shows the electrode positions used. Baseline gamma power is the highest in the occipital cortex, lower over the posteroparietal areas, and lowest in the antero-parietal cortex. Intrasaccadic gamma modulation is present over the posterior cortices, most prominently over the lateral-occipital (L34, R34) and parieto-temporal (LPT, RPT) cortices. Intrasaccadic gamma modulation shows an inverted U-shaped function peaking just before the new fixation. By 300 ms after the saccades completion gamma activity returns to baseline seen before the saccade, suggesting that intrasaccadic gamma power increase is real, it is not due to pre-, and postsaccadic gamma suppression.

186 shows the gamma power over the occipital and parietal cortices estimated by the statistical analysis for 10 subjects. Subjects executed saccades between 2 permanent markers subtending 40o in the light. 6 time windows were analyzed, from 300 ms before the onset of the saccade until 300 ms after saccade completion. Pre-, intra- and postsaccadic periods were divided into 2 time windows respectively. Figure 3 shows the electrode positions used. Baseline gamma power is the highest in the occipital cortex, lower over the posteroparietal areas, and lowest in the antero-parietal cortex. Intrasaccadic gamma modulation shows a inversed U-shaped function, peaking just before the new fixation (Bodis-Wollner et al., 2002). Intrasaccadic gamma modulation is the highest over the lateraloccipital and parieto-temporal cortices (Forgacs et al., 2004a). The 2 presaccadic time windows and the second postsaccadic time window is not different significantly, indicating that intrasaccadic gamma increase

A

is real; it is not due to pre- and postsaccadic gamma suppression. Intrasaccadic gamma modulation is present both in the light and when the observer is blindfolded (Bodis-Wollner et al., 2004) and codes saccade direction (Forgacs et al., 2004b). 3.3. Evidence that gamma is not an artifact of muscle activity during eye movements It is well known that eye movements are often accompanied by head and body movements involving neck muscle activity. Muscle activity may show up as low as 20 Hz in the power spectrum. To control the possibility of muscle activity contamination in perisaccadic EEG gamma modulation, we recorded electromyogram (EMG) in three subjects. EMG was recorded during eye movements over the sternocleidomastoid, temporal and nuchal muscles with the same type surface electrodes we used for the EEG. The amplifier

B

Fig. 3. Electrode positions on the scalp. (A) Vertical projection, 45° inclined toward nasion; (B) left profile. Z refers to midline position; L – left; R – right; P – parietal; T – temporal. Z5 – midline occipital electrode placed above the inion at 5% of the inion–nasion distance. L17, R17, L34, R34 – medial and lateral transverse occipital electrodes, numbers indicate the percentage of the distance between Z5 and the nasion. LP2, RP2 – posterior parietal electrodes placed between and above the medial and lateral occipital electrodes forming isosceles. LPT, RPT – parieto-temporal electrodes forming isosceles with lateral occipital and posterior parietal electrodes. LP1, RP1 – anterior parietal electrodes forming isosceles with posterior parietal and parieto-temporal electrodes. Reference electrode position was at the 63% of inion–nasion distance, ground electrode was placed at the forehead. Note that for control, occipital responses were derived simultaneously with the midfrontal and with a mastoid reference electrode to check that the occipital recording was truly active.

187 settings including the bandwidth were identical to the scalp recordings. For control purposes, EMG was also recorded during physiological movements of the muscles respectively (head turn, jaw closing, head bowing). During muscle movements the EMG showed physiological activity, hence the positions of electrodes and amplifier settings were appropriate for recording muscle activity. Frequency analysis of the EMG showed no gamma power increase during muscle activity. The EMG recordings during saccadic eye movements showed no muscle activity in any muscle group observed. The recordings underwent identical analysis in the same time windows as the EEG (see above). Average EMG gamma power was an order of magnitude lower in the EMG than in the EEG (Fig. 4). These findings are in concordance with previous studies evaluating the influence of EMG on gamma frequency (Sleigh et al., 2001). There was no evidence

Gamma power

1.5

1.0 Time window presac1 presac2

.5

intrasac1 intrasac2 postsac1

0.0

postsac2

EMG-Sternocleido EMG-Nuchal EMG-Temporal EEG-LPT

Fig. 4. Gamma power of EMG recordings (first 3 panels) and parieto-temporal EEG (4th panel) before, during, and after saccadic eye movements. The EMG was recorded during eye movements over the sternocleidomastoid, temporal and nuchal muscles with the same type surface electrodes we used for the EEG. The amplifier settings including the bandwidth were identical to the scalp recordings. EMG recordings during eye movements underwent identical analysis in the same time windows as the EEG. Average EMG gamma power was an order of magnitude lower in the EMG than in the EEG. There is no evidence for significant intrasaccadic increase of gamma power of the EMG over any of the recorded muscles.

of intrasaccadic modulation or lateralized gamma in the EMG over any of the muscles. 4. The functional significance of saccade-related gamma power modulation Binding of neuronal activity by a temporal code is one effective way the cortex integrates distributed activity (Gray and Singer, 1989; Engel and Singer, 2001). Gamma subserves cortical activity related to perception of closure and binding of visual features. Neurophysiological studies suggest that gamma reflects lateral interneuronal functional connectivity changes. Therefore our results can be interpreted as short-term (20–60 ms) activity-dependent plasticity in neuronal connections during the saccade, reflected in gamma power modulation. Here, we discuss the possible functional significance of intrasaccadic gamma power modulation based on its properties. The shape of gamma power modulation suggests that it is not the electrophysiological expression of voluntary decision to move the eye. It would be too late to achieve cancellation once the saccade started. Similarly, intrasaccadic gamma cannot be the direct correlate of preparatory motor signals, reflecting processes associated with breaking the fixation or postsaccadic perception. The fact that the hemispheric distribution of gamma power depends on saccade direction suggests that it is not a simple correlate of non-specific alertness or diffuse attention. More importantly, intrasaccadic gamma power modulation is similar when saccades were executed in the light and when the subjects were blindfolded. This finding makes it evident that parietooccipital gamma power during the saccades is not the direct reflection of the changing actual visual scene. We propose that the role of intrasaccadic gamma is to mediate functional plasticity in neural connectivity for updating visual representations (Nakamura and Colby, 2002). Current theories suggest that representations of the visual scene (presumably a salience map or reference frame) may contribute to transsaccadic visual memory and consequently the perception of a stable visual world across saccades (Duhamel et al., 1992; Nakamura and Colby, 2002). These representations are not direct correlates of the actual visual input;

188 therefore they could serve as an internal anchor in planning, initiation, and correct execution of voluntary saccades in the light and in the dark as well. Our results suggest that posterior cortical gamma participates or indexes sensorimotor (or motor-sensory) online information processing associated with saccades. There are several possible candidate mechanisms which may be indexed during saccades. One, it is possible that gamma is the result of the corollary discharge during the movement. Second, gamma represents a signal of the “internal monitoring of saccades” (Sommer and Wurtz, 2002). These possibilities are not mutually exclusive however, and in either case gamma may serve the purpose of preparing the visual cortex for the intended locus of fixation. This speculation insinuates gamma as an online mechanism during the saccade providing advance information about the upcoming visual information. Prediction about the target position and identifying it in the shortest possible time does have advantages. Current double saccade experiments suggest that information about the trajectory of the second saccade is already present before the first saccade is initiated (Caspi et al., 2004). Information about the intended locus of the target gives the possibility to the brain to act without waiting for visual feedback processes, when a considerable time, about 100 ms would be lost. An online, extraretinal mechanism would be efficient to detect deviations on saccade trajectory without losing this time; therefore the next movement can be initiated in a shorter reaction time. “Preemptive vision” gives parallel information for the motor system to act before visual feedback is available. Gamma has all the properties to be this mechanism. Irrespective of these speculations however, the properties of parieto-occipital gamma activity following a simple visual stimuli or in association with saccades demonstrate the “true” existence of gamma range activity in the human EEG. The evidence is the strong physiological dependence of high-frequency gamma different from concurrently analyzed lower frequencies. References Addison, P. (2002) The Illustrated Wavelet Transform Handbook. Institute of Physics Publishing, Bristol and Philadelphia.

