Biomagnetism: Proceedings. Third International Workshop, Berlin(West), May 1980 9783110863529, 9783110084030

Table of contents :
Preface
List Of Participants
Contents
Instrumentation
Instrumentation For Biomedical Applications
Construction And Performance Of The Otaniemi Magnetically Shielded Room
The Berlin Magnetically Shielded Room (BMSR), Section A: Design And Construction
The Berlin Magnetically Shielded Room (BMSR) Section B - Performances
The Berlin Magnetically Shielded Room (BMSR) Section C - Periphery
Installation Of A Biomagnetic Measurement Facility In A Hospital Environment
Thick-Walled Conducting Shield In Biomagnetic Experiments
An Aluminium Shielded Room For Biomagnetic Measurements
A Superconducting Helmet For Magnetoencephalography With A Squid
An Integrated System For Magnetic Assessment Of Cardiac Function
Biomagnetic Measurements In Unshielded, Normally Noisy Environments
Fundamentals
Medical Significance Of The Magnetic Activities Of The Human Body
Generation Of Magnetic Fields By The Human Body (Theory)
Biomagnetic Fields And Cellular Current Flow
Biomagnetic Fields. Cardiomagnetism
Cardiomagnetism
Theory Of The Pr Sequent Of The Human Magnetocardiogram
High Resolution Cardiomagnetism
Magnetocardiographic Study Of Some Human Cardiac Electrophysiological Phenomena: Preliminary Observations (+)
R And T Waves In The MCG And ECG
On The Recording Of Phase And Amplitude Relationships Between Electrical And Magnetical Cardiac Events
A Method Of Magnetocardiography For Clinical Use
Simultaneous Measurement Of The Magnetic Heart Vector Components With The Uni Positional Lead System
Enhancement Of Magnetocardiograms By Applied Magnetic Fields
The Influence Of The Torso On The Magnetic Field Of A Current Dipole
Properties Of An Ideal MCG-Recording System
Biomagnetic Fields: Neuromagnetism
Magnetic Fields Of The Cerebral Cortex
Transient Evoked Responses: Some Comments On The Use Of Magnetic Recording
Localization Of Neural Generators Underlying Auditory Evoked Magnetic Fields Of The Human Brain
Estimation Of The Magnetoencephalogram Power Spectrum
Evoked Magnetic Fields Reveal Different Visual Areas In Human Cortex
Application Of A Squid To Measurement Of Somatically Evoked Fields: Transient Responses To Electrical Stimulation Of The Median Nerve
Sensitivity Distribution In Magnetoencephalography
Magnetoretinogram And Magneto-Oculogram In Man
Susceptometry: Magnetopneumography
Practical Magnetopneumography Using Fluxgate Magnetometers
Magnetic Field Measurements On Lungs: A Comparison Between Two Methods
Evaluation Of Magnetopneumography For Assessing Thoracic Accumulation Of Welding Fume Particulate And Lung Dust Clearance
Localized-Field Magnetopneumographic Measurements Of Coal Workers And Of Freeze-Dried Lungs
Susceptometry: Thalassemia Susceptometry
Non-Invasive SQUID Diagnosis Of Liver Iron Overload
Susceptometer For In Vivo Measurements Of Iron Stored In Human Tissue
Miscellaneous
Proposal For An Improvement Of NMR-Imaging By Low-Temperature-squid- Detection Towards Molecular Kinetic Measurements Localized To A Small Sub-Region Of The Sample
Subject Index
Author Index

Citation preview

Biomagnetism

Biomagnetism Proceedings Third International Workshop on Biomagnetism Berlin (West), May 1980 Editors S. N. Erne, H-D. Hahlbohm, H. Lubbig

W DE G Walter de Gruyter • Berlin • New York 1981

Editors: Sergio Nicola Erne, Dr. Fis. Hans-Dieter Hahlbohm, Dr. phil. Heinz Lübbig, Dr.-Ing. Physikalisch-Technische Bundesanstalt Institut Berlin Abbestraße 2-12 D-1000 Berlin 10

CIP-Kurztitelaufnahme

der Deutschen

Bibliothek

Biomagnetism: proceedings/3. Internat. Workshop on Biomagnetism, Berlin (West), May 1980. Ed. S. N. Erné... - Berlin; New York: de Gruyter, 1981. ISBN 3-11-008403-1 NE: Erné, Sergio N. [Hrsg.); International Workshop on Biomagnetism

Library of Congress Cataloging in Publication Data

International Workshop on Biomagnetism (3rd, 1980, Berlin, Germany) Biomagnetism. Proceedings of earlier workshops not published. Bibliography: p. Includes indexes. 1. Body, Human-Magnetic fields-congress. 2. Biomagnetism-Congress. 3. Superconducting quantum interference devices-congress. I. Erne, S. N. (Sergio Nicola), 1944 II. Hahlbohm, H. D„ 1930 - III. Liibbig, H„ 1932 - IV. Title. QP345.I57 1980 61Z.0142 81-5457 ISBN 3-11 -008403-1 AACR2

Copyright ©1981 by Walter de Gruyter & Co., Berlin 30. All rights reserved, including those of translation into foreign languages. No part of this book may be reproduced in any form - by photoprint, microfilm or any other means nor transmitted nor translated into a machine language without written permission from the publisher. Printing: Karl Gerike, Berlin. - Binding: Dieter Mikolei, Berlin. Printed in Germany.

PREFACE Low temperature magnetic measurement techniques have provided improved f a c i l i t i e s f o r use in Biomagnetism. This has resulted in i n t e n s i f i e d a c t i v i t i e s to i n v e s t i g a t e magnetic f i e l d s generated in the human body and to make possible c l i n i c a l a p p l i c a t i o n s of biomagnetic e f f e c t s . Results have been discussed at Workshops on Biomagnetism in Boston (1976) rd and Grenoble (1978). The 3 Workshop on Biomagnetism was organized in p a r a l l e l with the IC SQUID '80 in B e r l i n (West). The delegates, representing 12 c o u n t r i e s , included 56 p a r t i c i p a n t s . rd This volume presents the Proceedings of the 3

Workshop. Special

lectures

and some papers on SQUID technology were published in the Proceedings of IC SQUID '80 " .

Since i t i s the f i r s t time that the Proceedings of a

Biomagnetic Workshop are published, there i s no precedent. We have arranged the contents using both physical and medical p r i n c i p a l s

reflecting

the i n t e r d i s c i p l i n a r y character of the workshop. The Subject Index should be taken as a f i r s t attempt to set up a guide of t h i s not yet f u l l y systematized subject. We are conscious that under these conditions some weak points are unavoidable. Our wishes are that t h i s volume may stimulate further and, in p a r t i c u l a r , more refined volumes on Biomagnetism, an area which appears to be of i n creasing importance. Our thanks go to the authors who have enabled us to produce t h i s volume. The publication of the proceedings affords a s u i t a b l e opportunity to acknowledge once more the generous support of the President of the Deutscher Bundestag, the B e r l i n Senate and the President of the P h y s i k a l i s c h - T e c h nische Bundesanstalt at the Conference. I t should a l s o be emphasized that without the cheerful and active p a r t i c i pation of the members of the I n s t i t u t B e r l i n of the Physikalisch-Technische Bundesanstalt in a l l aspects of the conference a c t i v i t i e s , the meeting would not have been p o s s i b l e . On behalf of a l l of them our thanks go to G. Sauerbrey, the head of the I n s t i t u t .

VI We are greatly indebted to Frau M. Bieber and Frau A. Lochmann for their loyalty and patient secretarial organization of this volume. We are also indebted to the de Gruyter Publishing Company for its valuable cooperation in preparing these proceedings.

Berlin, February 1981

S. N. Erne H.-D. Hahlbohm H. LUbbig

This volume was generously supported by VACCUMSCHMELZE GMBH HANAU

SQUID '80

Superconducting Quantum Interference Devices and their

Appi ications Editors: H. D. Hahlbohm, H. Lubbig Walter de Gruyter, Berlin • New York 1980

LIST OF PARTICIPANTS

Aittoniemi, K.

Department of Technical Physics, Helsinki University of Technology, 02150 Espoo 15, Finland

Barbanera, S.

Laboratorio di Elettronica dello Stato Solido, C.N.R., Via Cineto Romano 42, Roma, Italy

Barnard, B.

Cryogenic Consultants Limited., Metrostore Building, 231 The Vale, Acton, London, W3 7QS, England

Bastuscheck, C.M.

Neuromagnetism Laboratory, Physics Department, New York University, 4 Washington Place, New York, NY 10003 USA

Bergmann, W.

Kraftwerk-Union AG, R122/Bau 55, D-8520 Erlangen, Germany

Blum, Th.

Technische Universität Berlin, I n s t i t u t für Informatik, Einsteinufer 37-39. 1000 Berlin 10, Germany

Boesiger, P.

I n s t i t u t für biomedizinische Technik, Universität Zürich, Moussonstraße 18, 8044 Zürich, Switzerland

Borek, L.

Vacuumschmelze GmbH, Postfach 2253, D-6450 Hanau 1, Germany

Brauer, F.

Klinikum Westend, HNO-Forschungstrakt, Spandauer Damm 130, 1000 Berlin 19, Germany

Brenner, D.

Departments of Psychology and Physics, Neu York,University, 6 Washington Place, New York, NY 10003, USA

Vili Groupe P h y s i o l o g i e , French-German

Buck, K.

Research I n s t i t u t e , 12 rue de l ' I n d u s t r i e , F-68301 S a i n t - L o u i s , France Combasson

DRET/SDR, 26 Boulevard

Victor,

F-75996 Paris Aimces, France Costa R i b e i r o , P.

Departanento de F i s i c a ,

Pontificia

Universidade C a t ó l i c a , C.P. 38071 Rio de Janeiro, B r a s i l Crum, D.B.

S.H.E. Corporation, 4174 Sorrento Valley B l v d . , San Diego, CA 92121, USA

Durcansky, G.

I n s t i t u t für Festkörperforschung, Kerforschungsanlage J ü l i c h , Postfach 365, D-5170 J ü l i c h , Germany

Duret, D.

C.E.N./G - LETI/NCE, 85X, F-38041 Grenoble, France

Eghrari,

I.R.

Departamento de F i s i c a ,

Pontificia

Universidade C a t ó l i c a , C.P. 38071 Rio de Janeiro, B r a s i l Erné, S.N.

Physikalisch-Technische Institut Berlin,

Bundesanstalt,

Abbestraße 2 - 12,

1000 B e r l i n 10, Germany F a r r e l l , D.E.

Physics Department, I n s t i t u t e of Technol o g y , Case Western Reserve U n i v e r s i t y , Cleveland, Ohio 44106, USA

F e n i c i , R.

S e r v i z i o C a r d i o l o g i a , U n i v e r s i t a 1 Cattol i c a Del Sacro Cuore, P o l i c l i n i c o "A. Gemelli", Largo Gemelli 8, Roma, I t a l y

Freedman, A.P.

Di v i s i o n of Pulmonary Diseases, Hahnemann Medical College and H o s p i t a l , 230 North Broad S t r e e t , P h i l a d e l p h i a , Pa. 19107, USA

IX

Ganssen, A.

Siemens AG, UB Med., Abt. RKG 2, Henkestraße 127, D-8520 Erlangen, Germany

Gross, F.

Universität Graz, Institut für Experimentalphysik, Universitätsplatz 5, 8010 Graz Austria

Hahlbohm, H.-D.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany Hari, R.

Laboratory of Clinical

Neurophysiology,

University Hospital of Helsinki, Haartmanink 4, 00290 Helsinki 29, Finland Hillenbrand

Siemens AG, Forschungslabor FL MET 1, Postfach 32 40, D-8520 Erlangen 2, Germany

Hoenig, H.E.

Physikalisches Institut der Universität Frankfurt, Robert Mayer Str. 2, D-6000 Frankfurt 1, Germany

Jablonski, H.

S.H.E. GmbH, Maria-Theresia-Allee 22, D-5100 Aachen, Germany

Jakschik, J.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany Joseph, S.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany Katila, T.

Department of Technical

Physics,

Helsinki University of Technology, 02150 Espoo 15, Finland Karp, P.F.

University of Kuopio, P.O. Box 138, SF-70101 Kuopio 10, Finland

Technical Research Centre of F i n l a n d , Metall imiehenkuja 8, SF-02150 Espoo 15, Finland U n i v e r s i t ä t Frankfurt, Robert-MayerStraße 2, D 6000 Frankfurt/M, Germany Vacuumschmelze GmbH, Postfach 2253, D-6450 Hanau 1, Germany Department of P h y s i c s , Stanford U n i v e r s i t y , Stanford, CA 94305, USA Biomedical Engineering Laboratory, Department of E l e c t r i c a l

Engineering,

Tampere U n i v e r s i t y of Technology, P.O. Box 527, SF-33101 Tampere 10, Finland Dornier System, Abt. NTF, Postfach 1360 D-7990 Friedrichshafen, Germany Physikalisch-Technisehe

Bundesantalt,

I n s t i t u t B e r l i n , Abbestraße 2-12, 1000 B e r l i n 10, Germany Vaccumschmelze GmbH, Postfach 2253, D-6450 Hanau 1, Germany Biomedical Engineering Laboratory, Department of E l e c t r i c a l

Engineering,

Tampere U n i v e r s i t y of Technology, P.O. Box 527, SF-33101 Tampere 10, Finland I s t i t u t o di F i s i c a "G. Marconi", Università d e g l i Studi di Roma, Piazzale delle Scienze, Roma, I t a l y Klinikum Westend, HNO-Forschungstrakt, Spandauer Damm 130, 1000 B e r l i n 19, Germany

XI Naito, S.

Yokogawa,Electric Works, 9-32 Naka-Cho 2 Chome, Musashino-Shi, Tokyo 180, Japan

Neubert, D.

Physikalisch-Technische Bundesanstalt, Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany

Neumai er, K.

Zentralinstitut für Tieftemperaturforschung, Bayerische Akademie der Wi ssenschaften, Hochschul gelände, D-8045 Garching, Germany

Nink, R.

Physikalisch-Technische Bundesanstalt, Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany

Ohmichi, H.

Clinical Research Institute, National Medical Center Hospital, Toyama-cho 1, Shinjuku-ku, Tokyo 162, Japan

Okada, Y.

Neuromagnetism Laboratory, New York University, 6 Washington Place, New York, N.Y. 10003, USA

Palow, J.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12 1000 Berlin 10, Germany Peters, M.J.

Department of Technical Physics, Technical University Twente, P.O. Box 217, 7500 AE Enschede, The Netherlands

Plonsey, R.

Department of Biomedical

Engineering,

Case Western Reserve University, Cleveland, Ohio 44106, USA Robinson, S.E.

Division of Pulmonary Diseases, The Hahnemann Medical College and Hospital, 230 North Broad Street, Philadelphia, PA 19102, USA

XII Romani, G.L.

Laboratorio di Elettronica dello Stato Solido, C.N.R., Via Cineto Romano 42, 00156 Roma, Italy

Rushby, H.

School of Mathematical and Physical Sciences, Physics Division, University of Sussex, Brighton BN1 9QH, Great Britain

Swithenby, S.J.

Physics Discipline, The Open University, Walton Hall, Milton Keynes, MK7 6AA, Great Britain

Scheer, H.-J.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany Schlief, R.G.

Physikalisch-Technische

Bundesanstalt,

Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany Stroink, G.

Department of Physics, Dalhousie University, Halifax, Nova Scotia, B3H 3J5, Canada

Stuart, C.I.J.M.

Department of Physics, P124, University of Alberta, Edmonton, Alberta T6G 2J1, Canada

Teppner, U.

Physi kali sch-Techni sehe Bundesanstalt, Institut Berlin, Abbestraße 2-12, 1000 Berlin 10, Germany

Teszner, D.

Hopital Ambroise Pare, 9 Ave. Charles de Gaulle, 92100 Boulogne, France

Tripp, J.H.

Department of Physics, Case Western Reserve University, Cleveland, Ohio 44106, USA

XIII

T r o n t e l j , Z.

Physics Department, U n i v e r s i t y of Ljubljana, P.O. Box 543, 61001 Ljubljana, Yugoslavia

U l r i c h , B.

Max-Planck-Institute f o r Plasma P h y s i c s , D-8046 Garching/München, Germany

Varpula, T.