Ademoglu, A., Micheli-Tzanakou, E. and Istefanopulos, Y. (1997) Analysis of pattern reversal visual evoked potentials (PRVEPs) by spline wavelets. IEEE Trans. Biomed. Eng., 44: 881–890. Bas¸ar, E., Bagar-Eroglu, C., Karakas, S. and Schurmann, M. (1999) Are cognitive processes manifested in event-related gamma, alpha, theta and delta oscillations in the EEG? Neurosci. Lett., 259: 165–168. Bertrand, O. and Tallon-Baudry, C. (2000) Oscillatory gamma activity in humans: a possible role for object representation. Int. J. Psychophysiol., 38: 211–223. Bodis-Wollner, I., Davis, J., Tzelepi, A. and Bezerianos, T. (2001) Wavelet transform of the EEG reveals differences in low and high gamma responses to elementary visual stimuli. Clin. Electroencephalogr., 32: 139–144. Bodis-Wollner, I., Von Gizycki, H., Avitable, M., Hussain, Z., Javeid, A., Habib, A., Raza, A. and Sabet, M. (2002) Perisaccadic occipital EEG changes quantified with wavelet analysis. Ann. N.Y. Acad. Sci., 956: 464–467. Bodis-Wollner, I., Forgacs, P., Von Gizycki, H., Avitable, M., Amassian, V., Selesnick, I. and Syed, N. (2004) A comparison of posterior cortical gamma in man when saccades are executed to visual targets and without visual targets in the dark. Invest. Ophthalmol. Vis. Sci., 45: E-Abstract 2516. Brown, P. and Marsden, C.D. (1998) What do the basal ganglia do? Lancet, 351: 1801–1804. Brown, P., Salenius, S., Rothwell, J.C. and Hari, R. (1998) Cortical correlate of the piper rhythm in humans. J. Neurophysiol., 80: 2911–2917. Caspi, A., Beutter, B.R. and Eckstein, M.P. (2004) The time course of visual information accrual guiding eye movement decisions. Proc. Natl. Acad. Sci. USA, 101: 13086–13090. Duhamel, J.R., Colby, C.L. and Goldberg, M.E. (1992) The updating of the representation of visual space in parietal cortex by intended eye movements. Science, 255: 90–92. Engel, A.K. and Singer, W. (2001) Temporal binding and the neural correlates of sensory awareness. Trends Cogn. Sci., 5: 16–25. Engel, A.K., Kreiter, A.K., Konig, P. and Singer, W. (1991) Synchronization of oscillatory neuronal responses between striate and extrastriate visual cortical areas of the cat. Proc. Natl. Acad. Sci. USA, 88: 6048–6052. Fell, J., Fernandez, G., Klaver, P., Elger, C.E. and Fries, P. (2003) Is synchronized neuronal gamma activity relevant for selective attention? Brain Res. Rev., 42: 265–272. Forgacs, P.B., Syed, N., Von Gizycki, H., Avitable, M. and BodisWollner, I. (2004a) Time course of occipital gamma range EEG during voluntary saccades. Clin. Neurophysiol., 115: 250. Forgacs, P.B., Von Gizycki, H., Syed, N., Avitable, M., Amassian, V. and Bodis-Wollner, I. (2004b) Intrasaccadic gamma EEG in blindfolded subjects: Saccade-direction-dependent hemisperic differences. Perception, 33: 145–146. Gray, C.M. and Singer, W. (1989) Stimulus-specific neuronal oscillations in orientation columns of cat visual cortex. Proc. Natl. Acad. Sci. USA, 86: 1698–1702. Gruber, T., Keil, A. and Muller, M.M. (2001) Modulation of induced gamma band responses and phase synchrony in a paired associate learning task in the human EEG. Neurosci. Lett., 316: 29–32.

189 Humphreys, G.W. (2003) Conscious visual representations built from multiple binding processes: evidence from neuropsychology. Prog. Brain Res., 142: 243–255. Kaiser, J. and Lutzenberger, W. (2003) Induced gamma-band activity and human brain function. Neuroscientist, 9: 475–484. Munk, M.H., Roelfsema, P.R., Konig, P., Engel, A.K. and Singer, W. (1996) Role of reticular activation in the modulation of intracortical synchronization. Science, 272: 271–274. Nakamura, K. and Colby, C.L. (2002) Updating of the visual representation in monkey striate and extrastriate cortex during saccades. Proc. Natl. Acad. Sci. USA, 99: 4026–4031. Pfurtscheller, G., Neuper, C. and Kalcher, J. (1993) 40-Hz oscillations during motor behavior in man. Neurosci. Lett., 164: 179–182. Salenius, S., Salmelin, R., Neuper, C., Pfurtscheller, G. and Hari, R. (1996) Human cortical 40 Hz rhythm is closely related to EMG rhythmicity. Neurosci. Lett., 213: 75–78.

Sleigh, J.W., Steyn-Ross, D.A., Steyn-Ross, M.L., Williams, M.L. and Smith, P. (2001) Comparison of changes in electroencephalographic measures during induction of general anaesthesia: influence of the gamma frequency band and electromyogram signal. Br. J. Anaesth., 86: 50–58. Sommer, M.A. and Wurtz, R.H. (2002) A pathway in primate brain for internal monitoring of movements. Science, 296: 1480–1482. Tallon-Baudry, C., Bertrand, O., Delpuech, C. and Permier, J. (1997) Oscillatory gamma-band (30–70 Hz) activity induced by a visual search task in humans. J. Neurosci., 17: 722–734. Tzelepi, A., Bezerianos, T. and Bodis-Wollner, I. (2000) Functional properties of sub-bands of oscillatory brain waves to pattern visual stimulation in man. Clin. Neurophysiol., 111: 259–269.


191

Chapter 26

Functional mapping in epilepsy patients’ information from subdural electrodes Ronald P. Lesser* Department of Neurology and Neurosurgery, Johns Hopkins University, Baltimore, MD 21287-7247 (USA)

The standard concepts regarding the organization of the cortex were developed in the nineteenth and twentieth centuries and well summarized in the writings of Wilder Penfield and his collaborators. In brief, this model describes motor cortex anterior to, and sensory cortex posterior to, the rolandic sulcus, with the topographic localization represented by a homunculus. Schematically, the appearance of this is of a body more or less upside down, with the legs at the top, the trunk and arms closer to the sylvian fissure, and the face, mouth, and throat closer still. In this model, there are two separate primary language areas in the language-dominant hemisphere, an anterior area located in the frontal operculum, a posterior area in the region of the temporal–parietal–occipital junction, and an area near the supplementary motor area that is pertinent for vocalization but not language per se. This model has stood us well, but contemporary developments have enhanced and expanded our understanding of how language is organized in the brain. This review will focus on several aspects of this. First, however, we should review how cortical stimulation is used to explore function. We use platinum–iridium *Correspondence to: Ronald P. Lesser, M.D., 2-147 Meyer Building, Johns Hopkins University, 600 N Wolfe Street, Baltimore, MD 21287-7247, USA. E-mail: [email protected]

electrodes, with 2.3 mm surface areas exposed to the cortex, and imbedded in silastic. Other centers use stainless steel electrodes, or electrodes of different diameters and exposure sizes. In our center, the metal of the electrodes is about 0.1 mm below the surface plane of the silastic. The subdural electrodes come in a variety of arrays, from 1 × 4 strips (4 electrodes total), to 8 × 8 grids (64 electrodes total). The centers of the electrodes are 1 cm apart. Electrodes are implanted in the operating room, with the patient under general anesthesia. After recovery they come to the epilepsy monitoring unit. A main purpose of this is to record seizures, and interictal epileptiform activity. The second purpose is to localize motor, sensory, language, and other areas important for patient function. Patients are tested while comfortable in their hospital room. We test for several hours at a time. In our center, cortical stimulation is performed using alternating-polarity, 0.3 ms duration, square-wave pulses. We deliver 50 pulses/s, in trains routinely lasting from 2 to 5 s, but sometimes lasting less or more time than this. We start at a low intensity and gradually increase, so as to find the level that produces functional changes at a given site, and also so as to avoid after discharges, and possible stimulation induced seizures (Lesser et al., 1984a). Function can be tested in one of several ways. The simplest is to ask the patient to do nothing. One then

192 sees if a functional change occurs in response to stimulation. For example, if the patient does nothing during stimulation of the motor cortex, movement occurs. During stimulation of the sensory cortex, tingling occurs. The second way is to ask the patient to perform a task continuously, and see if this performance is altered during stimulation. For example, a patient can be asked to continuously wiggle the fingers, or hold the arms up and outstretched. With stimulation of a region adjacent to the primary motor strip, sometimes called the negative motor area (Lüders et al., 1988), the fingers stop wiggling or the arms drop. In the language area the patient might be asked to count, or read a passage, or speak continuously. One looks to see if performance is altered during stimulation. The third way is to begin stimulation and then assess function with probes. For example, during stimulation, a series of words could be spoken, or shown. Again, one looks to see if performance is altered during stimulation (Lesser et al., 1987). These kinds of testing often confirm the traditional functional–anatomic teachings: Broca’s area is in the frontal lobe, Wernicke’s in the temporal lobe (Lesser et al., 1984b, 1986). These tests also confirm information that exists in the literature, but which often had been overlooked by mainstream neurology. For example, stimulation within “Broca’s area” can alter comprehension (Schaffler et al., 1993). It finally has expanded our understanding of localization as this has existed in the literature. For example, Wernicke and others had speculated on the presence of an area in the inferior temporal lobe important for verbal function. Stimulation demonstrated the presence of an often extensive language region in the base of the temporal lobe (Lüders et al., 1986). This review will concentrate on three areas of the functional organization of the brain, giving examples of findings based upon studies with subdural electrodes: (1) Small regions of the brain can have very specific functions; (2) Small regions of the brain can support multiple functions; (3) Multiple regions may be needed for specific functions.

1. Specificity of function The standard homoncular maps imply that different functions map different parts of the brain. The idea that this might be so is not new. However, it is important to emphasize how specific this kind of functional representation can be. To give one example, Hart et al. (2004) reported on a patient whose subdural electrodes included coverage over the posterior temporal lobe and in whom category-specific impairment was found at a single site just inferior to the traditional posterior temporal language area (i.e. Wernicke’s area). In this patient, verbal size representation was impaired with stimulation, but size representation was normal when stimulation was not applied. Specifically, with stimulation the patient could not answer questions such as “Is a bee bigger than a house?” However, if pictures of two objects were shown, the patient could accurately assess their relative sizes in real life. Other measures of verbal and visual comprehension, for the categories of color, shape, orientation, movement, texture, and structure, were normal. The findings in this patient were interpreted as follows. Localized stimulation of a small region in the posterior temporal lobe caused impaired access between the visual and verbal systems. This impairment suggested that small brain regions can have restricted properties (that can be impaired by stimulation), and supported the idea that categorical modularity may be an organizing principle of cortical organization. 2. Multiplicity of function In addition, however, stimulation of a site can produce impairment of more than one function. For example, stimulation of a single electrode, or of a pair of adjacent electrodes, over the perirolandic region can alter motor or sensory function for more than one part of the body. Figure 1 gives an example of this. The pictures at each site indicate the functions impaired at that site. A second feature noticeable in this figure is that the overall distribution of responses more or less conforms to what might be predicted from the traditional homunculus, but the third feature to notice is that the details of this organization can vary considerably.

193

Fig. 1. The effects of stimulation in a patient with a history of intractable simple partial seizures affecting the left arm. The letters indicate the body part affected during localization testing. Upper case letters indicate that movement occurred in response to stimulation. Lower case letters indicate that inhibition of ongoing alternating movements occurred in response to stimulation. The presence of an asterix indicates that stimulation produced a sensation; this was usually a tingling sensation. In general stimulation occurred between adjacent electrodes, and the letters are placed between the electrodes that were stimulated. In some cases, stimulation occurred between a single electrode in the sensorimotor area and an electrode at which there were no effects with stimulation. Darkly shadowed circles indicate electrodes at which ictal onset patterns were seen; lightly shaded circles indicate electrodes to which there was spread of the ictal patterns.

The fourth feature that is noticeable is that a specific function can be represented at more than one electrode site. The distribution of responses varies widely among individuals. The literature indicates that this kind of variation of distribution is a common occurrence (Uematsu et al., 1992a, b). This finding is not restricted to stimulation. Kunieda et al. (2004) showed that cortical evoked potentials at single cortical sites occur in response to movement of different body sites (eye closure, lip pursing, shoulder abduction, middle finger extension, thumb abduction, foot dorsiflexion).

They found premovement potential shifts could occur at some electrodes regardless of the area of the body that moved. These electrodes were within or rostral to electrodes that identified the primary motor area, and often were adjacent to electrodes identified as within the negative motor area. These findings suggest that restricted sites could have widespread body representations, and most likely have widespread connections with other areas affecting movement. This finding is not restricted to the perirolandic cortex. For example, Morris et al. (1984) described

194 language changes in response to stimulation of the posterior temporal lobe and the temporal–parietal– occipital junction. They tested writing, calculation, finger recognition, right–left orientation, as well as naming, reading, constructional ability, spontaneous conversation, and spelling. They found that, at two electrodes, located over the angular gyrus, stimulation would impair the first four of these functions, consistent with the idea that Gerstmann’s syndrome might be due to impairment of a restricted portion of the cortex. However, at other electrodes, stimulation could impair elements of Gerstmann’s syndrome, together with other language functions. This would be consistent with the idea that Gerstmann’s syndrome could be due to more widespread impairment of the brain. Matsuhashi et al. (2004) recorded evoked responses at the temporal–parietal–occipital junction, using median and tibial nerve sensory evoked responses, responses to passive movement and pain, auditory evoked responses, and visual responses evoked by motion. Here again they found electrodes at which multiple modalities were affected by stimulation. 3. Multiplicity of functional representation The standard maps of the brain suggest that there are discrete areas for specific function and that the location of these areas is predictable. While representations may be discrete, they are not necessarily unique. In summary, a careful review of the results of cortical stimulation suggests that the representations of body parts in the sensory and motor cortices can vary among individuals (Uematsu et al., 1992a, c). Analysis of the results of stimulation also indicates this. Further confirmation of this idea comes from eventrelated spectral analysis (Crone et al., 1998a, b). With this technique, ongoing EEG is recorded while the patient performs motor, sensory, cognitive or other tasks, responding to stimuli. The EEG for a period of time before and after a stimulus is recorded digitally, using a referential montage. For the studies just referenced, the EEG was filtered as follows: 1–100 Hz, 6 dB per octave, 60 Hz notch. In order to assess the background variability of the EEG, including variation in arousal and attentiveness, we compared the

electrocorticography (ECoG) during the activation task to an ECoG sample during non-text conditions. EEG was segmented into 200 ms epochs, with a 100 ms overlap between epochs. Each epoch had a fixed time relationship to the time of stimulus onset. Fast Fourier transform was used to estimate power in specific frequency bands. We found that maximum power changes more or less conformed to the traditional mapping of the motor cortex, with leg area superior, hand area intermediate, and tongue area inferior. Nonetheless, somatotopic distribution of power changes during motor activation tasks showed considerable overlap among body parts. These data suggest that the functions we assessed may be distributed broadly in the cortex, rather than being topographically organized with restricted areas of representation for specific body parts or functions. Similarly, Matsumoto et al. (2004) stimulated electrodes in the language area with brief electrical pulses. They obtained responses from electrodes located in separate language areas. These observations give direct support for the idea that there are connections between cortical areas separated in space. In summary, our concept of how functions are represented in the human cortex is evolving. It continues to appear that specific regions exert a high degree of control over specific functions, but this control is not unique. Rather, multiple areas appear to cooperate towards common goals. Still to be defined is what the time course of this cooperation might be. We know, from evoked potential studies and studies of eventrelated depolarization and synchronization that the intensity of responses varies over time. We assume that these variations reflect the time course of involvement of a particular region in accomplishing the tested function. We are beginning to understand exactly how a specific region varies in its control of a function over time, and how regions cooperate in accomplishing the completed function (Georgopoulos, 2002; Crowe et al., 2004; Fortes et al., 2004; Naselaris et al., 2004). References Crone, N.E., Miglioretti, D.L., Gordon, B., Sieracki, J.M., Wilson, M.T., Uematsu, S. et al. (1998a) Functional mapping of human sensorimotor cortex with electrocorticographic spectral

195 analysis. I. Alpha and beta event-related desynchronization. Brain, 121: 2271–2299. Crone, N.E., Miglioretti, D.L., Gordon, B. and Lesser, R.P. (1998b) Functional mapping of human sensorimotor cortex with electrocorticographic spectral analysis. II. Event-related synchronization in the gamma band. Brain, 121: 2301–2315. Crowe, D.A., Chafee, M.V., Averbeck, B.B. and Georgopoulos, A.P. (2004) Participation of primary motor cortical neurons in a distributed network during maze solution: representation of spatial parameters and time-course comparison with parietal area 7a. Exp. Brain Res., 158: 28–34. Fortes, A.F., Merchant, H. and Georgopoulos, A.P. (2004) Comparative and categorical spatial judgments in the monkey: “high” and “low”. Anim. Cogn., 7: 101–108. Georgopoulos, A.P. (2002) Cognitive motor control: spatial and temporal aspects. Curr. Opin. Neurobiol., 12: 678–683. Hart, J.J., Lesser, R.P. and Gordon, B. (2004) Selective interference with the representation of size in the human by direct cortical electrical stimulation. J. Cogn. Neurosci., 4: 337–343. Kunieda, T., Ikeda, A., Ohara, S., Matsumoto, R., Taki, W., Hashimoto, N. et al. (2004) Role of lateral non-primary motor cortex in humans as revealed by epicortical recording of Bereitschafts potentials. Exp. Brain Res., 156: 135–148. Lesser, R.P., Lüders, H., Klem, G., Dinner, D.S., Morris, H.H. and Hahn, J. (1984a) Cortical afterdischarge and functional response thresholds: results of extraoperative testing. Epilepsia, 25: 615–621. Lesser, R.P, Lüders, H., Dinner, D.S., Hahn, J.F. and Cohen, L. (1984b) The location of speech and writing functions in the frontal language area. Results of extraoperative cortical stimulation. Brain, 107: 275–291. Lesser, R.P., Lüders, H., Dinner, D.S., Klem, G., Hahn, J.F. and Harrison, M. (1986) Electrical stimulation of Wernicke’s area interferes with comprehension. Neurology, 36: 658–663. Lesser, R.P., Lüders, H., Klem, G., Dinner, D.S., Morris, H.H. and Hahn, J.F. et al. (1987) Extraoperative cortical functional localization in patients with epilepsy. J. Clin. Neurophysiol., 4: 27–53.

Lüders, H., Lesser, R.P., Hahn, J., Dinner, D.S., Morris, H., Resor, S. et al. (1986) Basal temporal language area demonstrated by electrical stimulation. Neurology, 36: 505–510. Lüders, H., Lesser, R.P., Dinner, D.S., Morris, H.H., Wyllie, E. and Godoy, J. (1988) Localization of cortical function: new information from extraoperative monitoring of patients with epilepsy. Epilepsia, 29(Suppl. 2): S56–S65. Matsuhashi, M., Ikeda, A., Ohara, S., Matsumoto, R., Yamamoto, J., Takayama, M. et al. (2004) Multisensory convergence at human temporo-parietal junction – epicortical recording of evoked responses. Clin. Neurophysiol., 115: 1145–1160. Matsumoto, R., Nair, D.R., LaPresto, E., Najm, I., Bingaman, W., Shibasaki, H. et al. (2004) Functional connectivity in the human language system: a cortico-cortical evoked potential study. Brain, 127: 2316–2330. Morris, H.H., Lüders, H., Lesser, R.P., Dinner, D.S. and Hahn, J.F. (1984) Transient neuropsychological abnormalities (including Gerstmann’s syndrome) during cortical stimulation. Neurology, 34: 877–883. Naselaris, T., Merchant, H., Amirikian, B. and Georgopoulos, A.P. (2005) Spatial reconstruction of trajectories of an array of recording microelectrodes. J. Neurophysiol., 93: 2318–2330. Schaffler, L., Lüders, H., Dinner, D., Lesser, R. and Chelune, G. (1993) Comprehension deficits elicited by electrical stimulation of Broca’s area. Brain, 116: 695–715. Uematsu, S., Lesser, R.P. and Gordon, B. (1992a) Localization of sensorimotor cortex: the influence of Sherrington and Cushing on the modern concept. Neurosurgery, 30: 904–912 (discussion 912–913). Uematsu, S., Lesser, R.P., Fisher, R.S., Gordon, B., Hara, K., Krauss, G.L. et al. (1992b) Motor and sensory cortex in humans: topography studied with chronic subdural stimulation. Neurosurgery, 31: 59–72. Uematsu, S., Lesser, R., Fisher, R.S., Gordon, B., Hara, K., Krauss, G.L. et al. (1992c) Motor and sensory cortex in humans: topography studied with chronic subdural stimulation (see comments). Neurosurgery, 31: 59–71 (discussion 71–72).