Department of Technical

Physics,

H e l s i n k i U n i v e r s i t y of Technology, Otakaari 3 A, 02150 Espoo 15, Finland Vollmer, R.

Neurologische Abteilung, Krankenhaus Lainz, Wolkersbergenstraße 1, A-1130 Wien, A u s t r i a

de Waal, V.J.

Laboratorium voor Technische Natuurkunde Technische Hogeschool

Delft,

Lorentzweg 1, 2628 CJ D e l f t , The Netherlands Wevers-Henke, J . J .

Department of Technical P h y s i c s , Technical U n i v e r s i t y Twente, Postbox 217, Enschede, The Netherlands

Williamson, S . J .

Department of P h y s i c s , University of New York, 4 Washington Place, New York, N.Y 10003, USA

Witt, Th. J.

Bureau International Des Poids et Mesures P a v i l l o n de B r e t e u i l , 92310 Sevres, France

Zimmerman, J.E.

U.S. Department of Commerce, National Bureau of Standards, Cryoelectronic Metrology Group, Boulder .Colorado 80303 USA

CONTENTS

INSTRUMENTATION Instrumentation for biomedical a p p l i c a t i o n s . T. K a t i l a

3

Construction and performance of the Otaniemi magnetically shielded room.

V.O. Kelha

33

The B e r l i n magnetically shielded room (BMSR). Section A: Design and construction. Section B:

Performances.

A. Mager

51

S.N. Erne, H.-D. Hahlbohm, H. Scheer,

Z. Trontel j Section C:

79 Periphery. S.N. Erne, H.-D. Hahlbohm, J. Palow

89

I n s t a l l a t i o n of a biomagnetic measurement f a c i l i t y in a hospital environment.

C. Bercy, D. Duret, P. Karp, D. Teszner

95

Thick-walled conducting s h i e l d in biomagnetic experiments. J. Malmivuo, P. Heinonen, M. Tuomola, J. Lekkala An aluminium shielded room for biomagnetic measurements.

107 G. S t r o i n k ,

B. Brown, B. Blackford, M. Horacek

113

A superconducting helmet for magnetoeneephalography with a SQUID. H.E. Hoenig, C. Gassinger

117

An integrated system f o r magnetic assessment of cardiac function. M.C. L e i f e r , J.C. G r i f f i n , E.J. I u f e r , J.P. Wikswo, W.M. Fairbank, D.C. Harrison

123

Biomagnetic measurements in unshielded, normally noisy environments. S. Barbanera, P. C a r e l l i , R. Leoni, G.L. Romani, F. Bordoni, I . Modena, R. F e n i c i , P. Z e p p i l l i

139

FUNDAMENTALS Medical s i g n i f i c a n c e of the magnetic a c t i v i t i e s of the human body. A.P. Freedman

153

XVI Generation of magnetic fields by the human body (theory). R. Plonsey

177

Biomagnetic fields and cellular current flow. J.H. Tripp

207

BIOMAGNETIC FIELDS Cardiomagnetism Cardiomagnetism.

P. Karp

219

Theory of the PR segment of the human magnetocardiogram.

J.H. Tripp,

D.E. Far re 11 High resolution cardiomagnetism.

259 D.E. Farrell, J.H. Tripp,

C.L. VanDoren

273

Magnetocardiographic study of some human cardiac electrophysiological phenomena: Preliminary observations.

S. Barbanera, P. Carelli,

R. Leoni, G.L. Romani, F. Bordoni, I. Modena, R. Fenici, P. Zeppilli R and T waves in the MCG and ECG.

283

C.I.J.M. Stuart, S.B. Woods,

A.C. Crawford, N.R. Thomas, K.B. Newbound

.'

291

On the recording of phase and amplitude relationships between electrical and magnetical cardiac events.

I.R. Eghrari,

J.P. von der Weid, P. Costa Ribeiro, 0. Symko

303

A method of magnetocardiography for clinical use. H. Ohmichi, N. Ibuka, S. Naito, M. Kotani, Y. Uchikawa, K. Atsumi

309

Simultaneous measurement of the magnetic heart vector components with the unipositional lead system.

J. Lekkala, J. Malmivuo

319

Enhancement of magnetocardiograms by applied magnetic fields. M.J. Peters, Z. Dunajski, L.C. van der Marel

327

The influence of the torso on the magnetic field of a current dipole. M.J. Peters, A.P. van de Graaf, A. van Oosterom

337

Properties of an ideal MCG-Recording system.

343

J. Malmivuo

XVII

BI'OMAGNETIC FIELDS Neuromagnetism Magnetic fields of the cerebral cortex. S.J. Williamson, 353

L. Kaufman Transient evoked responses: Some comments on the use of magnetic recording.

D.E. Farrell, J.H. Tripp, J. Patrick, D. Hess

403

Localization of neural generators underlying auditory evoked magnetic fields of the human brain. K.'Aittoniemi, R. Hari, M.-L. Jarvinen, T. Katila, T. Varpula

415

Estimation of the magnetoencephalogram power spectrum. D.W. Hess ...

423

Evoked magnetic fields reveal different visual areas in human cortex. D. Brenner, Y.'Okada, E. Maclin, S.J. Williamson, L. Kaufman

431

Application of a SQUID to measurement of somatically evoked fields: Transient responses to electrical stimulation of the median nerve. Y.C. Okada, L. Kaufman, D. Brenner, S.J.Williamson Sensitivity distribution in magnetoencephalography.

445 J. Malmivuo ...

457

Magnetoretinogram and magneto-oculogram in man. K. Aittoniemi, M.-L. Jarvinen, T. Katila, R. Maniewski, T. Varpula

463

SUSCEPTOMETRY Magnetopneumography Practical magnetopneumography using fluxgate magnetometers. K. Aittoniemi, K. Kalliomaki, T. Katila, T. Varpula

475

Magnetic field measurements on lungs: A comparison between two methods. G. Stroink, D. Dahn

485

Evaluation of magnetopneumography for assessing thoracic accumulation of welding fume particulate and lung dust clearance.

A.P. Freedman,

S.E. Robinson

489

Localized field magnetophneumographic measurements of coal workers and of freeze-dried lungs. F.H.Y. Green

A.P. Freedman, S.E. Robinson, 497

XVIII

SUSCEPTOMETRY Thalassemia susceptometry Non-invasive SQUID diagnosis of liver iron overload.

D.E. Farrell,

J.H. Tripp, P.E. Zanzucchi, J.H. Harris, G.M. Brittenham, W.A. Muir

507

Susceptometer for in vivo measurements of iron stored in human tissue.

C.M. Bastuscheck, D. Brenner, S.J. Williamson, L. Kaufman..

519

MISCELLANEOUS Proposal for an improvement of NMR-imaging by low-temperatureSQUID-detection towards molecular kinetic measurements localized to a small sub-region of the sample.

W.H. Bergmann

535

Subject Index

549

Author Index

557

INSTRUMENTATION

INSTRUMENTATION FOR BIOMEDICAL APPLICATIONS

T. Katila Department of Technical Physics, Helsinki University of Technology, SF-02150 Espoo 15, Finland

Introduction B i o e l e c t r i c measurements have long been of great importance in both biophysics and medical science and also in clinical applications. Measurements of electric potential differences in human skin are well known routine procedures. Such registrations are called the electrocardiogram (ECG), the electroencephalogram (EEG), the electro-oculogram (EOG), etc. Various bioelectric phenomena, their amplitudes and typical measurement bandwidths are listed in Table 1. The Table also includes references to pioneering works in bioelectric measurements. B i o m a g n e t i c fields are, by definition, magnetic fields produced by electric currents associated with bioelectric activity or by magnetization of biological tissue. Table 1 includes magnetic counterparts of bioelectric signals. The first observation of these biomagnetic fields, namely that associated with bioelectric activity of the heart, was made by Baule and McFee in 1963 /l/. The corresponding registration is called the magnetocardiogram (MCG). Most magnetic observations have been made during the past decade. Although

some magnetic properties

known for a long time,

of human tissue

magnetic susceptibility of the tissue for biomagnetic studies

have been

the use of the magnetization or of the and of its contaminants

is relatively new.

Examples

kind of measurements are also mentioned in Table 1.

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

of this

4

BIOELECTRIC AND BIOMAGNETIC FIELDS BIOELECTRIC PHENOMENA

AMPL. (pT)

BANDW. (Hz)

MAGNETOCARDIOGRAM

50

.05-100

FETAL MAGNETOCARD.

1-10

.05-100

AMPL.

BIOMAGNETIC PHENOMENA

(JJV)

ELECTROCARDIOGRAM

1000

FETAL ELECTROCARD.

5-50

(ECG) WALLER 1887

(FECG) CREMER 1906

(MCGJ BAULE ET AL. 1963

(FMCG) KARINIEMI ET AL.1974

E LECTROENCEPHALOGRAM

50

MAGNETOENCEPHALOGRAM

1

.5-30

EVOKED POTENTIALS

10

EVOKED FIELDS

.1

DC-60

MAGNETOMYOGRAM

10

DC-2000 DC

(EEG) BERGER 1924.

(VEP) WALTER ET AL. 1946 (SEPI DAWSON ET AL. 195C (AEP) DAVIS ET AL. 1939

I MEG) COHEN 1968

(VEF) COHEN 1975 (SEP) BRENNER ET AL. 1978 IAEF) REITE E T A L . 1978

ELECTROMYOGRAM

1000

ELECTRO-OCULOGRAM

1000

MAGNET0-0CUL0GRAM

10

ELECTRORETINOGRAM

100

MAGNETORETINOGRAM

.1

(EMG) ADRIAN 1929

(EOGI DUBOIS-REYMOND 1849 (ERG) HOLMGREN 1865

(MMG) COHEN 1972

IMOG) KARP ET AL. 1976

(MRGI AITTONIEMI ET AL. 1978

.1-30

MAGNETIZATION EFFECTS MAGNETIC CONTAMINATION (LUNG) IPC) COHEN 1973

MAGNETIC SUSCEPTIBILITY PLETHYSMOGRAPHY (MSPG) WIKSWO ET AL. 1974

HUMAN IRON STORES HARRIS ET AL. 1978

Table 1. Various bioelectric phenomena and their biomagnetic counterparts. Corresponding bioelectric signals do not exist for pure magnetization phenomena. References to the pioneering works of various biomagnetic measurements are made in the reference list from no 1 to no 12. All known biomagnetic fields are not included in the Table. This paper

is

to

present

niques

of

biomagnetic fields

To-day

it

is not

yet

aspects on for

the measurement tech-

biomedical

quite clear

applications.

to what extent

magnetic

measurements

may give new information,

unattainable by elec-

tric means.

Hence clinical applications

can only be given as

potentional candidates for the future.

5 On the Origin of Biomagnetic Maxwell's

electrodynamic

Fields

field

equations

for

the magnetic

flux density B are: V • B = 0 ,

(1) 3D

V x B = u Q (jfc + V x M +

) •

(2)

Assuming that the time derivative of the electric flux density D is negligible as compared with the total

(true) current den-

sity j t and with the curl of the magnetization M, the magnetic vector potential A(r) becomes, A(r)

=

^

The first term caused by r

presents

+

— 4ir

V 1 x M(r') /r-rV

of

the magnetization

points

from

dv' .

(3)

in our application the magnetic

the bioelectric activity

contribution vector

/r-r'/

4tt

dv 1

and of

term

the second term

the

the biomaterial.

The

the chosen origin

to the measurement

point and the vector r' to the source element dv'. Hence knowing the total current density tissue field

enables us

law.

to calculate the corresponding

biomagnetic

(only j t is now considered to cause the field outside) B(r) =

Eq.

of the biological

" 4tt

x j. (r') /r-r'/

dv' .

(4)

(4) is in fact equivalent to the well known

Biot-Savart 1 s

Since a more detailed analysis on the generation of bio-

magnetic field is given by Plonsey in this volume

/13/,

only

a brief introduction is given here. It is assumed that ductor,

the biomaterial is an electric volume con-

which is at least piece by piece homogeneous.

depicts the situation schematically. 3

n-

Fig. 1

The volume conductor

is

6 Fig. 1. A simplified cross section of a biomaterial. It is assumed that the impressed current density j. due to the bioelectric activity is located in a volume v., resulting also in a veil ume current j . The total current density flows inside the outer surface S,, of the biomaterial. Mathematically, the boundaries between different conductivities can be taken into account by assuming secondary currents K s . different electric conductivities are the surfaces S^/ i = 1,2,...,N, only one of them (Sj) is shown in the figure. The impressed current density gives rise to volume current j in the whole volume conductor. Eq. (4) is,

The total current density

of

l t = li + j v •

(5)

Actually, the impressed current density consists only of transmembrane currents in the active cells, but often all the intracellular currents are constrained in this term. Since as any other vector field, is completely determined by its divergence and curl, it has become customary to divide the current densities correspondingly in flow sources (F) and vortices (V) /13/, It

=

•F

+

li

V +

V x jj = V x £ Since

in

2V

F

+

V ^V '

= V •

where

(6

>

= V • ¿J = 0 .

a quasistatic approximation

total current density vanishes,

the divergence

(7) of the

7 v

• J.t = V • 1 ± + v • 2 V = V

.F ¿1

+

V

lv =

0

'

(8)

and E is approximately irrotational, V x E =

3B

the electric field tial V (r) =

1 4 ira

^ 0 , E

(9)

can be obtained

/r-r'/

from the scalar poten-

dv'

(10)

Correspondingly, the magnetic field is obtained from: V, " x dv' /r-r'/

(11)

,F = homogeneous volume conductor V x 0 an 5 Hz). The dotted line of Fig. 11 is the noise spectrum of the same instrument in our remote measurement site. The difference of these two curves at low frequencies is due to the increased noise level in HUCH. The noise level at the line frequency and also at its harmonics is somewhat increased, too, but those probably can be removed with the aid of a comb-filter /26/. Typical results Fig. 12. upper

cardiac

The uppermost curve

MCG curve

26.5 Hz

from

measurements is the

are

ECG II for timing.

is a real time measurement result,

disturbing field was on.

shown

It shows

in The

when the

considerable con-

tamination also from 50 Hz. The lower MCG curve is the average of 50 signals. Its quality is acceptable for clinical studies.

23

1 mV

ECG n

10

b) 25 pT

25 pT

MCG E L

MCG £L

10'

FREQUENCY (Hz)

Fig. 11. The noise spectrum of a second-order gradiometer /2 7/ in a hospital (solid line) and in a magnetically quiet site (dotted line). The gradiometer is asymmetric, the diameter of the main sensing coil is 2.5 cm. The equivalent energy sens^ tivity obtained was 2-10 J/Hz above 5 Hz, except at the frequencies of the disturbance peaks seen in the figure, when a laboratory made point contactSQUID was used at 20 MHz. The noise is partly due to the rfshield surrounding the gradiometer.

Fig. 12. Results from cardiac measurements in a) The ECG II is used for timing. b) A real time MCG, the measurement band is from Hz. c) An average of 50 MCG complexes.

a hospital. 0.0 5 to 100

24 Since the main disturbances of the MCG of ic,

it is likely

remove

Fig. 12 are period-

that a suitable data handling procedure can

this noise contribution

and

thus

make averaging un-

necessary . For practical

real time magnetocardiography,

the noise level

should be below about 50 fT//Hz". This is easily achieved using first order gradiometers sites.

in

magnetically

It is not so easily achieved

quiet

measurement

in noisy laboratories or

hospitals even though second-order gradiometers were used (see also Ref. 26).

Comments on the Measurement Policy It was stated before that the electromagnetic inverse problem cannot generally be solved uniquely. In electric vectorcardiography, a corresponding problem of nonuniqueness of the inverse solution is avoided by the use of an equivalent dipole current source or by a more complete expansion of sources. The magnetic multipole expansion

is a unique equivalent pres-

entation of the measured magnetic field. The total measured field B(r) is presented with the aid of the even and odd "unit © o fields" H ^ of various multipoles n, multiplied by their pole strengths A„„, 3 nm or C nm: y » n (r) = — H e —(r) + Cnm—mn H° —(r) } . B Ly i y {A — nm—mn n ri 4. TT n=0 m=0

v

(22) '

The pole streng-ths up to the ocfcupole term (n=3) are listed in Ref. 35. The reference also shows examples where multipole expansions up to the octupole term were made for cardiac magnetic fields. We consider sion

the application

of the magnetic multipole expan-

to the magnetization effects of Table 1.