197

Chapter 27

Are the processes reflected by late and ultra-late laser evoked potentials specific of nociception? A. Mourauxa,* and L. Plaghkib a

Laboratoire de Neurophysiologie (NEFY), Université Catholique de Louvain, B-1200 Brussels (Belgium) b Unité de Réadaptation (READ),Université Catholique de Louvain, B-1200 Brussels (Belgium)

1. Introduction Both in fundamental and clinical research, cutaneous heat stimulation is the most frequently used form of natural nociceptive stimulation. Allowing selective and synchronous activation of Aδ- and C-fibers, CO2 laser heat stimulators have been extensively used to study electrophysiological brain responses elicited by phasic activations of nociceptors (see Chen et al., 1998 for a review). Albeit it is nociceptive specificity, brief laser stimuli directed to a non-glabrous area of the skin (e.g. dorsum of the hand) do not necessarily evoke a painful sensation (Bromm and Treede, 1984). Indeed, at stimulus intensities slightly above absolute detection threshold, perception is dominated by warmth and touch-like sensations which are detected with latencies above 800 ms. At higher intensities, the stimulus evokes a well-localized, predominantly painful and pricking sensation (i.e. “first pain”) which is detected with much shorter reaction times (~350 ms).

*Correspondence to: Dr. André Mouraux, Laboratoire de Neurophysiologie (NEFY), 54 Avenue Hippocrate, B-1200 Brussels, Belgium. Tel: +32(2)764.5449; Fax: +32(2)764.5465; E-mail: [email protected]

Most often, it is followed by a second, diffuse and long-lasting, burning sensation referred to as “second pain.” The dual nature of these perceptual responses may be interpreted as resulting from the activation of two distinct groups of nociceptive afferents (Lewis and Ponchin, 1937; Willis, 1985; Price, 1988). One has a low heat threshold (~40°C), a low conduction velocity (~1 m/s), and is related to non-myelinated C-fibers. The other has a high heat threshold (~46°C), a faster conduction velocity (~10 m/s), and is related to small myelinated Aδ-fibers. This dual sensation is most consistently evoked by brief (e.g. ≤ 50 ms), large (e.g. ≥ 80 mm2), and intense (e.g. ≥ 8 mJ/mm2) laser stimuli, producing very steep heating ramps. Indeed, under these conditions, and despite the fact that the heat threshold of Aδ-nociceptors is higher than that of C-nociceptors, the stimulus activates all polymodal nociceptors quasi-simultaneously. Due to the difference between Aδ- and C-fiber conduction velocities, Aδ-nociceptive input activates cortical projections well before C-nociceptive input (see left panel of Fig. 1). Depending on peripheral conduction distance, the difference in arrival time of both afferent volleys may thus vary from ~0.1 s (stimulation of the trigeminal area) to more than 1 s (stimulation of the foot dorsum).

198

Fig. 1. Left panel: the difference in conduction velocity of Aδ- and C-fibers explains why subjects most often report a double sensation (first and second pain). Indeed, the Aδ-fiber afferent volley arrives much faster at central projection sites than the slower C-fiber afferent volley. Reaction times reflect the latency of detection of the Aδ-fiber-mediated sensation. Right panel: an ischemic Aδ-fiber block of the superficial radial nerve was used to selectively activate C-fibers. Subjects report the disappearance of first pain but persistence of a delayed second pain sensation. Reaction times reflect the latency of detection of slowly conducting C-fibers. (Adapted from Nahra and Plaghki, 2003.)

When heating the skin with such a stimulus, laser evoked brain potentials (LEPs) reveal components whose latencies are compatible with the conduction velocity of Aδ-fibers (“late LEP”: ~160–390 ms). Although subjects clearly report the perception of both Aδ-fiber-related first pain and C-fiber-related second pain, no evoked potentials are recorded at latencies compatible with the conduction time of C-fibers. At first, the absence of C-fiber-related ultralate components was assumed to result from poor synchronization of the evoked afferent volley (Bromm and Treede, 1987; Arendt-Nielsen, 1990). However, avoiding the concomitant activation of Aδ-fibers does not only lead to the disappearance of first pain and its electrophysiological correlate (see right panel of Fig. 1). It also leads to the appearance of an ultra-late LEP whose latency (~750–1150 ms) is compatible with the

arrival time of C-fiber input (Bromm et al., 1983; Bragard et al., 1996; Magerl et al., 1999). To explain why C-fiber input elicits an ultra-late LEP only when concomitant activation of Aδ-nociceptors is avoided, several hypotheses have been proposed. The aim of this review was to outline these different hypotheses and their consequences regarding the functional significance of both late and ultra-late LEPs. Furthermore, an alternate hypothesis will be presented, challenging the nociceptive specificity which is often ascribed to LEPs. 2. Non-stationarity of the C-fiber afferent volley Microneurographic recordings have shown that the conduction velocity of C-fibers can vary greatly and moreover, is reduced by repetitive activation (Valbo et al.,

199 1979). Several investigators have therefore proposed that the poorly synchronous nature of C-fibers could explain why their activation does not elicit consistent ultra-late LEP components (Bromm and Treede, 1987; Arendt-Nielsen, 1990). Indeed, it is likely that an important latency jitter of C-fiber input would prevent conventional time-domain averaging from revealing C-fiber-related brain potentials. Latency-correction algorithms were therefore applied in an attempt to circumvent these limitations. However, applying these methods to the analysis of stimuli concurrently activating Aδ- and C-fibers did not reveal significant C-fiber-related brain potentials. In addition to evoked potentials, stimuli may induce transient enhancements or attenuations of ongoing EEG oscillations which may be revealed by methods based on the time-frequency decomposition of EEG epochs (for a review, see Lopes da Silva and Pfurtscheller, 1999). Using such methods, several studies have shown that laser stimuli may induce prolonged modulations of the EEG frequency spectrum (Arendt-Nielsen, 1990; Ploner et al., 2002). As these long-lasting cortical responses spread across the time window compatible with the conduction time of C-fibers, it was proposed that they may reflect central processing of C-fiber input. A recent study (Mouraux et al., 2003) compared EEG spectrum modulations induced by concomitant activation of Aδ- and C-fibers with that induced by selective activation of C-fibers. Results showed that the long-lasting EEG changes produced by co-activation of Aδ- and C-fibers were induced by Aδ-fibers and not by C-fibers. 3. Ad-fiber-mediated spinal inhibition of C-fiber afferent transmission Several investigators have proposed that Aδ-fiberinduced inhibition of C-fiber spinal transmission could explain why C-fiber activation does not elicit ultralate brain responses when Aδ-fibers are concomitantly activated. This hypothesis was based on studies showing that repetitive stimulation of peripheral nerves in primates could produce a sustained central inhibition of spinal transmission (Chung et al., 1984). Furthermore, these studies have shown that the greatest inhibition of C-fiber transmission was produced by Aδ-fibers.

Several behavioral studies in humans, using long duration tonic heat stimuli, have indeed shown that blocking the peripheral transmission of Aδ-fibers not only completely suppressed the first pain sensation but also consistently increased the second pain sensation (Landau and Bishop, 1953; Sinclair and Stokes, 1964; Price et al., 1977). However, if the C-fiber afferent volley is blocked at the spinal level, how is it that the laser stimulus clearly leads to the perception of both first pain and second pain? Consequently, one may question the ability of transient Aδ-fiber afferent volleys produced by brief laser heat stimuli to induce a similar inhibition of C-fiber spinal transmission. In fact, psychophysical studies using brief laser stimuli have failed to replicate the enhancement of second pain after Aδ-fiber blockade (Chakour et al., 1996; Nahra and Plaghki, 2003). 4. Refractoriness of LEP cortical generators The morphological and topographical similarities between Aδ- and C-fiber-related LEPs have been emphasized by several investigators (Bromm et al., 1983; Bragard et al., 1996; Magerl et al., 1999; Opsommer et al., 2001a). The most prominent components of both late and ultra-late LEP responses consist of a large negative–positive complex (N2–P2) whose maximum is recorded at the vertex (see Fig. 2). A smaller negativity, labeled N1, may precede the late N2 negativity (Treede et al., 1988; Valeriani et al., 1996). The topographical scalp distribution of this earlier component is maximal at contralateral temporal leads. In a recent study (Valeriani et al., 2002), it was shown that selective activation of C-fibers could also elicit an early negative component. Such as the Aδ-fiber evoked N1, it displays a contralateral temporal topography. In accordance with these observations, several studies have shown that similar dipole configurations could adequately explain both Aδ- and C-fiber-related LEP components (Opsommer et al., 2001b; Kakigi et al., 2003). Bilateral suprasylvian and anterior cingulate cortical regions have been consistently identified as generators of these responses (see García-Larrea et al., 2003 for a review). In addition to sharing similar morphologies and topographies, late and ultra-late LEP components appear to be equally