We repeat here

25 the dipole strengths (n=l): A

10

A

11

c

ll.

M z (r') =

M x (r')

m

d V

=

z" m x

My (r')

and the quadrupole strengths (n=2)

The

A

20

2z' -x* -y.

A

21

x'

C

21

A

22

0

0 z' 1 , 1 , 2* -2Y'

C

22

0

¿x' 2

=

multipole

y'

z' 0

strengths

M z (r')

independent measurements, contain all information The nonuniqueness

(24)

M (r') y -

describe

magnetization distribution.

d 3r ,I

M x (£')

various

In the limit

moments

of

of large number

the multipole strengths

the of

calculated

available from magnetic measurements.

of the inverse solution

fact that all moments cannot be solved.

is manifest in the

Thus

M(x,y,z) cannot

generally be calculated from experimental data. It is

worth

studying

one

special case,

magnetization effects of Table 1. magnetizing field

HQ

applicable

to the

We assume that the external

is homogeneous

and oriented

along the

z-direction. Then the only nonzero dipole component, A

10

=

m

z

- J

M z (r') d r'

equals total magnetization. tween the total magnetization

(25)

Quite often

the relationship be-

and the total amount of magnet-

izable material can be expressed as /16/,

26 = aHQ A. 10

p (r') d r' . a

The coefficient

(26)

is either known

(e.g. for blood) or it has

to be determined from calibration measurements ious contaminants in the lungs). Cij

(e.g. for var-

Other coefficients

A^

and

give information on the distribution of the magnetization

and hence on the distribution of the magnetizable material. PRE-

Fig. 13. The mapping of the biomagnetic field produced by the heart or by contamination in the lungs.

Most magnetometers are designed

in such a way that they natu-

rally measure the component of the magnetic field lar

to

the

(plane)

surface of the subject.

plane is then a natural choice, be shown that all components culated accuracy

as shown in

A i j and C ^

were measured

or

if measurements

Fig. 13.

It can

can be uniquely cal-

from the data measured in a plane. would be further improved

perpendicu-

A mapping in a

In practice,

the

if other field components were performed

in several

planes /16,35/. For the mathematical treatment be valid, measured

in principle,

of the multipole expansion

all the experimental data

outside a sphere

to

should be

containing all the sources

of the

magnetic field. Thus another natural choice is the measurement on the surface of a sphere.

Again, the perpendicular

field component contains all the magnetic

information.

(radial)

27 For certain geometries, the normal component

including

the plane

of the magnetic field

not influenced by the volume currents source is an impressed current dipole that

the bounded volume conductor

simple model works reasonably well conductors

and the sphere, on the surface is

, when the bioelectric /36/.

It is assumed

is homogeneous. also for

is shown in several papers

That this

practical volume

(see e.g. Refs. 22,37,

38,39) . Then the electric surface potential and the magnetic 'surface field' measurements are expected to show some orthogonality. For a homogeneous volume conductor, the electric and magnetic fields caused by a current dipole are everywhere orthogonal. The vortex current sources and volume currents of inhomogeneous volume conductors (Eq. 14) may change this picture. However, the orthogonality is often approximately valid, as has been experimentally shown in the cases of the cardiac electric and magnetic dipole sources /40,41/ or in the cases of the auditory evoked potentials and fields /38/, etc. Thus one may plan the experiments expecting approximate orthogonality between the electric and magnetic measurements.

Comments on Standardization Comparison of the results obtained by different research groups is easier, if certain standardization rules are accepted. Since research on biomagnetism is relatively young, it is not to be expected that invariable rules could be made for the future. Nevertheless, some proposals are made in the following. The sign of the magnetic flux

should be positive

is directed towards the subject tion is not free from ambiguity,

to be measured.

if the flux This defini-

but it seems to be the defi-

nition most often accepted in publications.

28 The type of measurement instrument coil magnetometers, main sensing coil

should be

the coil configuration,

mentioned.

For

the size of

the

and the distance between this coil

and the

biological surface should be stated. The type of measurement site shielded

quiet site,

laboratory or

Whenever an external magnetic field

has been used

or

expected

is

magnetically

(magnetically

hospital). it

enclosure,

should be defined

to have influence on

the results,

this

should be mentioned. Generally, the same frequency band should be used as in corresponding

electric measurements.

beneficial,

Extension of the band may be

while reduction of the band

facts in recordings.

may even cause arti-

The measurement band should be stated in

publications. Whenever possible, attention

primary data should be presented.

should be paid to the scales

of

Special

the data,

the SI

unit system should be used. For

cardiac magnetic measurements,

standardization

have been made

the measurement grid

some

/42/.

more proposals

for

An improved version of

for cardiac studies

is given

elsewhere

in these proceedings /4 3/.

Conclusion Almost all biomagnetic measurements have potential medical and clinical applications. Since conventional and accepted electric measurement techniques already exist for most bioelectric phenomena magnetic measurements will have to display definite advantages to be able to compete with electric measurements. One disadvantage is immediately obvious s the magnetic measurements are technically more difficult. Extremely small signals have to be measured in the presence of much stronger external

29 disturbing fields. Magnetic measurements show definite advantages as well. They are a new source of information. In some cases, data are obtained on phenomena which are not electrically measurable at all. The magnetic measurements are performed without any physical contact between the subject and the measurement instrument. The measurement band can be extended down to quasi-DC. In true field measurements no reference point is required. However, it is too early to predict, which biomagnetic measurements will eventually compete successfully with electric measurements.

Acknowledgements The author expresses his gratitude to MD Riitta Hari, Dr. P. Karp, Dr. R. Maniewski, MD P. Siltanen and Mr. T. Varpula for useful discussions and to Miss Maria Lindstrom and Mr. T. Tuomisto for their help in preparing the manuscript.

References 1. Baule, G.M.,

McFee, R.:

Am.

Heart

J.

6£, 95

2. Kariniemi, V. , Ahopelto, J., Karp, P.J., J. Perinat. Med. 2, 214 (1974).

(1963).

Katila, T.E.:

3. Cohen, D.: Science 161, 784 (1968). 4. Cohen, D.: IEEE Trans. Magn. MAG-11, 694 (1975). 5. Brenner, D., Lipton, J., Science 19^, 81 (1978).

Kaufman, L.,

Williamson, S.J.:

6. Reite, M., Edrich, J., Zimmerman, J.T., Zimmerman, J.E.: Electroenceph. clin. Neurophysiol. 4_5, 114 (1978) . 7. Cohen, D.,

Givler, E.:

Appi. Phys. Lett. 21, 114 (1972).

8. Karp, P.J., Katila, T.E., Mäkipää, P., Saar, P.: Digest of the 11th International Conference on Medical and Biological Engineering, p. 504 (Ottawa 1976).

30 9. Aittoniemi, K., Katila; T., Kuusela, M.-L., Varpula, T.: Digest of the 12th International Conference on Medical and Biological Engineering, ch. 94.6 (Jerusalem 1979). 10. Cohen, D.: Science 180, 745

(1973).

11. Wikswo, J.P. Jr., Opfer, J.E., Conference Proc. 18, 1335 (1974).

Fairbank,

12. Harris, J.W., Farrell, D.E., Messer, M.J., Brittenham, G.M., Danish, E.H., Muir, W.A.: 26, 504A (1978).

W.M.: AIP Tripp, J., Clin. Res.

13. Plonsey, R.: Proc. of this confer. 14. Geselowitz, D.B.: Biophys. J. 7, 1 15. Horacek, B.M.s

(1967).

IEEE Trans. Magn. MAG-9, 440

16. Aittoniemi, K., Kalliomäki, K., Proc. of this confer.

(1973).

Katila, T., Varpula, T.:

17. Livanov, M.N., Kozlov, A.N., Korinevskij, A.V., Markin, V.P., Sineijnikova, C.E., Holopov, Ju.A.: On the registration of human magnetic fields, Academy of CCCP (1978). 18. Lekkala, J., Malmivuo, J.: Proc. of this confer. 19. Denis, B., Matelin, D., Favier, Ch., Tanche, M., MartinNoel, P.: Arch. Med. Coeur 69, 299 (1976). 20. Zimmerman, J.E. : J. Appl. Phys. 48:, 702 21. Verster, N.F.: Appl. Sei. Res. Bl, 363

(1977). (1949).

22. Williamson, S.J., Kaufman, L.: Proc. of this confer. 23. Cohen, D.: Data presented (Florence, Italy 19 78) . 24. Farrell, D.E.:

at

the

INTERMAG-conference

Private communication,

25. Varpula, T.: Diploma Technology (1979) .

Thesis,

see also: Ref. 39.

Helsinki

26. Barbanera, S., Carelli, P., Leoni, R., Bordoni, F., Modena, I., Fenici, R., Proc. of this confer. 27. Kuusela, M.-L.: Diploma Technology (19 79) . 28. Stroink, G. Brown, Proc. of this confer.

B.,

Thesis,

University

of

Romani, G.L., Zeppilli, P.:

Helsinki University of

Blackford, B.,

Horacek, M.:

29. Stuart, C.I.J.M.: Private communication, see also: Stuart, C.I.J.M., Woods, S.B., Crawford, A.C., Thomas, N.R., Newbound, K.B.: Proc. of this confer. 30. Voss, R.F., Laibowitz, R.B., Ketchen, M.B. Broers, A.N.: Proceedings of the 2d International Conference on Superconducting Quantum Devices, Berlin 1980 (to be publ.).

31 31. Ahopelto, J., Karp, P.J., Katila, T.E., Lukander, R., Mäkipää, P.: Biocapt 75, Paris 1975, 1, 347. 32. Karp, P., Duret, D.: J. Appi. Phys. 51, 1267 (1980). 33. Järvinen, M., Katila, T., Maniewski, R., Varpula, T.: Proceedings of the IV National Conference on Biocybernetics and Biomedical Engineering (Poznan, Poland 1980) , to be pubi. Katila, T., Kuusela, M.L., 34. Aittoniemi, K., Karp, P.J., Varpula, T.: J. Physique 39, C6-1123 (1978). 35. Karp, P.J., Katila, T.E., Saarinen, M., Varpula, T.T.: Circuì. Res. (to be pubi.). 36. Grynszpan, F.: U.S.A. (1971).

PhD Thesis,

Siltanen, P.,

University of Pennsylvania,

37. Aittoniemi, K., Järvinen, M.-L., Katila, T., R., Varpula, T.: Proc. of this confer. 38. Aittoniemi, K., Hari, R., Järvinen, M.-L., Varpula, T.: Proc. of this confer. 39. Farrell, D.E., Tripp, J.H., VanDoren, C.L.: confer.

Maniewski, Katila, T.,

Proc. of this

40. Wikswo, J.P., Malmivuo, J.A.V., Barry, W.H., Leifer, M.C., Fairbank, W.M.: Adv. cardiovasc. Phys. 2, 1 (1979). 41. Denis, B., Machecourt, J., Favier, C., Ann. Cardiol. Angeiol. 27_, 81 (1978).

Martin-Noel, P.:

42. Saarinen, M. , Siltanen, P., Karp, P.J., Ann. Clin. Res. 10, S21, 1 (1978). 43. Karp, P.: Proc. of this confer.

Katila, T.E.:

CONSTRUCTION AND PERFORMANCE OF THE OTANIEMI MAGNETICALLY SHIELDED ROOM

Väinö 0. Kelhä Technical Research Centre of Finland, Instrument Laboratory, Metallimiehenkuja 8, SF-02150 Espoo 15, Finland.

Abstract A new magnetically shielded room has been constructed in Otaniemi, Finland.

Three different methods have been employed.

Ferromagnetic shielding has been made by three mumetal shells , eddy current shielding by six aluminium layers, and active shielding by a flux-gate controlled coil system. Construction has been discussed in detail.

The shielding factor for static

field is 10 4 , for 0,2 Hz 1,6-10 4 , for 0,7 Hz 1,8-10 5 and for 10 Hz more than 2-10 5 . for 0,1 Hz, 4 0 fT/Vhz

The noise in the room is 2 pT/\/Hz for 0,5 Hz.

Above 0,5 Hz the room

noise level is below that of the SQUID magnetometer used. Because of the low noise level the room makes an excellent facility for the study of extremely weak magnetic fields of very low frequencies.

1.

Introduction

According to the original plans the construction of the Otaniemi magnetically shielded room (MSR) in the Low Temperature laboratory of Helsinki University of Technology was to be accomplished in 1979.

Owing to the delay in the delivery of

the mumetal sheets the construction work could not be started until December 1979.

In April 1980 five months later, the

room was complete and tested, (Fig. 1). The design and

© 1981 Walter de Gruyter & Co., Berlin • New York Blomagnetism

34

Fig. 1

The Otaniemi Magnetically Shielded Room

construction of the HSR were made by the Instrument laboratory of the Technical Research Centre of Finland. The object of the design was to make an enclosure big enough for human beings to work inside, approximately 2,4x2,4x2,4 m^ , within the shielding factor for slowly varying magnetic field

(the ratio of the magnetic fields outside and inside) S=400 for 0,1 Hz and S=10

for 10 Hz to quarantee the feasibility

of the most sensitive magnetic measurements made with SQUIDmagnetometers.

35

200

Fig.

The

2

400

800

S

T h e m a s s o f m u m e t a l as a f u n c t i o n o f t h e s h i e l d i n g f a c t o r S in d i f f e r e n t m u l t i - l a y e r e d s t r u c t u r e s , (d^j = t h i c k n e s s o f t h e i n n e r m o s t s h e l l - I n o p t i m u m point for S=400 the dimensions of the cubes are 1-1=2,5 m , 1 2 = 2,8 m , 1 3 = 3,3 m , d 1 = d 2 = d 3 = 2 , 0 m m ) .

shielding

enclosures

factor was

/1,2,3/

the mass

for

analysis

1-, 2-, and

of the

/4,5,6/.

shells

in Fig.

superior

alternative was the

MSR.

2.

Construction

2.1

Shielding

2.

The threelayer

of The Otaniemi

methods

is

shell-thickmess

for the c o n s t r u c t i o n

MSR.

of

shielding

structure

to the o t h e r s , and an equal

c h o s e n as t h e b a s i s

3-shell were

The comparison

of m u m e t a l l r e q u i r e d to o b t a i n a c e r t a i n

is s h o w n

obviously

calculated

and the dimensions

optimised by numerical factor

600

of

36 Three different methods have been employed to produce a space of low magnetic field?

Fig. 3 /7/.

Ferromagnetic shielding

has been accomplished by enclosing the space by three cubic shells of mumetal.

The mumetal plates have been riveted

between six aluminium shells whose shielding effect is due to the induced eddy currents.

The third method is called active

shielding and involves control by negative feedback of currents

REASING

OF p

RROMAGNETIC EDDY CURRENT ACTIVE

Fig. 3.

(SHAKING)

SHIELDING SHIELDING

SHIELDING

(MUMETAL ) (AL)

( C L O S E D LOOP CONTROL)

Different shielding methods.

37 in coils around the room to counterbalance the slowly varying external field. 2.2

Shell construction

The magnetic shield is situated in a special concrete room 3 7x7x5,6 m , reinforced with stainless steel to avoid the inhomogeneous magnetic disturbances excited of normally reinforced walls. Each shell is made of two crossed layers of 1,0 mm thick mumetal sheets squeezed and riveted between 5- and 2 mm thick aluminium plates with 05x16 Al-rivets (stainless steel nails). The dimensions of the shells were determined by the availability of mumetal sheets long, wide, and thich enough. Therefore the dimensions of the cubes are 2450, 2800, and 3150 mm, respectively, instead of optimum dimensions (Fig. 2), and all the walls have been made of 3 50 mm wide mumetal sheets delivered by Vacuumschmelze GmbH.

The quaranteed incremental

permeability in the field 0,4 A/m was

= 30 000.

However,

the measured value was between 10 000 and 70 000, and the average relative permeability of the sheets in the innermost shell 21 000, in the middle shell 23 000, and in the outermost shell 51 000.

The total mass of the mumetal used in MSR is

3000 kg and the mass of the aluminium about 4000 kg /7/. The joints in the edges of the shells were made by using mumetal corner-pieces (1+1) (95x95x1500 mm) manufactured and annealed by the deliverer.

The corners were squeezed by

aluminium 90°-corners and riveted.