200

Fig. 2. Grand average of laser evoked potentials (LEPs) recorded in 9 subjects before and after applying an ischemic Aδ-fiber block to the superficial radial nerve. Four different stimulus intensities, ranging from 5.8 to 10.6 mJ/mm2 were used (labeled 1–4). L-LEP – the time window within which Aδ-fiber-related late LEP components are usually recorded after stimulation of the hand (160–390 ms). U-LEP – the time window within which C-fiber-related ultra-late LEP components are usually recorded (750–1150 ms). Note that unlike the amplitude of the late LEP recorded in the control condition, the amplitude of the ultra-late LEP recorded in the Aδ-fiber block condition was mostly uncorrelated with stimulus intensity. (Adapted from Nahra and Plaghki, 2003.)

modulated by attentional factors such as the subjects’ experimental surroundings, general level of arousal, and focus of attention (Qiu et al., 2002; Valeriani et al., 2002; Opsommer et al., 2003). There is, therefore, converging evidence indicating that both responses reflect the activation of common generators. To explain why C-fiber activation appears to evoke an ultra-late LEP only when avoiding the concomitant activation of Aδ-fibers, it was thus often hypothesized that LEP generators were subject to refractoriness (Bromm and Treede, 1987; Treede et al., 1988; Magerl et al., 1999;

Ploner et al., 2002; Kakigi et al., 2003). Consequently, when Aδ-fibers trigger a late LEP, refractoriness of LEP generators would explain why the later arriving C-fiber afferent volley does not elicit an ultra-late LEP. The “refractoriness hypothesis” was tested in a recent study (Mouraux et al., 2004), using two consecutively applied laser stimuli. The delay between the onset of the first and the second stimulus was set such that the second stimulus produced an Aδ-fiber afferent volley arriving at LEP cortical generators during their hypothesized refractory period. For inter-stimulus delays as short as 280 ms, results showed that the Aδ-fiber afferent volley produced by the second stimulus evoked a late LEP whose morphology, latency, amplitude, and topography was unaltered by the occurrence of the preceding stimulus (see Fig. 3). This study therefore clearly showed that the neural populations which generate N2 and P2 components can be reactivated as soon as 280 ms after a first activation. Consequently, refractoriness of LEP generators cannot explain why C-fibers do not elicit an ultra-late LEP when both Aδand C-fiber nociceptors are concomitantly activated. These results contrast with those of several studies reporting that repetition of the laser stimulus induces a significant reduction of LEP amplitude (Bromm and Treede, 1987; Raij et al., 2003; Truini et al., 2004). In particular the study of Truini et al. (2004), using a similar double stimulation paradigm, showed that the amplitude of the second stimulus could be significantly reduced at inter-stimulus intervals ranging from 250 to approximately 1000 ms. However, as in this study the inter-stimulus interval was kept constant across successive stimulus pairs, the expectancy of the second stimulus was likely to be greater than that of the first stimulus. The difference in amplitude of both LEP responses may therefore have been related to differences in stimulus expectancy. Indeed, studies have shown that the amplitude of late event-related vertex potentials decreases with increasing stimulus expectancy (Fruhstorfer, 1971; Bourbon et al., 1987). 5. What and where is the ultra-late LEP? It is commonly accepted that Aδ- and C-fibers mediate different qualities of pain perception (Bromm and

201

Fig. 3. Grand average of laser evoked potentials (LEPs) recorded in eleven subjects. Stimulus intensity was supraliminal for Aδ-nociceptors. To test the refractoriness of LEP components, trials consisted of either two consecutively applied stimuli (SOA-X) or of a single applied stimulus (SINGLE). Stimulus onset asynchrony was 280, 600, 1100, or 2100 ms. Trials were presented in random order and equal probability. Subjects were asked to rate the intensity of both stimuli on two visualanalogue scales displayed side-by-side. At SOA-280, LEP components elicited by the first and second stimulus overlapped. Therefore, LEP components elicited by the first stimulus were cancelled out by subtracting the SINGLE waveform (dashed curve). S1 – onset of the first stimulus. S2 – onset of the second stimulus. Odd and even trials were averaged separately (grey curves). (Adapted from Mouraux et al., 2004.)

202 Treede, 1984; Willis, 1985; Price, 1996). Aδ-fibers appear to convey sensations which are mostly implicated in phasic pain. The faster conduction velocity of Aδ-fibers agrees with the assumption that they could serve primarily as early warning signals. In contrast, slow unmyelinated C-fibers induce sensations which appear more implicated in tonic pain and are generally associated with protective behaviors. A number of psychophysical studies have compared the perception of nociceptive heat before and after applying an ischemic nerve conduction block selectively affecting myelinated Aδ-fibers. Most studies using tonic heat stimuli of long duration and relatively large surface areas showed that C-fibers were the main mediators of heat pain (Landau and Bishop, 1953; Sinclair and Stokes, 1964; Price et al., 1977). In contrast, studies using laser heat stimuli of much shorter duration and much smaller surface areas showed that C-fibers contributed much less to the sensations evoked by these more phasic stimuli (Nahra and Plaghki, 2003). This phenomenon could be related to C-fiber afferents requiring substantial temporal and spatial summation to produce consistent perceptions. Given the sparse contribution of C-fibers to sensations evoked by brief laser stimuli, it is not surprising that the amplitude of the ultra-late LEP evoked by selective activation of C-fibers is usually of lower amplitude than that of the late LEP evoked by Aδ-fibers. Indeed, a positive correlation between intensity of the laser stimulus, intensity of perception, and LEP amplitude has been repeatedly described (Carmon et al., 1978; Kakigi et al., 1989). Furthermore, studies examining laser evoked responses under different attentional settings have shown that the amplitude of LEPs is more closely related to subjective pain sensation than to the actual stimulus intensity (Plaghki et al., 1994; García-Larrea et al., 1997). The intrinsically nonphasic nature of C-fiber sensory input might therefore explain why ultra-late LEPs are recorded only under specific conditions which are fulfilled, for instance, when C-fibers are activated in isolation. In most laser stimulation paradigms, subjects are asked to pay close attention to the upcoming stimulus. Therefore, when methods are used to selectively activate C-fibers, attention is entirely focused on the

isolated sensation of second pain. When the laser stimulus concomitantly activates Aδ- and C-fibers, even though subjects report the sensation of both first and second pain, attention may be mostly focused on the more salient sensation of first pain. It may therefore be hypothesized that a necessary condition for C-fibers to elicit an ultra-late LEP would be that attention be entirely focused on that specific “sensory channel.” In accordance with this hypothesis, Bromm and Treede (1985) showed results suggesting that when attention was shifted from first to second pain, C-fiber activation elicited an ultra-late LEP even when Aδ-fibers were concomitantly activated. However, this observation should be interpreted with caution as it was not replicated, concerned with a single trained subject, and relied on the use of latency-correction algorithms. It is often assumed that LEPs reflect processes directly related to the perception of noxious stimuli. However, this assumption does not account for the fact that C-nociceptor activation produces a sensation of second pain both in the presence and absence of concomitant Aδ-nociceptor activation but elicits an ultra-late LEP only when activated in isolation. Indeed, this dissociation implies that the processes underlying ultra-late LEPs are not required to perceive C-fiber-mediated sensations. The fact that co-activating Aβ- and C-fibers using transcutaneous electrical stimuli of high intensity produces a clearly painful sensation without evoking LEP-like components at latencies compatible with the conduction velocity of Aδ-fibers, suggests the existence of a similar dissociation between perception and electrophysiological responses produced by Aδ-fiber stimulation (Boulu et al., 1985; De Broucker and Willer, 1985). As pointed out by a number of studies, the laser evoked N2–P2 complex shares many similarities with the late vertex potentials elicited by other sensory modalities (Kunde and Treede, 1993; García-Larrea et al., 1997). Indeed, both electrophysiological responses exhibit a strikingly similar morphology and topography, and seem equally sensitive to attentional factors. Therefore, it could well be that the greater part of LEPs reflects activity common to the processing of sensory input in general. It has been proposed that LEP-related processes could serve to trigger

203 involuntary reorientations of attention. In accordance with this view, it was shown that at least the later part of the P2 component was connected to processes of attentional switching (Legrain et al., 2003). Therefore, these processes should best be triggered by sudden and salient changes in the ongoing stream of sensory input. This would explain why C-fiber input, unlike the more phasic and salient Aδ-fiber input, appears to struggle at reaching the threshold required to trigger LEP responses. 6. Conclusion Activation of Aδ- and C-fiber nociceptors produces a sensation of first pain and second pain without necessarily evoking late and ultra-late LEP responses. This observation indicates that the processes underlying LEPs may not be directly related to the perception of nociceptive input. Occurrence and amplitude of LEPs seem strongly conditioned by the salience of the evoking afferent input, suggesting that LEPs reflect processes related to mechanisms of involuntary attentional capture. The inherently non-phasic property of C-fiber input may therefore explain why a particular setting seems required for C-fibers to elicit LEP responses. The morphological, topographical, and behavioral similarities between N2 and P2 LEP components and the late endogenous vertex potentials which can be evoked by any kind of sensory stimulation suggests that these “vertex” components reflect non-specific processes common to all sensory modalities. In other words, the nociceptive specificity of LEPs would be a consequence of the fact that they are evoked by a stimulus which selectively activates receptors projecting onto the extralemniscal pathways involved in nociception. References Arendt-Nielsen, L. (1990) Second pain event related potentials to argon laser stimuli: recording and quantification. J. Neurol. Neurosurg. Psychiatr., 53: 405–410. Boulu, P., De Broucker, T., Maitre, P., Meunier, S. and Willer, J.C. (1985) Somatosensory evoked potential and pain. I. Late cortical responses obtained at different levels of stimulation. Rev. EEG Neurophysiol. Clin., 15: 19–25.