The 9 5 mm overlap of

mumetal along the edges accomplished a joint with very low reluctance, thus avoiding degration of the overall shielding. 2.3

Eddy-current shielding

38 The induced eddy-currents are generated in the six aluminium shells between which the mumetal plates have been riveted. The shielding effect caused by the eddy currents has been calculated to be 22 dB for 1 Hz disturbance and 75 dB for 10 Hz, if the shield is a closed cube without doorway. 2.4

Shaking

The effective permeability of the mumetal can be increased by shaking /8,9,10/.

The process of shaking involves superimpos-

ing a relatively strong (Hs = 2...5 A/m, rms) alternating magnetic field on the disturbance field in the shield.

The

shaking field is generated by the shaking coils wound along 2

the edges of the three cubic shells (12 turns, 0,8 mm every edge).

along

The shaking is made in two perpendicular

directions with line frequency SO Hz.

The coils can also be

used for the demagnetization of the walls.

By using a shaking

current of 6 A, the shielding factor of the MSR increased by the amount of 5,6 dB for a 0,5 Hz disturbance. 2.5

Active shielding

The system for active compensation of the external disturbing field consists of two coils surrounding the MSR (4800x4800 mm 2

square, 15 turns each, 2,5 mm ), a flux-gate transducer outside of the room, a PID-controller and a power amplifier 111.

The location of the flux-gate transducer has been

carefully chosen in order to reach the balance between the compensating field, created by the coil current, and the disturbance simultaneously on the roof of the shield and inside the MSR.

The frequency range of the controlling system

is f=0...10 Hz and the compensation range 0...27 000 nT. active shielding increased the shielding factor 3 5 dB for 0,1 Hz disturbance and 20 dB for 1,0 Hz.

The

39 2.6

Supporting structure

The framework of the MSR has been made by welding the 5 mm aluminium plates of the walls together along the edges 111. The walls have been strengthened by aluminium z-beams 5x50x175 mm, and the walls have been bolted together with stainless steel bolts A1S1 316, M 8 x (25...50).

Both the

magnetic shield and the floor inside are mounted directly on the bedrock in order to isolate the room from vibrations.

The

wooden inner floor with aluminium frame lies on 4 feet, which are also used as inlets for fresh air and electric wires. 2

The doorway dimensions are 700x1400 mm construction.

and the door is under

On the frequencies higher than 1 Hz a properly

made mumetal-door is necessary and will increase the shielding factor by 60 dB for 100 Hz disturbance. 3

Performance of the Otaniemi MSR

3.1

Shielding factor measurements

All the shielding factor measurements were made with the aid of a known magnetic signal source, consisting of a large multiturn (N=300) coil that was fed at various frequencies by one kilowatt DC-coupled power amplifier. the source to the room was

The distance from

15m.

To obtain reliable reference values for shielding factor calculations a series of field measurements were carried out at the site of the MSR before any y-metal was brought in. At low frequencies (< 2 Hz) the measured field strength proved to be consistent with the values calculated from the dipole radiation formula.

At higher frequencies the measured values

were higher than the calculated ones.

The measured values

were used in the shielding factor calculations.

40

Fig. 4

The measured shielding factors of the Otaniemi magnetically shielded room as a function of frequency a) one shell, b) two shells, c) three shells, d) three shells and one door, e) three shells and two doors, and f) three shells and three doors, g) three shells and three doors with active shielding. Each shell contains two layers of y-metal and aluminium.

Magnetic fields were measured with a fluxgate magnetometer and a SQUID magnetometer.

The fluxgate was sensitive enough in

41 the measurements of the first and second shells, but after the construction of the third shell the SQUID-magnetometer was the only feasible device. Measurements for the first and second shell were also performed with the SQUID system and the results conformed with those of the flux-gate measurements. When measuring the highest shielding factor with the temporary doors closed the signal source had to be removed closer the room to distinguish the signal from the magnetometer noise. The SQUID-magnetometer equipment had a total equivalent flux noise of 8-10 (J> /v/ifz, where (f>o is the flux quantum. This corresponds to a field noise of 36 fT / \/hz . The results are shown in figure The shielding factors at different frequencies were measured always immediately after the installation of a new shell. It is seen that the performance of the room above 0,4 Hz is limited by the doorway when the doors are open. With all the doors closed the calculated values given in the introduction are reached. Using active shielding the improvement in the shielding factor at 0,1 Hz has been demostrated to be 3 5 dB. Static field inside the complete room was more than 1200 nT with a direction opposite to that of the earth's field. After careful demagnetization of the innermost shell with a portable demagnetizing coil and of the other two shells by a strong shaking current, the internal field was reduced to 5 nT. This corresponds to a static shielding factor of 10000. The history of the static field within the room is presented in Fig. 5.

42

Fig. 5

The history of the remanent static magnetic field in Otaniemi MSR.

3.2

Noise measurements

The ambient magnetic noise for high and low daytime noise levels is seen in figure 6.

The corresponding noise level

inside the MSR with open doors are seen in figure 7.

A

typical measured noise spectrum is also given-

pT/vHz

Fig. 6

The ambient daytime magnetic noise outside of the Otaniemi magnetically shielded room for a) low, and b) high measured noise levels.

fT/VHZ

10 5 -

10 4 -

10^-

102-

10

0,1 i 0,1

. 7

1

1 1

1 10

1— 10 2

10 3

f/Hz

The internal daytime magnetic noise in the Otaniemi magnetically shielded room with the doors open. Curves a) and b) correspond to low and high ambient noise in fig. 6, and curve c) gives a typical measured noise spectrum limited by the SQUID magnetometer noise above 100 Hz. Curve d) shows the noise spectrum of the squid magnetometer and e) the estimated thermally induced noise.

45

Fig. 8 a) The internal daytime magnetic noise in the Otaniemi magnetically shielded room with closed doors and with active shielding. Curves b) and c) correspond to low and high ambient noise in fig. 6, without active shielding and curve d) the noise spectrum of the SQUID.magnetometer. Curve e) shows the estimated thermally induced noise. Below approximately 50 Hz the noise inside the room stems from the attenuated ambient noise.

When the doors are closed the

46

Fig. 9

The decrease of the noise by using the active shielding. a) Passive shielding b) Passive shielding with active compensation.

noise spectra reduce to those seen in figure 8.

Above 0,5 Hz

the room noise level is below that of the SQUID magnetometer except the 50 Hz peak of 200 fT.

The decrease of the noise

by using active shielding is clearly shown in fig. 9.

The

noise level near 0,1 Hz can be further decreased by shaking and by more careful adjustment of the active shielding. Thermally induced currents in the aluminium walls generate thermal magnetic noise in the MSR.

A rough estimate of the

level of this noise is also given in figures 7 and 8. Measurements made with a small-scale model show that these values are overestimated by at least a factor of three. The comparison of the external and the internal noise of

47 Otaniemi MSR is shown in fig. 10.

Fig. 10 Comparison of the external and internal magnetic noise of the Otaniemi MSR.

48 PT 10 •

, L U N G CON TAMINATIOf*

10 MC G

MIV G

10 •

FMC G MEG SE F, AEF VEF MRG

0,1

MSR

0,01

1

0,1 Fig. 11

3.3

10

10

10

t/Hi

The amplitudes of various biomagnetic phenomena and the bandwidths needed to measure them are indicated. Also, the noise level expressed per V~Hz vs. frequency in the Otaniemi magnetically shielded room with active shielding is shown.

Other measurements

The high frequency performance of the MSR was tested by measuring the signal strength of some nearby radio stations inside the room.

The level of some VHF-stations near 100 MHz

49 decreased only 12 dB when the receiver was moved into the room using a short piece of wire (0,5 m) as antenna.

When one door

was then properly closed the signal dropped below the -110 dBm (50 fi) noise level of the spectrum analyzer.

No signal

exceeding that level could be detected between 10 MHz an 1 GHz.

4.

Conclusions

Because of low internal noise the Otaniemi magnetically shielded room is well suited for magnetometer testing, demagnetization studies, measurements of geophysical samples, extremely sensitive physical experiments, measurements of weak magnetic fields produced by natural ion currents in human beings such as the field of heart, brain, and muscles, and mapping magnetic contamination in humans.

The frequency

ranges and field amplitudes of some of these signals are given in figure 11.

The MSR is also suitable for studies that

require effective shielding against radiofrequency signals.

References 1.

Cohen, D., Large-volume conventional magnetic shields. Revue de Physique Appliquee, _5, Fevrier, 53... 58 (1 970 ).

2.

Cohen, A., A Shielded facility for low-level magnetic measurements. J.Appl.Phys. _38, 3 , 1 295-6 ( 1 967 )

3.

Hager, A., Magnetic shields, IEEE Trans. MAG., MAG-6, 1 67-75 (1970)

4.

Lequain, P., Berkhout, J.A. and Schouten, J., Design and construction at ESTEC of a magnetically shielded enclosure for VLF, ESA TN-125, January (1977)

5.

Kaden, H., Wirbelströme und Schirmung in Nachrichtentechnik. Berlin, Springer-Verlag. OGH (1959)

6.

Schweizer, F., Magnetic shielding of a system of concentric spherical shells. J.Appl.Phys. 3 , 1 001 ... 1 003 (1976)

50 7.

Kelhä, V., Peltonen, R and Pukki, J., Construction of a magnetically shielded room, Espoo 1980, Technical Research Centre of Finland, Instrument Laboratory, Tiedonanto 4, 1 ... 45 (1980) (in Finnish)

8.

Kelhä, V., Peltonen, R. and Rantala, B., The effect of shaking on magnetic shields. To be published IEEE Trans. Mag., July (1980)

9.

Cohen, D. , Enhancement of ferromagnetic shielding against low-frequency magnetic fields, Appl.Phys. Letters, 10, 67...69 (1967).

10. Bozorth, R., Ferromagnetism. Princeton, N J, D van Nostrand Co, p. 549 (1968)

THE BERLIN MAGNETICALLY SHIELDED ROOM (BMSR), SECTION A: DESIGN AND CONSTRUCTION*^

A. Mager

VACUUMSCHMELZE GMBH D-6450 Hanau, W-Germany

1. Introduction A magnetically shielded room was designed and constructed for the Physikalisch-Technische Bundesanstalt Institut Berlin

' by VACUUM-

SCHMELZE GMBH Hanau in close co-operation with Prof.H.D. Hahlbohm and his co-workers to measure extremely weak magnetic fields with SQUID-magnetometers. The present section A describes the design and construction, section B the performance /1/ and section C the periphery /2/. The shielding effect of the room extends from DC to GHz fields. An optimum design under technological and economical aspects was found in a cube shaped construction. The inner dimensions of the cube are 2.25 x 2.25 x 2.25 m 3 , the outer dimensions 4.6 x 4.6 x 4.6 m 3 . The entire shield has seven shells, the inner shell acts as an eddy current shield and is constructed of 15 mm welded copper plates, followed by six concentric magnetic shells of double-layered high permeable sheets and non-magnetic spacings (foam-plastic) between the magnetic shells. Formulas to calculate the shielding factor are presented and a first measured value of the shielding factor is given - further details in /1/. The shielding factor is much higher than that of existing magnetically shielded walk-in rooms /2-10/. The new shielded room is suitable for bio-magnetical measurements and can solve delicate EMC-problems in the field of medicine. in An extract of this paper was presented by L.Borek, VAC Hanau to the Third Workshop on Biomagnetism, Berlin. 1980 **)

Physikalisch-Technische Bundesanstalt (PTB) Institut Berlin, Berlin 10, Abbestr. 2-12; by order of Bundesbaudirektion

© 1981 Walter de Gruyter & Co., Berlin New York Biomagnetism

52 2. History Several magnetically shielded walk-in rooms with ferromagnetic shells were built in the years 1962 /3/ and 1967/70 /4,5/ in the USA. A very high shielding effectiveness was reported by D. Cohen in 1970 /7/ at MIT's Francis Bitter National Magnet Laboratory on a tri-layered magnetic shield with a shielding factor of 900 using the method of shaking, but only about 150 without shaking /II,12/. The first magnetically shielded room with high permeability nickel-iron alloys in Europe was built by VACUUMSCHMELZE GMBH Hanau (VAC) for the Technische Hochschule Darmstadt, Institut fLir Arbeitswissenschaft /8/.

(ft

VAC also delivered sheets of MUMETALI^for the magnetically shielded room in Otaniemi, Finland /9/. The principles of the design were discussed with V. Kelha et al. and the construction of the shells with two crossed layers was taken from the Darmstadt room /8,13/.

3. The Required Shielding Factors Shielding against electrical noise or electromagnetical noise with higher frequencies is relatively easy with conducting materials, like eddycurrent shields of Cu or Al. The higher the frequency the higher the demands on the sealing of the enclosing conductive shell. The shielding of low frequency magnetic fields is much more difficult. The different methods of shielding against magnetic noise are: shielding with soft magnetic materials, very thick and heavy conducting shields (ineffective for static magnetic fields), active shielding with compensating currents and superconductive shielding. For a walk-in magnetically shielded room the last method seems too expensive, see discussion in /5/. A combination of all the other conventional methods of shielding was requested by the PTB to reach a high shielding efficiency. The design and the construction of the magnetic and the conductive shield was assigned to VACUUMSCHMELZE GMBH. Magnetic noise has different sources in the environment /14/. In a

53 laboratory like PTB Berlin with big power supplies and magnetic highfield measuring systems, the disturbing magnetic noise is relatively high. In such an urban environment there is an additional noise level, especially caused by the subway system less than a kilometer away. Nevertheless the field disturbances in PTB /2/ are not higher than the fields measured in the laboratory of VAC /8/ with maximum values of approx. + 0.1 to 0.2 A/m in the vertical direction. At night - without trains running and other activities - only a tenth of these values was measured; the 50 Hz noise-field was about ten times as high as the above given values /2/. This is not very critical. The ac rms noise at the location of the MIT magnetically shielded room during an average evening is at about 0.01 A/m a little better /5/. With the intended measuring sensitivities and the given noise values the following minimum shielding factors without shaking should be warranted by VACUUMSCHMELZE (Table 1): frequency range

shielding factor S

in Hz

equal or more than

0

-

2

1 000

2 - 1 0

3 000

10

-

20

10 000

20

-

104

30 000

104

-

105

100 000

Table 1: The warranted shielding factors Initially the PTB requested higher values, especially for the low frequency region, e.g. a shielding factor of about 5 000 in the frequency range 0.01 to 1 Hz. With the relatively poor results of other constructors and based on our own experiences with the Darmstadt room it appeared to be unrealistic to warrant this value for the extrapolated low frequency shielding factor. There were numerous uncertainties: the field amplitude in the inner shells is extremely low, nobody has exactly measured the initial permeability in this range; the large dimensions and the very high weight of the shield give rise to mechanical stresses in the sensitive magnetic

54 material; the heat treatment of large sheets from 2 m to more than 4 m in length required extreme precautions. Therefore, the warranted values could not be set at a higher level. Nevertheless VAC has attained higher values as a result of tenacious work in research and development.

4. Mechanical Magnetic Noise, Thermal Induced Stress and Problems with Thermoelectric Currents The disturbing fields in the interior of the room are not only determined by outside noise and the shielding factor, but parts of the shielding walls themselves can be sources of disturbances. A magnetic shell with a magnetic flux from the outside should not be mechanically disturbed. Mechanical shocks or vibrations of parts of such a magnetic wall can cause corresponding magnetic shock and vibration fields. This is most critical for the outermost shell and therefore,less critical for the interior room. But with unavoidable inhomogeneitieE in the shields (edges, connections, mechanical stresses and others) there are locally enhanced values of remanent magnetism in the material. In order to attain a low noise level, vibrations of the walls have to be avoided, i.e. the magnetic sheets should be well fixed on the bearing material, and the bearing wall-elements should have adequate rigidity. Moreover the activities of the people in the room should not affect the magnetic parts of the shielding, the floor should be supported separately from the shielding walls /5,8/. The relatively high sensitivity of the high permeability alloys to all mechanical stresses requires that these stresses are low enough not to diminish the high permeability of the material on the one hand and not to produce a perturbation field with variable stresses on the other hand, in a similar manner as for vibrating parts of the wall. Relatively high stresses can be effected by different coefficients of thermal expansion of two materials which are mounted together. To avoid this the coefficients of thermal expansion should only have small differences /8/, and the variations of the temperature should also be

55 small. When under construction the temperature of the room should be the same as the working temperature of the completed shield. Other possible influences of temperature gradients are perhaps the thermoelectric currents in closed electric loops between different metals. When the shielded room was under construction the chief designer and leader of the working groups was greatly alarmed by experiments of PTB on the first magnetic shell of the room. Placing a hand on the outside of the magnetic shield was followed by a relatively big field variation around this spot on the inside of the room. Not quite as big a variation was also found with thermal influence from the inside of the room. At first thermoelectric influences were thought to be the cause. An electrical

insulation

between the conducting shield and the first magnetic shell, all around the inner copper cube, was inserted. But there was no detectable improvement. It is supposed that the observed effects were due to the stress sensitivity of permalloy in the state of relatively high flux-density caused by the highly concentrated earth-field in that shell as mentioned above. A reduction of this effect was expected with each shell that was added to the shielding - realizing this.the problem appeared to be solved.