Bourbon, W.T., Will, K.W., Gary Jr., H.E. and Papanicolaou, A.C. (1987) Habituation of auditory event-related potentials: a comparison of self-initiated and automated stimulus trains. Electroencephalogr. Clin. Neurophysiol., 66: 160–166. Bragard, D., Chen, A.C. and Plaghki, L. (1996) Direct isolation of ultra-late (C-fibre) evoked brain potentials by CO2 laser stimulation of tiny cutaneous surface areas in man. Neurosci. Lett., 209: 81–84. Bromm, B. and Treede, R.D. (1984) Nerve fibre discharges, cerebral potentials and sensations induced by CO2 laser stimulation. Hum. Neurobiol., 3: 33–40. Bromm, B. and Treede, R.D. (1985) Evoked cerebral potential changes accompanying attention shifts between first and second pain. Electroencephalogr. Clin. Neurophysiol., 61: S117. Bromm, B. and Treede, R.D. (1987) Human cerebral potentials evoked by CO2 laser stimuli causing pain. Exp. Brain Res., 67: 153–162. Bromm, B., Neitzel, H., Tecklenburg, A. and Treede, R.D. (1983) Evoked cerebral potential correlates of C-fibre activity in man. Neurosci. Lett., 43: 109–114. Carmon, A., Dotan, Y. and Sarne, Y. (1978) Correlation of subjective pain experience with cerebral evoked responses to noxious thermal stimulations. Exp. Brain Res., 33: 445–453. Chakour, M.C., Gibson, S.J., Bradbeer, M. and Helme, R.D. (1996) The effect of age on Aδ- and C-fibre thermal pain perception. Pain, 64: 143–152. Chen, A.C., Arendt-Nielsen, L. and Plaghki, L. (1998) Laserevoked potentials in human pain: I. use and possible misuse. Pain Forum, 7: 174–184. Chung, J.M., Lee, K.H., Hori, Y., Endo, K. and Willis, W.D. (1984) Factors influencing peripheral nerve stimulation produced inhibition of primate spinothalamic tract cells. Pain, 19: 277–293. De Broucker, T. and Willer, J.C. (1985) Comparative study of the nociceptive reflex and late components of the evoked somatosensory potential during stimulation of the sural nerve in healthy subjects. Rev. EEG Neurophysiol. Clin., 15: 149–153. Fruhstorfer, H. (1971) Habituation and dishabituation of the human vertex response. Electroencephalogr. Clin. Neurophysiol., 30: 306–312. García-Larrea, L., Peyron, R., Laurent, B. and Mauguière, F. (1997) Association and dissociation between laser-evoked potentials and pain perception. Neuroreport, 8: 3785–3789. García-Larrea, L., Frot, M. and Valeriani, M. (2003) Brain generators of laser-evoked potentials: from dipoles to functional significance. Neurophysiol. Clin., 33: 279–292. Kakigi, R., Shibasaki, H. and Ikeda, A. (1989) Pain-related somatosensory evoked potentials following CO2 laser stimulation in man. Electroencephalogr. Clin. Neurophysiol., 74: 139–146. Kakigi, R., Tran, T.D., Qiu, Y., Wang, X., Nguyen, T.B., Inui, K., Watanabe, S. and Hoshiyama, M. (2003) Cerebral responses following stimulation of unmyelinated C-fibers in humans: electro- and magneto-encephalographic study. Neurosci. Res., 45: 255–275. Kunde, V. and Treede, R.D. (1993) Topography of middle-latency somatosensory evoked potentials following painful laser stimuli and non-painful electrical stimuli. Electroencephalogr. Clin. Neurophysiol., 88: 280–289.

204 Landau, W. and Bishop, G.H. (1953) Pain from dermal, periosteal, and fascial endings and from inflammation; electrophysiological study employing differential nerve blocks. AMA Arch. Neurol. Psychiatry, 69: 490–504. Legrain, V., Guérit, J.M., Bruyer, R. and Plaghki, L. (2003) Electrophysiological correlates of attentional orientation in humans to strong intensity deviant nociceptive stimuli, inside and outside the focus of spatial attention. Neurosci. Lett., 339: 107–110. Lewis, T. and Ponchin, E.E. (1937) The double pain response of the human skin to a single stimulus. Clin. Sci., 3: 67–76. Lopes da Silva, F.H. and Pfurtscheller, G. (1999) Basic concepts on EEG synchronization and desynchronization. Event-related desynchronization and related oscillatory phenomena of the brain. In: F.H. Lopes da Silva (Ed.), Handbook of Electroencephalography and Clinical Neurophysiology, Revised Edition. Elsevier, Amsterdam, pp. 3–11. Magerl, W., Ali, Z., Ellrich, J., Meyer, R.A. and Treede, R.D. (1999) C- and Aδ-fiber components of heat-evoked cerebral potentials in healthy human subjects. Pain, 82: 127–137. Mouraux, A., Guérit, J.M. and Plaghki, L. (2003) Non-phase locked electroencephalogram (EEG) responses to CO2 laser skin stimulations may reflect central interactions between A partial partial differential- and C-fibre afferent volleys. Clin. Neurophysiol., 114: 710–722. Mouraux, A., Guérit, J.M. and Plaghki, L. (2004) Refractoriness cannot explain why C-fiber laser-evoked brain potentials are recorded only if concomitant Aδ-fiber activation is avoided. Pain, 112: 16–26. Nahra, H. and Plaghki, L. (2003) The effects of A-fiber pressure block on perception and neurophysiological correlates of brief non-painful and painful CO2 laser stimuli in humans. Eur. J. Pain, 7: 189–199. Opsommer, E., Weiss, T., Miltner, W.H. and Plaghki, L. (2001a) Scalp topography of ultralate (C-fibres) evoked potentials following thulium YAG laser stimuli to tiny skin surface areas in humans. Clin. Neurophysiol., 112: 1868–1874. Opsommer, E., Weiss, T., Plaghki, L. and Miltner, W.H. (2001b) Dipole analysis of ultralate (C-fibres) evoked potentials after laser stimulation of tiny cutaneous surface areas in humans. Neurosci. Lett., 298: 41–44. Opsommer, E., Guérit, J.M. and Plaghki, L. (2003) Exogenous and endogenous components of ultralate (C-fibre) evoked potentials following CO2 laser stimuli to tiny skin surface areas in healthy subjects. Neurophysiol. Clin., 33: 78–85.

Plaghki, L., Delisle, D. and Godfraind, J.M. (1994) Heterotopic nociceptive conditioning stimuli and mental task modulate differently the perception and physiological correlates of short CO2 laser stimuli. Pain, 57: 181–192. Ploner, M., Gross, J., Timmermann, L. and Schnitzler, A. (2002) Cortical representation of first and second pain sensation in humans. Proc. Natl. Acad. Sci. USA, 99: 12444–12448. Price, D.D. (1988) Psychological and Neural Mechanisms of Pain. Raven Press, New York. Price, D.D. (1996) Selective activation of Aδ and C nociceptive afferents by different parameters of nociceptive heat stimulation: a tool for analysis of central mechanisms of pain. Pain, 68: 1–3. Price, D.D., Hu, J.W., Dubner, R. and Gracely, R.H. (1977) Peripheral suppression of first pain and central summation of second pain evoked by noxious heat pulses. Pain, 3: 57–68. Qiu, Y., Inui, K., Wang, X., Tran, T.D. and Kakigi, R. (2002) Effects of attention, distraction and sleep on CO2 laser evoked potentials related to C-fibers in humans. Clin. Neurophysiol., 113: 1579–1585. Raij, T.T., Vartiainen, N.V., Jousmaki, V. and Hari, R. (2003) Effects of interstimulus interval on cortical responses to painful laser stimulation. J. Clin. Neurophysiol., 20: 73–79. Sinclair, D.C. and Stokes, B.A. (1964) The production and characteristics of “second pain.” Brain, 87: 609–618. Treede, R.D., Kief, S., Holzer, T. and Bromm, B. (1988) Late somatosensory evoked cerebral potentials in response to cutaneous heat stimuli. Electroencephalogr. Clin. Neurophysiol., 70: 429–441. Truini, A., Rossi, P., Galeotti, F., Romaniello, A., Virtuoso, M., De Lena, C., Leandri, M. and Cruccu, G. (2004) Excitability of the Aδ nociceptive pathways as assessed by the recovery cycle of laser evoked potentials in humans. Exp. Brain Res., 155: 120–123. Valbo, A.B., Hagbarth, K.E., Törebjörk, H.E. and Wallin, B.G. (1979) Somatosensory, proprioceptive and sympathetic activity in human peripheral nerves. Physiol. Rev., 59: 919–957. Valeriani, M., Rambaud, L. and Mauguière, F. (1996) Scalp topography and dipolar source modelling of potentials evoked by CO2 laser stimulation of the hand. Electroencephalogr. Clin. Neurophysiol., 100: 343–353. Valeriani, M., Restuccia, D., Le Pera, D., De Armas, L., Maiese, T. and Tonali, P. (2002) Attention-related modifications of ultra-late CO2 laser evoked potentials to human trigeminal nerve stimulation. Neurosci. Lett., 329: 329–333. Willis, W.D. (1985) The Pain System. Karger, New York.