5. Thermomagnetic Noise Fields in a Conducting Cavity Another problem which to be considered was the magnetic field of the thermal noise inside a conductive cavity like the shielded room. Here only some results of the estimates from a more elaborate internal report are given /15/. In the center of the inner room and for very low frequencies f (e.g. for frequencies lower than about 2 Hz, with a copper wall-thickness d=0.015 m and an edge-length of a=2.25 m), the amplitude of the mean field TT in A/m rises with the square root of the frequency f as follows: R = 1(1^1)1/2

f 1/2

§

-23 with a and d in m, the Boltzmann constant k = 1.38 10

VAs/grd, the

temperature T in K (here about 300 K), f in Hz and the resistivity J Ohm m of the enclosing conductive shell.

in

56 With higher frequencies^ reaches an upper constant value: H =

(1AI) M 0 a3

1

/

2

.

with the permeability constant p Q = 4 IT 10" 7 Vs/Am. Approaching the wall and for low frequencies (in our example with frequencies lower than about 20 Hz) the field amplitude rises with the reciprocal value of the distance x from the wall, according to: TT = 0.11 ( A A I The transition to the higher frequencies is a little more complicated than for the fields in the center of the cube. At the higher frequencies -3/2 the increase closer to the wall is steeper and is proportional to x . Rough estimates for a frequency of about 10 Hz are very simple and they are given here for the above mentioned dimensions of a copper cube in Table 2. In Fig. 1 the results are given for frequencies from 0.1 to 1 000 Hz distance from the wall

amplitude of the field

x in meters

TT in A/m

1 (center region)

2 10~ 8

5 10" 8

0.1

2 10" 7

5 10" 7

6

5 10" 6

0.01

2 ID'

max.field,rough values H m a x in A/m

Table 2: The mean thermal noise field amplitude TT in a cubical room with walls made of copper, a wall-thickness of 15 mm and an edge-length of 2.25 m, at about 10 Hz. Estimated according to /15/. The expected maximum values of the thermal noise field in a fixed measuring interval should be about two or three times as high as the mean values, in normal, not too long observation times /15/.

57

Fig. 1: The frequency_response of the thermal generated magnetic mean noise fields H in a cubic copper room with wall-thickness of d = 15 mm andan edge-length of a = 2.25 m, in the center region and for different distances x from the wall. & is the skin-depth. Estimated according to /15/. Noise for total range 0Hz to f With a very optimistic total shielding factor at 10 Hz of perhaps 50 000100 000 (with shaking and compensation) the disturbing fields in the shielded room coming from outside fields should have values of about 1 - 4 10" 6 A/m in the day and 1 - 4 10~ 7 after midnight, the best time no trains running. Even with this optimistic assumption the statistical thermal noise would not interfere with the measurements in the whole room down to a distance of 10 cm from the wall. In the best time after midnight the thermal noise fields in this distance can be higher than the disturbing fields from the outside. For full sensitivity measurements during this time the center region of the room is recommended.

58 6. Magnetic Shielding with Soft Magnetic Materials, Static Magnetic Fields The magnetic shielding effect is based on the high magnetic conductivity of the soft magnetic materials. Contrary to the electrostatic case, air or vacuum and other materials are not insulators. Measured as rel. permeabilities they have the permeability 1 or infinitely small differences compared with this figure, only the ferromagnetic materials (the ferromagnetics included) have permeabilities up to approx. 100 000. This is comparable with the electrostatic case without insulators: shielding in an environment of diluted H ? S0. with a poor conducting material such as Pb. Thus magnetic shielding is more complicated than electrostatic shielding. Other problems are non-linearity, saturation and remanent magnetism. The shields are relatively heavy and expensive. Stringent requirements can only be met with very massive shells or with multiple shields. Single shields also have an upper limit, given by the value of permeability and their outer form (or more specifically: the demagnetizing factor). In shielding the soft magnetic material acts as a parallel conductor to the magnetic flux with low resistance relative to the enclosed room, and only a small part of the flux enters the room. Fig. 2 gives an example of long cylinders in transverse fields, the field lines are calculated for relative permeabilities of \i = 20 and a very high permeability with M — » oo .

a Fig. 2: Cylinders with inner/outer diameters of D^/D Q = 0.7 in transverse fields, with permeabilities of M = 20 (a) and m - > o o ( b ) . The distances between the lines in the air give the field strengths.

59 The shielding factor is the quotient of the undisturbed external field H e to the field inside the shield H^: S = He/Hr It is also usual to use the attenuation: a $ = 20 lg S

(dB).

The shielding factors of some simple shapes are as follows /16,17/: long cylinder (transverse field)

l) frequency dependence. In addition to these "macroscopic" movements the vibrations of the building walls and floors induced by ventilation, street traffic, etc. can bring about magnetic noise at certain frequencies, typically above 1 Hz. These noise sources can be extremely difficult to detect. Automobile, train and tram traffic in the hospital neighbourhood contributes to the magnetic noise: a disturbing field of

97 50 nT can be produced by a small car at a few meters' distance, and as high as several hundred nanoteslas by some electric train and tram power systems at a distance of a few hundred meters (1).

2.1. Magnetic measurements Measurements of magnetic noise were performed with a threeaxis 30 0 MHz SQUID magnetometer system, designed by Duret et al. (2). The SQUIDs were thin-film devices used without a flux transformer and placed in a commercial liquid helium container. In some cases a thin-film ferromagnetic magnetometer (3) was used. The noise level of this device was about -1/2

10 pT-Hz at 1 Hz. The three-axis magnetic noise was plotted in situ and also recorded on a magnetic tape and subjected to an FFT analysis. The hospital is an eight floor concrete building, surrounded by additional smaller buildings (see Fig. 1).

N. Fig. 1. Map of the hospital, showing the measurement sites : A: ground floor B & C: 8th floor D: ground floor, hospital entrance E: nurses' school (basement) F: day nursery (basement)

98 A typical result from a magnetic measurement is shown in Fig. 2. The vertical noise spectra from all the measurement sites are shown in Fig. 3. Only vertical noise components are shown since generally they were superior or equal to the horizontal ones. All the spectra show a low frequency behavior —ci with an f slope (a>l).

frequency(Hz) Fig. 2. Noise recording at site B (8th floor). The lifts were a few meters away and produced a peak noise of about 10 0 nT.

Fig. 3. Vertical noise spectra at different measurement sites. The large amplitude in basement site E compared with F can be explained by the presence of the highway tunnel.

Fig. 4 shows, in comparison, the spectra obtained at the Helsinki University Central Hospital (from ref. (1)). The results obtained are comparable, but the noise seems to be larger there, probably due to the difference in the hospital size.

Fig. 4. Vertical noise spectra obtained at the Helsinki University Central Hospital, from ref. (1) •

fr«qu»ncy(Jz)

99 Fig. 5 shows the vertical magnetic noise measured during a long period from Sunday evening to Monday morning at site D. On Sunday the peak to peak 50 Hz gradient level was about 20 nT/m at this site.

Fig. 5. Magnetic noise over a long period, showing 1) a period of medium activity from Sunday evening to Monday 1.30 a.m. 2) a quiet period from Monday 1.30 a.m. to 4.30 a.m. 3) a transition period from 4.30 a.m. to 7.00 p.m. 4) a full activity period from Monday 7.0 0 a.m.

2.2. Acoustic and seismic measurements The acoustic noise was measured with a commercial sound level meter (weighting A). The results are summarized in table 1. The seismic noise was measured as the vertical component of the speed of the floors. It was recorded from 0.1 to 20 0 Hz and subjected to an FFT analysis. Fig. 6 shows a typical result. € Fig. 6. Seismic speed spectrum at site C. A 3 Hz vibration is also visible in the time record in the upper right corner.

10

20

30

frequency (Hz)

100 The rms values at 10 Hz are summarized in Table 1, which contains also the magnetic results.

site

acoustic speed rms vertical rms magnetic noise -1/2, noise at 10 Hz (nT Hz dBA ymS- W 1 / 50 Hz 10 Hz 1 Hz 0.1 Hz

A

6

1.2

0.07

15

B

13

1.1

0.05

4

C

3

0.7

0.02

6

D

14

0.4

0.02

40

E

8

1.2

0.02

10

F

1.5

0.15

0.01

20

65 (inside) 43 (inside)

2

2 1.5 1

65 61 (outside^ 60

6 0.7 0.6

Table 1. Summary of all measured values. All the parameters are coarsely correlated at same sites.

2.3. Conclusion of the noise measurements The final site was chosen taking into account the measurements described above and arguments other than technical ones: availability, price of the installation, etc. The site D seemed to be the best and only possible compromise between these factors.

3. Measurement Site Equipment 3.1. Magnetometer After the external conditions of the measurement site were determined, its instrumentation could be specified. Since one

101 should be able to study even the cerebral magnetic fields in an eddy-current shield which provides poor attenuation of disturbances at low frequencies, the choice of the magnetometer becomes obvious: the most sensitive SQUID magnetometer available combined with a gradiometer providing the highest degree of external noise rejection. For cerebral magnetic fields, for instance, the inherent noise level of the magnetometer should -1/2

be below 10 fT*Hz

and its balance should be better than

10 ppm. Such a magnetometer will give reasonable results in the shield specified below. 3.2. Possibilities of shielding Fig. 7 shows the noise spectrum of a commercially available second order gradiometer measured in a noisy environment from Brenner et al. (4), and theoretically achievable results placing different shields around this gradiometer. These results demonstrate that shielding is a way to obtain the maximum sensitivity of the gradiometer.

Fig. 7. Noise spectra of a 2nd order gradiometer. a) in a noisy environment from ref. (4), b) inside the eddy-current shield of ref. (5), c) inside the shield described below, d) inside a ferromagnetic shield (inherent noise level of the gradiometer) . frequency ( H z )

Historically, the (6). The magnetic shield, providing inside the shield

best known shield is the ferromagnetic one field lines are conducted through the an attenuation down to dc. The noise level of Cohen (7) reaches a few tens of fT.

102 The second kind of shield is based upon generation of eddy currents in a conducting material that create a field opposing the excitation. For this type of shield, the material employed has to be cheap, highly conductive and light. Aluminium fulfills most of these requirements and is generally employed. These shields are efficient above a cut-off frequency f

and

they can be considered as a first order low-pass filter (8,9) with an attenuation slope of 6 dB/octave above f c . Above the skin thickness, the attenuation becomes even faster. This kind of a shield can be modelized by the equivalent electric circuit of Fig. 8.

Fig. 8. Equivalent electric circuit of an eddycurrent shield. L and R are the resistance and inductance of the shield H q is the applied field and H. is the field inside the shield. Three geometries are of interests sphere, truncated cylinder and cube. The two last cases are somewhat difficult to modelize. As in ref. (5) one can approximate them considering an infinite tube in a longitudinally applied field H q . Table 2 summarizes the interesting parameters for the three shapes considered. At a given frequency far above the 3dB cut-off frequency f c , the attenuation is given by S = f / f

c

- This value is propor-

tional to the thickness, side and conductivity. The third kind of shield is active: it uses a measurement of the magnetic field (from one to three directions) and generates an error current which drives a set of coils (10,11).

103 L sphere cylinder cube

R

T = L/R

f c = 1/27TT

2Trya 9

2ir/3ta

yata/3

3/2iryata

2 y I 2j ya d I

2ira/a It

yata/2

l/'iryata

4a/a£t

yata/4

2/iryata

TTa

Table 2. Equivalent circuit parameters for three types of eddy current shields, a is the radius of the sphere or cylinder or the side of the cube, I the length of the tube, a the conductivity, y the absolute permittivity and t the thickness. In the case of a rectangular section tube, the side a must be substitued by 2£L/£+L where I and L are the rectangle sides. After the detailed study described an eddy-current shield considering to dc was of secondary interest in ferromagnetic shields are too much

above, our final choice was that 1) the spectrum down the first studies, 2) the expensive.

3.3. Model experiments of eddy current shield geometries Before determining the final geometric parameters, a study on small scale models was performed. Two shapes were tested on a 1/5 scale: rectangular and hexagonal, approximating a cylinder, remembering that this form is more efficient theoretically. The openings were also different. Another hexagonal shape was tested on a 1/10 scale (see Fig. 9). The models were placed in a three-axis field facility (12) and the magnetic field was measured with an induction coil magnetometer to determine the cut-off frequency and attenuation as a function of position within the shield. From these measurements a zone of maximal attenuation within 2dB (20 %) was plotted (Fig. 9).

104

Fig. 9. Zones of maximal attenuation for the three models. The doorway is longer from left to right. It was also verified if the theoretical value (from table 2) was reached. Table 3 gathers this data, scaled up to real sizes and compared with the realizations of Zimmerman (5) and Malmivuo et al. (13). shield design thickness mm weight tons shield areas (except2opening) m opening 2 areas m

A

B

C

(5)

(13)

50

50

50

38-44

45

4.1

3.1

3.8

4

3.5

3.5

4

3.5

1.4

1.3

2

0.5

0.5

3

theoretical effectiveness at 50 Hz dB

52

vertical max effectiveness at 50 Hz dB

52

50

52

zone of max. eff. (% of total.area)

54%

30%

76%

35%

price US $ 19 79

31000

28000

34000

30000 (1978)

53 53 (cyl(cylinder) inder)

50

44

45

105 In conclusion, we noticed, as in ref. (13) , that the shape of the opening had a large influence on the field homogeneity. The hexagonal design with a small doorway was the worst one. As a conclusion on the shield study, a cubic geometry with a larger doorway was finally chosen for practical reasons: more usable surface inside, better fitting in the room where it is supposed to be placed, and reduced length of weldings. Finally, the correct approach to choose the dimensions is the following: 1) check the geometry and outside dimensions imposed by the room and installation, 2) deduce the thickness from the desired efficiency and/or the available quantity of money.

4. Installation of the Shield in the Room The final site at the ground floor at the entrance of the hospital permits the shield to be assembled in the factory and then be transported to the site. The shielded room will be installed on a suspended concrete platform (without magnetic armature) to decrease the mechanical resonance frequency of the system (Fig. 10). shielded room rubber damping

\ x \ \

x X X Fig. 10. Shielded room suspension system.

106 5. Conclusion This facility will offer an opportunity to carry out biomagnetic measurements in a hospital, i.e. in a place where normal and pathological patients are available. 6. Acknowledgements The authors wish to thank all the persons who helped to perform the measurements. We are indebted to Prof. Liot, head of Department of Functional Explorations, for his support of our work. This work was partially supported by D.R.E.T.

References (1) Kuusela, M.-L.: communication given at the 2nd Workshop on Biomagnetism, Grenoble, August 1978. (2) Duret, D., Zenatti, D., and Bernard, P.: Rev. Sci. Instr. 46, 474 (1975) . (3) Chiron, G., Delapierre, G.: IEEE Trans. Magn., MAG-15, 1815 (1979). (4) Brenner, D., Kaufman, L., and Williamson, S.J.: IEEE Trans. Magn., MAG-13, 365 (1977). (5) Zimmerman, J.E.: J. Appl. Phys., £8, 702 (1977). (6) King, L.V.: "Electromagnetic shielding at radio frequencies", London, Edinburgh, and Dublin, Phil. Mag. and J. Science, ser. 7, 15_, (1933). (7) Cohen, D.: IEEE Trans. Magn., MAG-11, 694 (1975). (8) Miller, D.A. and Bridges, J.E.: IEEE Trans. El. Comp., EMC-10, 52 (1968) . (9) Cooley, W.W.: IEEE Trans. El. Comp., EMC-10, 34 (1968). (10) Marzetta, L.A.: Rev. Sci. Instr., 22, 1192 (1961). (11) Freedman, M.S., Wagner, Jr., F., Porter, F.T., and Day, P.: J. Appl. Phys., 38, 1856 (1967). (12) Duret, D., Chiron, G., and Karp, P.: to be published (13) Malmivuo, J., Heinonen, P., Tuomola, M., aad Lekkala, J.: these proceedings.

THICK-WALLED CONDUCTING SHIELD IN BIOMAGNETIC EXPERIMENTS

J. Malmivuo, P. Heinonen, M. Tuomola and J. Lekkala Biomedical Engineering Laboratory, Department of Electrical Engineering, Tampere University of Technology SF-33101 Tampere 10, Finland

A traditional method to build a magnetic shield is to use highly permeable y-metal sheets in combination with aluminium sheets, and to construct a multi-layer wall /l/. This kind of shield gives an excellent attenuation with constant and lowfrequency magnetic field. The main disadvantages in this kind of construction are the high price of the y-metal and the complexity of the construction. Additionally the permeability of the y-metal is sensitive to mechanical stress. Another method is to use highly conducting material in the construction. The shielding effect in this kind of shield is based on the induced eddy currents. Zimmerman /2/ constructed this kind of shield from surplus aluminium plates with 5 cm wall thickness. At Tampere University of Technology we decided to build a magnetically shielded room according to the principle of conducting wall. We selected this alternative because of the limited size of the laboratory and because this construction was much more economic. This paper describes the attenuation properties of this shield in 50 Hz magnetic field. Our shield is constructed from 4 5 mm thick aluminium plate by welding the seams•completely through the plate. The material is 99.5 % pure aluminium with conductivity 36-10 S/m. The outer dimensions of the shield are 2x2x2 m. To decrease the loss of attenuation due to the door a long corridor is con-

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

108 structed inside the shield, Fig. 1. The total weight of the shield is 3000 kg. Theoretically the attenuation of the shield without the door should be 50 dB at 50 Hz. Because the shield gives no attenuation for the earth's static field the mechanical movement of the detector may induce considerable noise. Therefore it is important to attenuate the mechanical vibration of the shield. Therefore the shield is mounted on four pneumatic dampers. To attenuate the vibrations caused by the personnel inside the shield, the shield is equipped with a plywood floor which stands on the laboratory floor on brass rods through holes in the bottom of the shield.

109 The magnetic field measurements were performed with a system composing of a coil, an amplifier, and a spectrum analyzer. The coil diameter was 14.8 cm, it had 2000 turns of 0.5 mm 2

copper wire and its effective area was 23.6 m .

It was cov-

ered with a thin aluminium foil to shield it against electric field. Three orthogonal measurements of the magnetic field were performed outside the shield in 70 points at distances from 20 to 150 cm from the wall. Inside the shield the measurements were performed at 100 points, all at 1 m height. The distance between these points was 20 cm. The attenuation of the magnetic field at 50 Hz inside the shield is presented in Figs. 2, 3 and 4. The field strengths are given in rms values and the dimensions are in nanoTeslas outside the shield and in picoTeslas inside the shield. The field strength in the laboratory varies from 5 nT to 15 nT. The field values inside the shield were corrected according to a reference measurement outside the shield during the measurement procedure. The shield seems to achieve the full theoretical attenuation of 50 dB in only one component of the magnetic field, the Zdirection. The maximum attenuation in X- and Y-components are 44 dB and 46 dB, respectively. This indicates that the doorway is an important question in designing magnetically shielded enclosures.

References 1. D. Cohen: "Large-volume conventional magnetic shielding", Revue de Physique Appliqueé, Février 1970, p. 53 2. J. Zimmerman: "SQUID instruments and shielding for lowlevel magnetic measurements", J. Appl. Phys., £8, 1977, p. 702

Fig. 2.

The rms value of the X-component of the 50 Hz magnetic field. The dimensions are in nT outside the shield and in pT inside the shield.

111

Fig. 3.

The rms value of the Y-component of the 50 Hz magnetic field. The dimensions are in nT outside the shield and in pT inside the shield.

112

Fig. 4.

The rms value of the Z-component of the 50 Hz magnetic field. The dimensions are in nT outside the shield and in pT inside the shield.

AN ALUMINIUM SHIELDED ROOM FOR BIOMAGNETIC MEASUREMENTS

G. Stroink, B. Brown, B. Blackford, M. Horacek Department of Physics, Dalhousie University Halifax, Nova Scotia, Canada B3H 3J5

Introduction In order to observe details in the magnetic signals of the heart, typically 10 pT peak to peak in a bandwidth of .05-100 Hz, noise levels of about 0.1 pT//Hz are required.

In an

attempt to obtain such low noise levels in the lab environment, we have constructed a 2nd order pick-up coil (gradiometer) attached to a commercial SQUID sensor (1).

This,

however, was not, by itself, sufficient to reduce the 60 Hz noise detected by the gradiometer to acceptable levels.

We

found that the fields present consisted, for a large part, of field gradients with large amplitude variations in time.

To

reduce the 60 Hz fields, a large eddy current shielded room of high conductivity aluminium was constructed. In the first part of this summary we will give a short description of the probe with the gradiometer. In the second part, we present details of the room as well as its performance as measured by the gradiometer.

The Gradiometer The central part of the gradiometer is a cylinder with a diameter of 2.2 cm and 10.1 cm long, made of a fine grade of tufnol. The cylinder is attached to a rod, also of tufnol, which screws into the SQUID housing. A niobium wire, of

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

114

diameter 0.13 mm, is wound around the cylinder.

A total of 8

windings are used to create two first order gradiometers back to back.

The baseline is 3.8 cm; this is similar to that used

in the second order gradiometers described before (2,3). Three rods, containing small pieces of superconducting material, are used to adjust the balance of the coils in the x, y and z directions.

The balancing of the probe takes place

in the field of 3 square coils, mounted on the inside of the shielded room.

Using the field generated by the coils, we

balanced our probe to within 2 x 10 s in the z-directionf the direction of the axis of the probe. fibreglass cryostat (1).

The probe is located in a

The distance between the lower coil

and the outside of the cryostat is 1.5 cm.

The Shielded Room The welded aluminium room (i, = 3.66 m, w = 2.44 m, h = 2.44 m, t = 1.88 cm) was made from high purity aluminium having a conductivity a = 0.36 x 108 (fl-m)-1. Before welding the sheets together, the conductivity of several types of welds was tested. It was found that by bevelling the sides of the sheets and filling the v-grooves created between adjoining sheets with several welding layers, the conductivity across the weld was essentially the same as the of the sheets. In one of the long sides of the room, an opening, .83 m wide and 2.44 m high, gives access to the room. This opening is partitioned off by an sheet 2.44 m high and 1.20 m long, creating a small corridor outside the main room. The floor space of the main room is 6.6 m 2 - large enough to contain the wooden bed and wooden dewar support. The design is similar to that of the aluminium enclosure of Zimmerman (4), but is larger and does not contain an additional soft iron enclosure.

115 Performance The response of an eddy current shielded room to a uniform applied field can be approximated by (4): H ext(u) H(u) = -S5E (1+iwT)

where T =

(1)

„ y0CTwht 2(w+h)

Using the values given above, we find x = .52 s. From Eq. 1 it follows that the room should provide an attenuation of 4 6 dB at 60 Hz. x was measured by applying a step function to the large coils inside the room. From the response of the gradiometer a t = .49 s was measured. Assuming Eq. 1 is valid, we then obtain an attentuation of 45 dB at 60 Hz. This value agrees reasonably well with measurements of the 60 Hz noise level in and outside the shielded room, and with the measurements taken before the room was installed. We conclude that Eq. 1 can be used with confidence to predict the response of an eddy current shielded room to uniform fields. In order to measure the response of the room to non-uniform fields we measured the response of the room to a step function applied to a small magnetic coil located outside the room. T increased from .2 s just outside the room to .49 s, 6 m away from the centre of the room. The frequency response was measured as well; it is given approximately by Eq. 1. The data suggests that in order to make full use of the shielding properties of the room, instruments should be located at least 5 m away from the centre of the room. One of our concerns was that mechanical resonances in the shielded room could generate unwanted signals during MCG measurements. In one series of experiments the room was deliberately set in resonance by hitting it. The Fast Fourier transform of the noise was recorded. Resonances of about

116 equal amplitude occurred at 15 and 43 Hz. Next MCG's were recorded under normal conditions and Fourier analysed as well. No additional signals could be detected at the resonance frequencies. The data indicate that the contribution of the room to the MCG's at the resonant frequencies is less than .05 pT. The measured component of the Fourier spectrum of the MCG signals near 43 Hz. The Fourier spectrum of the room between 5 x 10"3 and 500 Hz was recorded. On a log-log plot the r.m.s. noise drops linearly from 10 pT//Hz to .1 pT/v/Hz" in the frequency range of 5 x 10"3 to 1 Hz. From that point, the noise remains constant, with the exception of peaks at 60 and 180 Hz, where the amplitude of the noise signals is typically .6 and .2 pT respectively. In addition to the noise described above, there were also DC drifts against which the room does not shield. Occasionally these can be as large as 10 pT in a 2 min. interval. However, it should be pointed out that after the MCG's are recorded, the data is further processed by a computer programme which digitizes and then signal averages the data; it also corrects for base line drifts.

References 1.

From S.H.E. Corporation, San Diego, California, U.S.A.

2.

Opfer, J., Yeo, Y. , Pierce, J., Rorden, L., I.E.E.E. Trans. Mag. MAG-10, 536 (1974). Brenner, D., Kaufman, L., Williamson, S., I.E.E.E. Trans. Mag. MAG-13, 365 (1977). Zimmerman, J., J. Appl. Phys. 48, 702 (1977).

3. 4.

A SUPERCONDUCTING HELMET FOR

MAGNETOENCEPHALOGRAPHY

WITH A SQUID

H. E. Hoenig and C.

Gassinger

P h y s i k a l i s c h e s I n s t i t u t der U n i v e r s i t ä t F r a n k f u r t R o b e r t M a y e r S t r . 2, D 6G00 F r a n k f u r t , West Germany

Introduction Up to now m a g n e t o e n r e p h a l o g r a p h y on g r a d i o m e t e r s

(WEG) using S Q U I D s is b a s e d

of 1. and 2. o r d e r for signal pick up

W i t h e x p e r i m e n t s on a 1:5 s c a l e model for a real MEG

(1). cryostat

we haue d e m o n s t r a t e d t h a t a h e l m e t like s u p e r c o n d u c t i n g to be p l a c e d over the head w i t h one coaxial up coil is s u p e r i o r to the g r a d i o m e t e r s ciple h i g h e r s e n s i t i v i t y

practice

closeby.

W h e r e a s the e f f i c i e n t s h i e l d i n g of c o a x i a l f i e l d s is

evident

this is not the case for t r a n s v e r s e f i e l d s . To see this has to r e a l i z e that a signal coil p o s i t i o n e d p r o p e r l y s h i e l d from o u t s i d e . We have shown t h a t this can be

one

relative

to the s h i e l d does not pick up the f i e l d reaching into in the

pick

b e c a u s e of in p r i n -

to " i n s i d e " fields and in

b e t t e r a t t e n u a t i o n of " o u t s i d e " f i e l d s

shield

(adjustable)

the

achieved

experiment.

Exp e r i m e n t a l The t e s t a p p a r a t u s mainly

c o n s i s t s of a l e a d - t i n c o a t e d

c y l i n d e r of 50 mm d i a m . and 75 mm h i g h t , open at the w i t h an a d j u s t a b l e pick up coil i n s i d e coil is p a r t of a s u p e r c o n d u c t i n g

brass

bottom,

(Fig. 1). The pick up

flux t r a n s f o r m e r and

inter-

c o n n e c t e d w i t h a two hole SQUID w h i c h is p l a c e d w i t h i n

its

own s u p e r c o n d u c t i n g

holder

s h i e l d on top of the c y l i n d e r . The

of the coil is a d j u s t e d r e l a t i v e to the s h i e l d w i t h i n

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

one

118

Fig. 1 S k e t c h of t h e t e s t a p p a r a t u s . A c y l i n d r i c a l s u p e r c o n d u c t i n g s h i e l d m a d e of l e a d - t i n c o a t e d b r a s s i n c l o s e s an a d j u s t a b l e p i c k up c o i l of 9 mm OD 10 to 15 mm a p a r t f r o m t h e t o p , i n t e r connected with the SQUID, and a test c o i l . A d j u s t m e n t is done by t h r e e Nb s c r e w s . T h e p l a s t i c h o l d e r of t h e p i c k up c o i l is h e l d in p l a c e by a C u - B e s p r i n g . T h e s h i e l d is p l a c e d in a regular glass dewar. minute

of a r c f r o m t h e o u t s i d e

screws

( b e t t e r by p i e z o e l e c t r i c

the response nally

to t r a n s v e r s e

means)

homogeneous

glass

around the

w i t h an a m p l i t u d e

a x i s of an e x t e r n a l of a b o u t

a few s t e p s t h u s m i n i m i z i n g in a m p l i f i e r .

and second

fields

the SQUID

minimize

applied

homogeneous response

exter-

consists ac

in

field

the screws

in

as m o n i t o r e d

by

f i e l d s or h o m o g e n e o u s

a r e g e n e r a t e d by c y l i n d r i c a l

65 cm d i a m e t e r s e t up in H e l m h o l t z - or g r a d i e n t respectively.

by t h r e e 1Mb

procedure

Q»5 ^uT and adjusting

Homogeneous

derivatives

dewar

in o r d e r to

(the m a i n p r o b l e m ) . T h e a d j u s t m e n t

turning

a lock

of t h e

first

coils

of

arrangements

119 Results

and

By p l a c i n g

Discussion

a coil

of the

SQUID

to

% which

3.6

coil of

were

the

In t a b l e ation and

the

(with 1 we

field

area

coil

determined

improved. ac

of

the

on a n d

off

field

the

transverse

response

field

20

Hz by

signal

deriva-

symmetry

was m o n i t o r e d

or at

transfer

adjusted

and

to

flux

experimentally

With

fields

or t r a n s v e r s e

SQUID

the

axis

was directly

a lock

in

difference).

have

collected from

of the

derivatives

was be

amplitude

no

signal

dc and

coaxial

D.5 ^ u j . The

(calculated

and the

can

homogeneous

switching

amplifier

the

transformer

applied

to

onto

certainly

shield. The

limited by

flux

external

tives

just

the

pick

up

B'(B'')

coefficients

SQUID

signal,

coil)

applied

as

for the

function

externally.

the

field

flux of The

attenu-

transfer the

fields

B

coefficients

Table 1 E x p e r i m e n t a l a t t e n u a t i o n c o e f f i c i e n t s OC a n d |5 o f t h e a d j u s t e d test apparatus for homogeneous fields or field d e r i v a t i v e s . T h e f i e l d d e t e c t e d by the p i c k u p coil f o r c o a x i a l e x t e r n a l f i e l d is d e f i n e d by I I

s

I

=

Fig.

6.

MEG: spontaneous activity. Bw: 4+16 Hz. The first 6 traces correspond to 6 different spots over the scalp 1 cm apart from each other from the temporal bone to the inion. The last trace is the noise level without the subject in the same experimental conditions

A Fourier transform of the sition c) is shown in Fig. 2 v 34 Hz and an average on the signal to noise ratio. clearly visible.

magnetic signal obtained in the po7. In this case the bandwidth is 50 samples has been used to improve A 10.5 Hz component (the a-rythm)is

Fig. 8a shows auditory evoked response recorded over the auditory cortex 2 cm above the ear. The transient stimulus was a

146

grms

V Fig.

7.

(Hz)

Fouri-er trans form of the magnetic signal due to spontaneous brain a c t i v i t y . Bw:2+34Sz.The Fourier transform was obtained averaging over 50 samples 8 seconds long. The lower trace is the noise spectrum without the subject in the same experimental conditions

average was made over 100 stimuli. Bw:l±40Tiz. The lower trace is the noise level averaged in the same experimental conditions b) Visually evoked brain response averaged over 100 flashes. The lower trace is the noise level with no stimuli. Bw:l+40Hz

147 short tone of ~lKHz repeated at an average frequency of ~lHz. The bandwidth was 1 i 40Hz. The resulting trace was obtained averaging over 100 stimuli. Fig. 8b shows visually evoked neuromagnetic fields recorded over the visual cortex. The transient stimulus was a strobe lamp flashing randomly at an average frequency of -1Hz. The trace shown was obtained averaging over 100 flashes using a bandwidth of 1 T 4 0 H Z . At last Fig.9 shows two magnetomyograms recorded during muscle contraction. The bandwidth was 0 . 1 i - 2 6 0 H z .