205

Chapter 28

Interaction of somatosensory evoked potentials within the same nerve and between the two different nerves: including high-frequency oscillation (HFO) Thoru Yamada*, Yoshikazu Azuma and Toshiyuki Yanagisawa Division of Clinical Electrophysiology, Department of Neurology, University of Iowa, College of Medicine, Iowa City, IA 52242 (USA)

1. Background The sensory pathways consist of complex and intricate network connections within the same nerve and among different nerves. One sensory modality may affect another modality or activation of one nerve may alter another nerve input if they share the synaptic connections. The sensory system also interacts with the motor system. The two stimuli delivered in close sequence within the same nerve or between the two different nerves undergo refractory process, occlusion, or inhibition through synaptic connections. Studying refractory periods, occlusion, and inhibition using somatosensory evoked potentials (SEPs) could explore the complex mechanisms of these synaptic connections. Our series of SEP experiments started from the anecdotal observation that the amplitude of common peroneal nerve (CPN)-SEP was smaller than that of the posterior tibial nerve (PTN)-SEP or sural (SN)-SEP. This was contrary to the “intuitive” expectation that

*Correspondence to: Thoru Yamada, M.D., Department of Neurology, Division of Clinical Electrophysiology, 200 Hawkins Drive (0150 RCP), Iowa City, IA 52242, USA. Fax: +1-319-356-1209; E-mail: [email protected]

CPN-SEP with stimulation at the knee would produce larger amplitude response than PTN stimulation at the ankle, because CPN stimulation at the popliteal fossa activates larger fibers receiving greater peripheral receptors than the PTN stimulation at the ankle. 2. Experimental studies 2.1. The effect of stimulus rates upon common peroneal, posterior tibial, and sural nerve somatosensory evoked potential The first study examined the effect of stimulus rates upon CPN-, PTN-, and SN-SEPs (Onishi et al., 1991). The amplitude of P40–N50 and N50–P60 in the PTNSEP and corresponding amplitude of CPN-SEP and SN-SEP at the rate of 2.3, 3.4, 4.1, and 5.1 Hz were measured. When the stimulation rate was increased from 2.3 to 5.1 Hz, the P40–N50 amplitude decreased by 50% for the CPN-SEP and 20% for the PTN-SEP. Also, the N50–P60 amplitude was reduced by 30% in the CPN-SEP and 20% in the PTN-SEP. In contrast, increased stimulus rate produced no significant amplitude decline in the SN-SEP. Blocking the peroneal nerve with lidocaine at just distal to the stimulating electrodes eliminated the descending peroneal nerve

206 volley and abolished the leg movement. This resulted in the disappearance of amplitude attenuation observed with the faster stimulus rate in CPN-SEP. The findings suggest that at higher rates of stimulation, the afferent volleys induced by the movements that follow mixed nerve stimulation interfere with the SEP produced by electrical activation of the sensory afferents. The interference was greater when the more proximal site of the mixed nerve was stimulated. The above changes were similar to the “gating” effect of movement-induced interference (Seyal et al., 1987). 2.2. Recovery functions of CPN-, PTN-, and SN-SEPs If the effects of stimulus rate differ among different nerves, we expect the recovery curves of SEP would be different. The recovery functions of the CPN-, PTN-, and SN-SEPs were studied using a paired conditioning test paradigm (Saito et al., 1992). The interstimulus interval (ISI) of paired stimuli ranged from 2 to 400 ms. In all SEPs, the amplitude recovered once at ISI of 12–20 ms after 2–10 ms refractory period. Further increases of the ISI resulted in another phase of depression of SEP (2nd phase suppression) between 30 and 50 ms ISI, most markedly in CPNSEP (Fig. 1A), less prominently in PTN-SEP, and least in SN-SEP. After a longer than 50 ms ISI, there was progressive recovery of SEP, but full recovery differed depending on the nerve stimulated, 400 ms ISI for CPN-, 250 ms for PTN-, and 100 ms for SN-SEP. The peroneal nerve block by local anesthetic injected just distal to the stimulus electrodes abolished the 2nd phase but not 1st phase SEP suppression observed before the nerve block (Fig. 1B). These findings suggest that the initial phase suppression is due to “refractory period” by stimulating the same nerve in close sequence, but the 2nd phase SEP suppression is attributable to the interference from “secondary” afferents as a result of activation of peripheral receptors (muscle, joint and/or cutaneous) by the efferent volley initiated from the stimulus point. The greater and the longer duration of peripheral receptor activation in CPN than in PTN or SN stimulation could explain the more pronounced and the longer duration of 2nd phase suppression in CPN-SEP.

2.3. The effect of stimulus rates on median, ulnar, and radial nerve SEP From the finding that stimulus rates affect differently among different nerves of lower extremity, we expect the same in upper extremity SEP. We then examined the effect of stimulus rates on the SEP amplitude following stimulation of the median nerve (MN) and the ulnar nerve (UN) at the elbow or wrist, and the radial nerve (RN) at the wrist (Fujii et al., 1994). Because SEP of upper extremity has more specific topographic characteristics than the lower extremity SEP, more detailed analysis was possible in the study of upper extremity SEP. We measured the amplitude of frontal (P14–N18–P22–N30) and parietal peaks (P14–N20–P26–N34). The amplitude attenuation was found at frontal P22 and N30 and to a lesser degree at parietal N20 and P26 peaks with an increasing stimulus rate from 1.1 to 5.7 Hz. The amplitude attenuation was the highest at the elbow when compared to the wrist stimulation in both MN and UN. The attenuation was the least for wrist stimulation for the RN. The UN block by local anesthesia just distal to the stimulus electrode at the elbow abolished the amplitude attenuation caused by the fast stimulus rate. The amplitude reduction with the faster stimulus rate is due to the interference from the “secondary” afferent inputs arising from the activation of peripheral receptors (muscle spindles, joint and/or cutaneous) activators. These frontal dominant changes were similar to the movement-induced gating interference (Conquery et al., 1972; Lee and White, 1974; Rushton et al., 1981; Cheron and Borenstein, 1987). It is also intriguing to note some of the movement disorders such as patients with Parkinson’s disease (Rossini et al., 1989), Huntington’s chorea (Yamada et al., 1991), and dystonia (Reilly et al., 1992) showed frontal N30 abnormality. 2.4. The influence of interfering input from common peroneal nerve on the tibial nerve SEP If motor afferent affect sensory input within the same nerve, one would expect that the interference would occur between the two nerves if they share the synaptic connections. Using a conditioning test paradigm,

207 Recovery of Common Peroneal N. SEP C Z′

L1

N35

Control

(%)

N50

160 P40

ISI (ms)

N35–P40

P60

140

P40–N50

2

N50–P60

4 8 12 16 20 30

120

100

80

40 50 75 100 150

60

40

20

250

mean of 6 subjects

400

0 4 5 μV

8 12 16 20 30 40 50 75 100 150 250 400 (ISI) ms

A

25 ms

Changes of CPN–SEP before and after CPN block at fibular head Before

After

CZ′

L1

CZ ′

L1

N35 N50

Control P40 P60 Primary afferent

ISI (ms) Stim.

4

Stim. Peroneal nerve block

Movement (+)

50 5 μV

10

Primary afferent

70 10 30 10

7010

Secondary afferent (+)

Movement (−) Secondary afferent (−)

30 ms

B

Fig. 1. (A) The recovery of CPN-SEP by conditioning and test paired stimuli showed dual phase suppression with the initial suppression at ISI of 2–12 ms followed by recovery at 6–20 ms ISI and 2nd phase suppression at 30–50 ms ISI, with the final recovery at 400 ms ISI. Note the depressed spinal potential (L1 spine) at 2 ms ISI (secondary to refractory period) but no depression at 30–50 ms ISI when cortical SEP was markedly suppressed. (B) Local nerve block by lidocaine just distal to the stimulus site eliminated the leg movement associated with the stimulus, which in turn blocked the secondary afferent. This resulted in the recovery of SEP at 50 ms ISI but not at 4 ms ISI. (Modified with permission from Saito et al., 1992 and Yamada et al., 1992.)