Fig. 9.

"o

MMG: elbow and palm contractions.

100

200

Bw: 0.1 *260

360

Hz

400

"fHiJ Fig. 10. Fourier trans form of the magnetic signal due to the elbow muscle contraction. No average was made in this case. Bw: l±400Hz. The lower trace is the noise spectrum reported in the same scales and obtained keeping the muscle relaxed.

148 In Fig. 10 we show a Fourier transform of the signal obtained during a single muscle contraction. The spectrum is explored with a bandwidth of 400 Hz. CONCLUSIONS From an experimental point of view, the results reported show that, provided an accurate construction and fine balance of the detector are made, it is possible to detect biomagnetic signals also in a normal working environment. This provides good feasibility for routine use in hospitals or research laboratories where the use of magnetically shielded rooms would be impractical. From a clinical point of view the detection of biomagnetic signals seems to be most desirable because of its absolute non-invasHzity. This is particularly true for magnetomyographic signals where the S/N ratio achievable in unshielded environments is in fact satisfactory enough to allow any requirable analyses both in the time and frequency domains. As long as MCG maps are concerned, our preliminary results are consistent with the detailed study performed by other authors (12, 13) even if the detection grid used by us is not, the stan dard one (12). Moreover we believe that, due to its high spatial resolution the MCG method is quite suitable for the investigation of ele£ trophysiological phenomena like, for instance, ventricular repolarization, which are not fully clarified with surface ECG. Besides high resolution MCG may be employed in non-invasive investigation of the conduction of the cardiac impulse. Preliminary results concerning these topics are presented in another communication to this workshop. (14)

ACKNOWLEDGEMENTS Thanks are due to Prof. J. Zimmermann for helpful discussions and suggestions. The authors thank also Mr. M. Maggi and Mr. S. D'Angelo for their skilled technical assistance.

149 REFERENCES 1. 2. 3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14.

G.L. Romani, International Journal of Refrigeration, 2 215 (1 979) D. Cohen, Rev. de Phys. Appliquée, 5^, 53 (1970) J.E. Zimmermann, J. Appi. Phys., 48_, 702 (1977) J.E. Zimmermann, N.V. Frederick, Appi. Phys. Letters, 1_9 16 (1971) S. Barbanera, P. Carelli, I. Modena and G.L. Romani, J. Phys. E: Sei. Instrum., V\_, 297 (1978) J.E. Opfer, Y.K. Yeo, J.M. Pierce, L.H. Rorden, IEEE Trans. Magn., 9, 536 (1974) S.J. Williamson, L. Kaufman, D. Brenner, in Superconductor Applications: SQUIDS and Machines, B.B. Schwartz and S. Foner eds., (Plenum Pubi. Corp., New York 1977) pp. 355-402 J.A. Overweg, M.J. Walter Peters, Cryogenics, 1_8, 529 (1978) S.M. Rubens, Rev. Sei. Inst., 1_6, 243 (1 945) Manufactured by S.H.E. Corporations, San Diego, California U.S.A. The scheme of the comb filter was kindly supported by the biomagnetic group working at New York University, directed by Prof. S.J. Williamson M. Saarinen, P. Siltanen, P.J. Karp, T.E. Katila, Annals of Clinical Research, Vol. 10, Suppl. 21 (1978) P.J. Karp, T.E. Katila, M. Saarinen, P. Siltanen, T.T. Varpula, Ann. Cardiol. Angeiol. 27^, 65 (1 978) S. Barbanera, P. Carelli, R. Leoni, G.L. Romani, F. Bordo ni, I. Modena, R. Fenici and P. Zeppilli, Magnetocardiographic study of some human cardiac electrophysiological phenomena: preliminary observation. Contribution to this workshop.

FUNDAMENTALS

MEDICAL SIGNIFICANCE OF THE MAGNETIC ACTIVITIES OF THE HUMAN BODY

Allan P. Freedman, M.D. Division of Pulmonary Diseases Hahnemann Medical College and Hospital 230 North Broad Street Philadelphia, PA, USA

Introduction Magnetic phenomena have long fascinated man. The association of magnetism with biologic processes is not a recent concept. Lodestone, either applied externally or taken internally, has been used therapeutically by medical practitioners for nearly two thousand years (1). It was a physician, William Gilbert, who in 1600 wrote the first scientific treatise discussing magnetism and electricity as separate entities (1). Dr. Franz Mesmer, a physician living in the eighteenth century, and his followers believed that manipulation of the magnetic force in man formed the basis of his therapeutic success with what we now know to be hypnosis (2). The coupling of magnetism and biology has been mostly associated with cults and pseudoscientific quackery until the relatively recent recording of biomagnetic phenomena. The measurement of biomagnetism, magnetic fields associated with biologic activity, began in earnest in the early 1960's with Baule and McFee's recording of the magnetocardiogram (3). Zimmerman's construction of the first SQUID magnetometer immensely facilitated biomagnetic measurements (4). Magnetic fields associated with ion currents, including the magnetocardiogram, the magnetoencephalogram, the magnetomyogram, and others, have now been recorded. Other aspects of magnetism have been explored. The detection of ferrimagnetic inhaled

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

154 particles within the thorax has made it possible to assess occupationally acquired lung dust and to study the kinetics of dust clearance in humans without using radiotracers. Also, magnetic susceptability measurements allow cardiac output and body iron stores to be non-invasively monitored. The number of investigators exploring biomagnetic phenomena has increased markedly since the First Workshop on Biomagnetics held in 1976. Despite the exponential increase in data and continuing improvements in technology, our understanding of biomagnetic phenomena is in its infancy. It is hoped that physiologists and physicians will actively collaborate with physicists and biomedical engineers to achieve new insights into basic physiologic mechanisms. The clinical diagnostic applications of biomagnetics are within reach. In this paper, I will attempt to review the scope of biomagnetic measurements from a physicians point of view.

Magnetocardiogram The magnetocardiogram (MCG) was the first recorded biomagnetic phenomenon. It is a relatively large signal (up to 50 picoTesla) that can be recorded in real time either with a single coil in a magnetically shielded room or with second derivative coil arrangements in an unshielded environment (5,6). A sample tracing recorded by us with a second derivative SQUID gradiometer in a magnetically noisy urban hospital laboratory is shown in figure 1. Sodium/potassium flux across cellular membranes induces a change in cellular transmembrane electric potential that propagates first along conducting fibers, and then through the striated muscle of the myocardium. The P wave of atrial depolarization, the QRS complex of ventricular depolarization, and the T wave of subsequent ventricular repolarization are recorded via skin electrodes in the electro-

155

MCG i

0 -100 Hz

10 pT

I < BRUSH ACCUCHAflT

200 mS Figure 1:

The top tracing is an MCG recorded on line with a second derivative gradiometer in an urban hospital laboratory. No magnetic shielding was used. Below is a representative ECG tracing for comparison.

cardiogram (ECG). Differences in electrical potential between skin electrodes reflect both the primary cardiac current and the secondary volume currents induced in surrounding low resistance tissues (7). The magnetic correlate of this, the magnetocardiogram (MCG), sees these events with a different weighting. It weights the signals according to distance of currents from the sensing coil (8). As a result, cardiac current generators are seen with a greater degree of spatial resolution, since secondary volume currents contribute little (8). An additional advantage of the MCG is its ability to easily detect dc changes in electric potential, information that in the ECG can be obscured by lead contacts with the skin (9). The following types of information are present in the ECG; cardiac rhythm, conduction defects, atrial enlargement or

156 hypertrophy, ventricular hypertrophy, myocardial ischemia, and myocardial infarction. Vectorcardiography may be a better indicator of the changing direction of current flux during the cardiac cycle (10). Does the MCG offer new or additional information? The normal MCG has been reported in detail and some features differ from the ECG (6,11-13). The polarity of the T wave with respect to the R wave, the relative sensitivity to P waves and to the U wave that may follow the T wave all differ from the ECG (11,14). Some patients with cardiac disease have also been studied. Patterns of left and right ventricular hypertrophy as well as patterns of bundle branch blocks in conduction have been recorded (8,11,14). Signal averaging allows the usually obscured components of the PR interval (during which current propagates in the atrioventricular node and conducting bundle of HIS) to be discerned (15), though this can be done with the vectorcardiogram (16). Finally, dc shifts of the ST segment, similar to those seen in the ECG, have been noted in patients with myocardial infarction (17). Only in recording dc shifts do the advantage of the MCG seem obvious. The source of the ST segment elevation or current of injury associated with acute myocardial infarction, ventricular aneurysmal dilatations, and pericarditis has been variously attributed to incomplete polarization or to incomplete depolarization in the injured myocardium (18). The latter would shift the ST segment (during which the entire myocardium is depolarized) above baseline, and the former would shift all wave forms other than the ST segment below baseline. Either process would appear the same in the ECG. Cohen's model of animal ischemia lends support to the theory of incomplete polarization. As the SQUID probe nears the heart, other waveforms shift downward while the ST segment remains isomagnetic to the baseline (shown in figure 2) (19).

157 Robinson has postulated that lactate ion generated in the periinfarcted ischemic area during anaerobic metabolism may have a role in this (20). The MCG will allow this to be explored further. Furthermore, it may be possible to use the greater spatial resolution of the MCG to better delineate the extent of infarcted tissue. MCG patterns must be correlated with standard techniques for studying physiologic and anatomic alterations such as the ECG and vectorcardiogram, ultrasound imaging 4- radiotracer studies of the myocardium, etc.(21,22). The spatial resolution of the MCG for infarct mapping must be compared to anterior chest lead grid mapping by the ECG and gamma camera studies of radiotechnetium uptake or radiothallium exclusion by the myocardium. Evaluation of ventricular wall hypertrophy by the MCG should use M mode ultrasound as a reference. M mode ultra~ sound is much better in evaluating the left than the right ventricle (22), and the MCG may be helpful here. Only with studies incorporating all available diagnostic modalities can the additional or complementary information supplied by the MCG be appreciated.

ORS Figure 2:

This is a pattern of ST segment elevation (arrows) or current of injury seen with myocardial infarction. Cohen found the true isomagnetic(and by inference isoelectric) baseline to correspond to the ST segment (19).

158 A standard MCG lead system should be adopted for ease of data comparison. Until the relative advantages of multiple grid measurements and unipositional vectorial recording have been clarified, it would be prudent to study pathologic events with both systems (14). In addition, the distance of the sensing coil from the myocardium should be mentioned. If ultrasound is not available or is not helpful, then distance to the overlying skin should be noted. Though the freedom from electrodes and electrode paste has caused some investigators to envision rapid MCG screening measurements, I am not sure what clinically useful information would be found at the present time. The ability of the MCG to preferentially see tangential vs. radial currents with a sensing coil parallel to the body surface and its decreased sensitivity to secondary volume currents are unique. The rate of signal fall off with distance and the ability to vary the rate of fall off with different coil configurations hint at the potential of the MCG for determining the depth of' current generators. The availability of superconducting instruments capable of real time measurements in the hospital environment will allow the usefulness of MCG recording to be rapidly evaluated.

Hagnetoencephalography The changes in electric potential between the scalp electrodes used to record the electroencephalogram (EEG) represent the summation of excitatory and inhibitory potentials. These are usually generated in the 3mm deep multilayered gray matter of the cerebral cortex (23). Only for epilepsy are abnormalities of spontaneous activity diagnostically specific (24). Epilepsy represents synaptically synchronized activity trig-

159 gered by an area of spike generation, and is clinically manifest as a seizure.

However, destructive lesions in

cerebral white matter, such as tumors or infarcts, often manifest continuous polymorphic slow wave or delta activity with focal depression of usual backgroud activity (25 ,26) .

The

EEG can also be used to record potentials evoked by various sensory stimuli.

These evoked potentials can be used to

explore afferent sensory neuronal circuits to cortical areas, and define normal physiology or pathologic processes(27) The magnetic correlate of the EEG, the magnetoencephalogram (MEG), was first described by Cohen in 1968, using signal averaging

(28) .

With the use of more refined SQUID magneto-

meters and magnetic shielding, real time recordings are now made of the MEG (29,31).

EEG recording via scalp electrodes

is subject to artifacts from muscle contraction, ECG potentials at the scalp, and changes in skin resistance from sweats ing

(32).

In addition, the insulating effect of the skull

makes for poor spatial resolution when recording from overlying scalp electrodes.

In contrast, the low

conductivity of the

skull makes it relatively transparent to magnetic fields. The signals of the MEG are up to several picoTesla in amplitude and have a slightly different power spectra than the EEG. Though alpha waves

(8-13Hz) are recorded

with equal facility

and demonstrate the same suppression with eye closure, slow delta

(0.5-3Hz) and theta

relative sensitivities well.

(6^7Hz) waves are seen with different

(28,29,31).

Sleep spindles are not seen

Though it has not been studied, the radially oriented

electrical vector of the evanescent spike and sharp waves from underlying seizure foci would be poorly detected magnetically.

Where might the MEG provide new information with regard to spontaneous events?

160 1. Seizure activity; During motor seizure, motion artifact makes it very difficult to study propagation of the seizure activity through the cortex with the EEG. During the interictal period, it might be possible to record non-radial components of the spike and sharp wave bursts (associated with focal irritability of motor, sensory, behavioral, or other cortical areas) with the higher spatial resolution offered by the MEG* if a pickup coil perpendicular to the scalp were used. 2. Areas inaccessible to EEG recording; Temporal lobe recording from scalp electrodes is difficult, and the other approaches using nasopharyngeal electrodes or sphenoid sinus electrodes can be uncomfortable, Similarly, the study of cerebellar disorders of gait and brainstem disorders of respiratory control, even with the use of signal averaging, is not feasible with the EEG but might be possible with the MEG, 3. Localization of destructive subcortical lesions; The slow waves and backgroud depression recorded by the EEG at best only allow localization to a single quadrant, and they are not always maximal over the anatomic lesions (25,26) , MEG might allow more precise localization. 4. DC magnetic shifts. These have been recorded over the brain by Cohen and are of unknown significance(33). However, he has recently presented evidence that they may actually arise from skin structures(9). These should be explored further in human pathologic conditions and experimentally induced lesions in animals. It is possible that currents of injury similar to those seen in the heart may exist in cerebral structures. It is important that comparable "lead" systems and filtering schema be used by different investigators so data can be compared.

Finally, just as with the MCG, other modalities

161 must be used for correlative study. cordings of functional changes.

MEG phenomena are re-

Comparison must be made not

only with the EEG, but also with careful neurologic examination and computerized transaxial tomography (an anatomic study)(34). The MEG may also offer new insights into cerebral function through study of evoked potentials. Signal averaging is necessary to record these

Dusts containing silicon dioxide are especially fibrogenic and may be encountered in various types of mining and quarrying where there is magnetite or haematite (y Fe 2 0 3 ) in the respired dust.

Asbestos is both fibrogenic and carcinogenic.

The use of magnetic alignment in counting of asbestos fibers attests to their ferrimagnetic properties (65), and magnetic minerals may be found in association with the fibers.

Coal

workers' pneumoconiosis is the basis for many disability claims in the eastern United States.

We have found that coal

mine dust has several tenths of a percent FM by weight, a concentration sufficient to allow detection of high lung dust levels (61).

Hemosiderosis of welders and iron workers is

more a radiologic than a physiologic problem (66) , but the large percentage of FM in the causative dust (25-70%) allows easy monitoring of lung dust levels (67). Of what use are such measurements? diagnosis of disease.

Certainly not for the

MPG does not detect lung pathology.

Measurements cannot even be used to prove a workers' exposure to a specific dust, as FM is present in many dusts.

MPG does

allow serial monitoring of workers exposed to any toxic or potentially toxic dust containing even trace amounts of FM. Though the radiologic abnormalities of pneumoconiosis are associated with higher remanent fields, thoracic FM can be measured in worker^ with normal chest x-rays (57,61,62).