208 we then studied the recovery function of PTN-SEP conditioned by preceding CPN stimulation at the popliteal fossa (Yamada et al., 1992). The ISIs ranged from 0 to 400 ms, where 0 ms indicated simultaneous arrival of tibial and peroneal nerve volleys at the L1 spine. The recovery curve was W-shaped, showing two peaks of SEP suppression, maximum at 6 ms ISI (1st phase) and 50–75 ms ISI (2nd phase), and recovery between two phases of suppression (Fig. 2A and 2B, right column). At 0–2 ms ISI, P40–N50–P60 amplitude decreased. At 4–6 ms ISI, all peaks were markedly depressed. Unlike the initial phase suppression due to refractory period in case of single nerve stimulation (Fig. 2A and 2B, left column), we attribute the 1st phase suppression to two processes; initial “occlusive effect” at 0 ms at ISI and subsequent “inhibitory effect,” each mediated via a central synaptic network between the two nerves (Fig. 2B). This differed from the 1st phase suppression in case of PTN-paired stimuli in which refractory period dominates the suppression (Fig. 2B). The abolishment of the 2nd but not the 1st phase suppression by CPN block distal to the stimulation electrodes provided the evidence that the 2nd phase suppression resulted from interfering afferent signals generated by the activation of CPN peripheral receptors induced by foot movements. 2.5. The interaction of SEPs between mixed–sensory and sensory–sensory nerves In this study, the interactions of mixed nerve (PTN) and sensory nerve (SN), and also sensory (SN) and sensory (saphenous) nerves were examined (Hasegawa et al., 2001). We found that the mixed nerve (PTN) exerted similar dual phases of suppression (as was seen in PT – peroneal nerve study) on to the SN SEP, but the reverse was not true. Also the sensory and sensory nerve interactions were not mutually equal; the SN stimulation caused suppression of two phases but the reverse condition did not show significant suppression. The above findings suggest (1) interference input from the sensory nerve to the mixed nerve is much weaker than the reverse condition, and (2) sensory and sensory nerves interactions occur but two nerves’ interference inputs are not necessarily equal and one could dominant the other. The results also support that

the pure sensory input plays a role for “gating” of cortical SEP (Kakigi and Jones, 1986; Jones et al., 1989) but it is weaker than motor input. 2.6. The interference of ulnar nerve input on median nerve SEP We then examined the interaction of two nerves in upper extremity SEP. Using UN conditioning stimulus, MN SEP change was measured (Ando, 1991). This also showed dual phase suppression of recovery curves. Most prominent change was noted in frontal (P20–N30) rather than parietal peaks (N20–P26). The 1st phase suppression was maximal at 4 ms, which was followed by partial or full recovery at 10–20 ms ISI. The 2nd phase suppression was maximal at 30–50 ms ISI and final recovery took place at 200–250 ms ISI. Similar to the changes of fast stimulus in MN SEP (Fujii et al., 1994), the most prominent change was decreased frontal P20–N30 components. As compared to the PTN-SEP conditioned by CPN stimulation, degree of suppression was less and the recovery was faster. As was shown in CPN–PTN paradigm, anesthetic local nerve block at the cubital tunnel distal to the stimulus electrode, abolished 1st phase suppression but not 2nd phase suppression. In contrast, the local nerve block at Guyon’s canal proximal to the stimulus electrodes abolished 2nd phase suppression but not 1st phase suppression. These evidences confirm that the primary sensory afferent ascending from the stimulus site exerts the 1st phase suppression and the 2nd phase suppression is brought out by the secondary afferent after activation of peripheral receptors evoked by descending impulse from the stimulus site. 2.7. The change of N20 and high-frequency oscillation (HFO) of median nerve SEP affected by ulnar nerve input Independent functions of first cortical potential of N20 and HFOs overriding N20 “slow” potential have been shown by the disappearance of HFO and persistent N20 in sleep (Yamada et al., 1988; Hashimoto et al., 1996). In our most recent experiment, recovery of HFO was compared with that of N20 in MN SEP conditioned

209 PTN(conditioning)–PTN(test) paired St. Recovery function N35

CPN(conditioning)–PTN(test) paired St. Interference function

N50

Control P40

ISI (ms)

ISI (ms)

P60

N50

N35

Control

P40

P60

0

2

2

4

4

8

6

12

12

16

16 20

20

30 30

40

50

50

75

75

100

100 150

150

2 μV

250

250 2 μV

1 μV

70 ms 10

10

70 ms

1 μV

400

A

70 ms 10

10

70 ms

Recovery Curves of Tibial N.SEP Tibia–Tibia N. Paired Stim.

Peroneal–Tibia N. Paired Stim. (%)

(%) 100

100

Occlusion

Refractory

Inhibition

80

80

60

60

40

40

N35–P40 20

20

P40–N50 N50–P60

Mean of 6 subjects

Mean of 6 subjects 0

0 4

8

12 16 20 30 40 50

ISI (ms)

75 100 150

250

400

0

B

2

4

6

12 16 20 30 40 50 75 100 150 250 400

ISI (ms)

Fig. 2. (A) The PTN-SEP conditioned by PTN showed the initial suppression maximally at 2–4 ms ISI followed by recovery at 12–20 ms ISI (left column). This contrasted to the PTN-SEP suppression when conditioned by CPN, which showed relative suppression at 0 ms ISI and maximum suppression at 6 ms ISI, followed by recovery at 20 ms ISI (right column). Both showed 2nd phase suppression after 20 ms ISI, which lasted longer with CPN than with PTN-conditioning stimulation. (B) Note the difference of the recovery curves (6 subjects) during the initial phase suppression between the tibial nerve–tibial nerve and peroneal nerve–tibial nerve paired stimuli paradigms. The maximum suppression was at the shortest ISI with progressive recovery with the increase of ISI until ISI of 16 ms, which was due to refractory period in the former paradigm (left). Whereas in the latter paradigm, the suppression progressively increased from 0 to 6 ms ISI which was due to initial occlusion followed by inhibitory effect via connection synapse between two nerves (right column). (Modified with permission from Saito et al., 1992 and Yamada et al., 1992.)

210 First phase suppresion of N20 and HFO Median N. Test and Ulnar N. Conditioning stim. HFO

N20

Control ISI (ms) 0 2 4 5 33

3 ms

33

3 ms

Recovery Curves of HFO vs N20 & P24 Ratio (vs Control) 1.2

1.0

0.8

0.6

0.4

#

P

Functional Neuroscience: Evoked Potentials and Related Techniques: Presentations from the 8th International Evoked Potentials Symposium (Fukuoka, ... 59 (Supplements to Clinical Neurophysiology)

Evoked Spinal Cord Potentials: An illustrated Guide to Physiology, Pharmocology, and Recording Techniques

Evoked Spinal Cord Potentials: An illustrated Guide to Physiology, Pharmocology, and Recording Techniques

Quantitative EEG, Event-Related Potentials and Neurotherapy

Electrode potentials

ECOLOGY - POTENTIALS AND LIMITATIONS

Molecular Electrostatic Potentials

Transcranial Magnetic Stimulation: Supplement to Clinical Neurophysiology Series, Volume 56 (Supplements to Clinical Neurophysiology)

Electromagnetic Fields Derived from Non-Commuting Potentials

The Critical and Dissociation Potentials of Hydrogen

The Clinical Neurophysiology Primer

The Clinical Neurophysiology Primer

Next Generation Networks: Perspectives and Potentials

Next Generation Networks: Perspectives and Potentials

The Critical Potentials of Molecular Hydrogen

Clinical Neurophysiology (Contemporary Neurology)

Color Atlas of Neuroscience - Neuroanatomy and Neurophysiology

PT-Symmetric Schrödinger Operators with Unbounded Potentials

Clinical Neurophysiology, Third Edition

Globalization and the Least Developed Countries: Potentials and Pitfalls

Neurology and Clinical Neuroscience

A Synopsis of the Ionization Potentials of the Elements

Series and Ionization Potentials in the Iron Spectrum

Potentials of Disorder (New Approaches to Conflict Analysis)

Electrophysiology of Mind: Event-Related Brain Potentials and Cognition (Oxford Psychology)

Dietary Supplements and Functional Foods

Dietary Supplements and Functional Foods

Multiculturalism and Social Cohesion: Potentials and Challenges of Diversity

Functional Neuroscience: Evoked Potentials and Related Techniques: Presentations from the 8th International Evoked Potentials Symposium (Fukuoka, ... 59 (Supplements to Clinical Neurophysiology)

Evoked Spinal Cord Potentials: An illustrated Guide to Physiology, Pharmocology, and Recording Techniques

Evoked Spinal Cord Potentials: An illustrated Guide to Physiology, Pharmocology, and Recording Techniques

Quantitative EEG, Event-Related Potentials and Neurotherapy

Electrode potentials

ECOLOGY - POTENTIALS AND LIMITATIONS

Molecular Electrostatic Potentials

Transcranial Magnetic Stimulation: Supplement to Clinical Neurophysiology Series, Volume 56 (Supplements to Clinical Neurophysiology)

Electromagnetic Fields Derived from Non-Commuting Potentials

The Critical and Dissociation Potentials of Hydrogen

The Clinical Neurophysiology Primer

The Clinical Neurophysiology Primer

Next Generation Networks: Perspectives and Potentials

Next Generation Networks: Perspectives and Potentials

The Critical Potentials of Molecular Hydrogen

Clinical Neurophysiology (Contemporary Neurology)

Color Atlas of Neuroscience - Neuroanatomy and Neurophysiology

PT-Symmetric Schrödinger Operators with Unbounded Potentials

Clinical Neurophysiology, Third Edition

Globalization and the Least Developed Countries: Potentials and Pitfalls

Neurology and Clinical Neuroscience

A Synopsis of the Ionization Potentials of the Elements

Series and Ionization Potentials in the Iron Spectrum

Potentials of Disorder (New Approaches to Conflict Analysis)

Electrophysiology of Mind: Event-Related Brain Potentials and Cognition (Oxford Psychology)

Dietary Supplements and Functional Foods

Dietary Supplements and Functional Foods

Multiculturalism and Social Cohesion: Potentials and Challenges of Diversity

Recommend Documents