A

rapid rate of dust accumulation may result from both slow clearance of inhaled dust and a high level of dust exposure (68,69)

Prospective serial monitoring of workers could detect

those workers beginning to accumulate large quantities of dust

166 from either impaired clearance mechanisms or very dusty working conditions. This-aot only protects"the worker, but is cost effective for government and industry. Measurement at only a single point in time cannot yield quantitative data on lung dust content. The fraction of FM in inhaled dust can vary and must be monitored along with serial MPG measurements. Also, quantitative data cannot be acquired with current techniques of uniform-field MPG (magnetizing the entire thorax at once), but may be possible with the new technique of localized-field MPG (70). Until such quantitation is available, the measurements obtained can be used for crude ranking of workers' thoracic FM concentrations. It has been suggested that the magnetization-relaxation curve may be effected by the biologic compartment (intracellular, extracellular free or extracellular bound in collagen, etc.) in which particles reside (55) . No controlled studies with pathologic correlation have been reported as yet. Finally, inhalation of trace quantities of magnetite allows clearance mechanisms and clearance kinetics to be studied. Cohen has studied the effect of smoking on long-term or alveolar clearance in humans (71). Halpern at NYU has studied clearance kinetics in the donkey (72), and we have done similar measurements in the guinea pig with localized-field MPG (62). The use of localized-field MPG with partial erasure, to ascertain particle depth, may allow dust clearance to be separated from particle translocation or redistribution within lung compartments (73). This cannot be done with radiotracer techniques. The effect of environmental pollutants, smoking, drugs, and pathologic conditions on particle clearance is now open for non-invasive study. Additionally, extrapulmonary destinations of cleared dust particles can now be readily investigated. Radiotracer limitations in dose have made it difficult until now to address these questions in man. Ethical considerations apply here. There seem to be no risks

167 from short term application of dc magnetic fields of the magnitude used for MPG (74,75). obtain a written consent.

It would still be prudent to

It is with magnetite inhalation for

the study of clearance that I see potential problems.

Purity

and sterility of the preparation used must be established.

It

must be free of trace metals, pyrogens, or toxic organic compounds.

The dose of even pure inert magnetite must be kept

at a minimum by use of state of the art detectors, since nonspecific irritant effects are a theoretical possibility. Though magnetite pneumoconiosis, a radiologic rather than a clinical entity, has been reported, it only occurs with massive exposures (76).

Magnetic Susceptometry The electron configuration of compounds confers upon them certain magnetic properties that distort magnetic fields applied to them.

Faraday noted over a century ago that most

substances, including tissue and blood, were diamagnetic

(77).

They attenuate the magnetic lines of force passing through them. high

Oxygen and organic iron compounds, found in relatively concentrations within the body, are paramagnetic.

They

augment magnetic fields passing through them, in contrast to most biologic substances. There are normally 3-4 grams of iron in the body, with one quarter stored in the liver and reticuloendothelial system as soluble ferritin and insoluble hemosiderin

(78).

Areas of

tissue injury may preferentially accumulate these iron protein complexes due to increased phagocytic cell activity.

However,

diffuse reticuloendothelial accumulation of these compounds is quite rare, occurring only in two circumstances.

The first is

an extremely rare congenital avidity for gastrointestinal iron uptake, termed idiopathic hemochromatosis

(78).

Hemosiderosis

refers to an increase in total body iron reflected in increased

168 depositions of hemosiderin, while hemochromatosis refers to functional impairment of body organs by massive generalized hemosiderin accumulation. This leads to cardiac failure, hepatic cirrhosis, diabetes mellitus, and a distinctive bluish bronze discoloration of the skin. The second condition leading to hemochromatosis is the massive iron accumulation caused by the repeated blood transfusions required by patients with homozygous thalassemia. These patients have impaired synthesis of the 3 globulin of hemoglobin, resulting in ineffective erythropoiesis and severe anemia (79). One patient, with a history of 404 blood transfusions, is reported to have triggered an airport security device (80). Of all the effected organs, the liver is the largest and most accessable. Bauman and Harris, first described in 1967 animal experiments showing a linear relationship of liver iron stores and m a g n e t i c

susceptability

(81).

A b n o r m a l i t i e s in f o u r

patients,

with f a m i l i a l hemochromatosis haye s u b s e q u e n t l y been reported

(82). Again we ask what advantages such measurements have over current tests? Current therapy of idiopathic hemochromatosis is with frequent phlebotomy. The secondary hemochromatosis of thalassemia is treated with a chelating agent, desferoxamine. Therapy is monitored by serum ferritin levels, urinary content of chelated iron after injection of desferoxamine, and by liver biopsy. Only the serum ferritin level can compete with susceptometry for being relatively noninvasive. The ability of magnetic susceptometry to detect small increases in hepatic iron stores suggests great potential for a clinical role. The possibility of using it to assess the distribution of the excess iron within the body should be explored in an animal model. However, a careful study comparing susceptometry to the other indices of iron storage and to indices of target organ function is first needed. Information on distribution is not available with any other test.

169 Another clinical application of magnetic susceptometry is in measuring cardiac output.

The lungs surrounding the heart

are much less diamagnetic than the heart due to lower water content and the presence of a higher concentration of paramagnetic oxygen (83)•

The change in cardiac size during each

heart beat therefore causes a change in magnetic susceptability that reflects cardiac output (84).

Though this is a clinically

useful application, two dimensional ultrasound already gives this information non-invasively

(21,22).

One potential application that would be of great importance is the monitoring of increased pulmonary interstitial water in critically ill patients with shock lung.

This presents a

real challenge, as magnetic susceptability of the lung will change with cardiac cycle (pulmonary vascular blood volume increases during systole) and respiration (air entry and consequent expansion of the lung reduce diamagnetic water and increase magnetic oxygen).

the concentration of

the concentration of para-

However, there is no non-invasive way of

accurately measuring lung water other than the cumbersome multiple-indicator techniques (85).

Ethical and Legal Implications of Clinical Measurements What is non-invasive?

Procedures are considered non-invasive

when the risks approach zero.

However, immense benefit may so

weight the risk/benefit ratio that a small risk is acceptable. The MCG and the MEG are non-invasive.

Magnetopneumography is

non-invasive in that there is no known risk of the magnetizing field.

Inhalation of trace quantities of sterile pure magne-

tite for study of dust clearance is similarly without known risk.

Even non-invasive procedures should not be carried out

on human subjects without their consent.

In the case of new

170 procedures which are not standard or routine tests, written consent should be obtained.

This protects both the subject

and the investigator. In dealing with human subjects, especially those who are ill, the laboratory setting must be suitable. The fears, stated and unstated, of patients and their sensitivities must be considered. We warn subjects that they may feel static electricity from the plexiglass covering our magnetizing coil, or hear the air compressor powering the patient bed turn on, or hear a loud noise as the relay to our magnetizing coil closes. Patients hear and see everything. Disorderly facilities and careless dress are not reassuring. Loose conversation such as "Boy, he's hot," "Look at this, will you," or "Oops," may prompt the worried patient to either get chest pain or write his last will and testament. A final caveat converns the interpretation of data. Since interpretations may have diagnostic, therapeutic, and even legal implications, they should not be taken lightly. Tests should not be put in into routine clinical use prematurely. The correlation of biomagnetic measurements with various disease states must first be carefully validated. The exact specificity and sensitivity (the incidence of false positive and false negative observations) must be known before deviation from normal can be interpreted. Only then can criteria for interpretation be formulated. For example, a mistaken diagnosis of cardiac disease could change a well man into a psychologic cripple. Finally, as with all clinical information gathered with consent of the patient, confidentially must be assured. Only with the individual's consent or with a subpoena can information be released to the company, the labor union, or the government.

171 Summary Magnetic measurements in biologic systems can now be made with sophisticated sensitive equipment applicable to the hospital environment.

Biomagnetic measurements of differences

in electric potential and current flow in the heart, brain and other organs may offer new insights into function and have clinical applications.

Magnetopneumography is a unique tool

for monitoring the level of dust acquired by the worker from his environment and studying clearance mechanisms.

Magnetic

susceptability measurements may be applied to iron storage disease, cardiac output monitoring and possibly other problems. Biomagnetic applications merit more intensive research. is inadequate at the present to warrant routine measurements.

Data

clinical

Research studies should be carried out on a

scale large enough to firmly establish clinical correlations. Investigation of basic physiologic processes and disease mechanisms must also be pursued.

It is hoped that the joint

efforts of basic science researchers and clinicians will speed these efforts.

Acknowledgements The constant stimulation and biomagnetic education provided by my colleague, Stephen E. Robinson, Ph.D. is gratefully acknowledged.

I thank Gary Mintz, M.D. and Neal Schaul, M.D.

for their informative discussions on non-invasive cardiac and neurologic tests.

I also thank Ms. Debbie Wal and

Ms. June Fean for their patience in preparing this manuscript.

172 References 1. 2.

Roth, N.: IEEE/EMBS Newsletter 18, 7 (1979). Freedman, A.M., Kaplan, H.I., Sadock, B.J.: Comprehensive Textbooks of Psychiatry II, Second edition, Williams and Wilkins, Baltimore 1975 p.45.

3.

Baule, G., McFee, R. : Am. Heart J. 5!5 , 95-96 (1963).

4.

Zimmerman, J.E., Frederick, N.V.: Appi Phys. Lett. 19, 16-18 (1971) .

5.

Cohen, D., Edelsack, E.A., Zimmerman, J.E.: Appi. Phys. Lett. 16, 278-280 (1970) .

6.

Saarinen, M., Siltanen, P., Karp, P.J., Katila, T.E.: Ann. Clin. Research, 1£, Suppl. 21 (1978).

7.

Wikswo, J.P., Jr.: Non-Invasive Magnetic Measurement of the Electrical and Mechanical Activity of the Heart, Ph.D. Thesis, Department of Physics, Stanford University, (1975) .

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18.

Mosaka, M., Cohen, D., Cuffin, B.N., Horacek, B.M. : J. Electrocardiology 9, 418-425 (1976). Cohen, D., Palti, Y. , Cuffin, B.N., Schmid, S.J.: Proc. Natl. Acad. Sci. USA, In press Waltham, A.C.: In Hurst, J.W., Logue, R.B., Schlant, R.C. Wenger, N.K., Eds., The Heart, Fourth Edition, McGraw Hill New York 1978, p.373 Cohen, D.: Report of the Low Field Group: The Magnetocardiogram, Report MIT/FBNML 76-1, Cambridge, 1976 Barry, W.H., Fairbank, W.M., Harrison, D.C., Lehrman, K.L. , Malmivvo, J.A.V., Wikswo, J.P. Jr.: Science 198, 1159-1162 (1977). Matelin, D.: La Magnetocardiographie - Realisation D'un Magnétocardiographe a#Bobines. Premiers Resultats Obtenus en Milieu Urbain, Ph.D. Thesis, Grenoble,1974 Wikswo, J.P. Jr., Malmivvo, J.A.V., Barry, W.H., Leifer, M.C., Fairbank, W.M. : Adv. Cardiovasc. Phys. 2_, 1-67 (1979) . Tripp, J.H., Farrell, D.E.: Conference Digest IC SQUID 1980, pp. 206-208 Berbari, E.J., Lazzari, R., Samet, P., Sherlag, B.J.: Circulation £8, 1005-1013 (1973). Saarinen, M., Karp, P.J., Katila, T.E., Siltanen, P.: Cardiovasc. Res. 8, 820-834 (1974). Castellanos, A., Myerburg. R.J.: In Hurst, J.W., Logue, R.B., Schlant, R.C., Wenger, N.K. Eds. Op: cit, p» 302.

173 19.

Cohen, D., Norman, J.C., Molokhia, F., Hood, W. Jr.: Science 172, 1329-1333 (1971).

20.

Robinson, S.E.:

21.

Bodenheimer, M.M., Banka, V.S., Helfant, R.M.: Am. J. Card. 4_5, 661-684 (1980). Kotler, M.N., Mintz, G.S., Parry, W.R., Segal, B.L.: Am. J. Card. In press.

22.

Personal communication (1976).

23.

Goldensoln, E.S.: In Klass, D.W., Daly, D.D., Eds, Current Practice of Clinical Electroencephalography, Raven Press, New York 1979, pp. 421-450

24.

Daly, D.D.:

25. 26.

Goldensohn, E.S.: Ibid pp. 307-341 Gloor, P., Ball, G., Schaul, N.: Neurology 27, 326-333 (1977).

27.

Low, M.D.: In Klass, D.W., Daly, D.D., Eds. Op cit, pp. 441-450

28.

Cohen, D.:

Science 161, 784-786 (1968).

29.

Cohen, D.:

Science 175, 664-666 (1972).

30.

Ahopelto, J., Karp, P.J., Katila, T.E., Lukander, R., Mäkipää, P.: Report TKK-F-A233, Helsinki University of Technology 1974

31.

Reite, M., Zimmerman, J.E., Edrich, J., Zimmerman, J.: Electroenceph. Clin. Neurophysiol. 4(), 59-66 (1976).

32.

Saunders, M.G.: pp. 38-67

33.

Cohen, D.: Francis Bettes National Magnet Laboratory Annual Report, Cambridge 1977 Weisberg, L.A.: Ann. Int. Med. 91, 87-105 (1979).

34.

Ibid pp. 221-268

In Klass, D.W., Daly, D.D., Eds. Op cit,

35.

Brenner, D., Williamson, S.J., Kaufman, L.: 190, 480-482 (1975).

36. 37.

Cohen, D.C.: IEEE Trans. BME-24, 508-512 (1980) Williamson, S.J., Kaufman, L. Brenner, D.: In Schwartz B.B., Foner, S., Eds., Superconducting Applications: SQUIDS and Machines, Plenum, New York 1977, pp. 355-402

38.

Hari, R. Aittonieme, K. , Jarvinen, M.-L., Kätila, T., Varpula, T.: Report TKK-F-A399, Helsinki University of Technology 19 79

39.

Brenner, D., Lipton, J., Kaufman, L., Williamson, S.J.: Science 199, 81-83 (1978). Penfield, W., Rasmussen, T. : The Cerebral Cortex of Man, MacMillan, New York 1955, p. 214

40.

Science

174 41.

Kariniemi, v., Karp, P.J., Katila, T.E.: Med. 2 (1974)

42.

Cohen, D.:

43.

Hukkinen, M.S., Kariniemi, V. , Katila, T., E. , Laine, H., Lukander, R., Mäkipää, P.: Am. J. Obstet. Gynecol. 125, 1115-1120 (1976).

44.

Shenker, L.:

45. 46.

Cohen, D., Givler, E., Appl. Phys. Lett. 2JL, 114 (1972). Robinson, S.E., Freedman, A.P.: Unpublished observations Cohen, D.: Science 180, 745-748 (1973).

47.

J. Perinat

Personal communication (1974).

OB and GYN Survey 21, 367-388 (1966).

48.

Cohen, D.: Report of the Low-Field Group: The Magnetic Field of the Lung, Report MIT/FBNML 78-1, Cambridge 1978

49.

Davenport, H-W.: Physiology of the Digestive Tract, Year Book Medical Pub., Chicago 1977 Sleisenger, M.M., Fordtram, J.S.: Gastrointestinal Disease, Second Edition, W.B. Saunders, Philadelphia 1978, p. 1006 Karp, P.J., Katila, T.E., Mäkipää, P., Saar, P.: Digest of the XI International Conference on Medical and Bio- . logical Engineering, Ottawa 1976, pp. 504-505 Aittoniemi, K., Katila, T., Kuusela, M., -L., Varpula, TJ Digest of the XII International Conference on Medical and Biological Engineering, Jerusalem 1979

50. 51. 52. 53.

Wikswo, J.P., Barach, J.P., Freeman, J.A.: 53-55 (1980).

Science 208,

54.

Saha S., Lakes, R.S.: (1980).

55. 56.

Cohen, D.C.: IEEE Trans. MAG-11, 694-700 (1975). Kalliomäki, P.L., Karp, P.J., Katila, T., Mäkipää, P., Tos-Savainen, A.: Scand. J. Work Environ. & Health 232-239 (1976).

57.

Kalliomäki, P.L., Alanka, K., Korhonan, 0., Mattson, T., Vaaranen, V. , Kalliomäki, K., Koponen, M.: Scand. J. Work Environ & Health 4, 122-130 (1978).

58.

Kalliomäki, P.L., Korhonen, 0., Vaaranen, V. , Kalliomäki, K., Koponen, M.: Int. Arch. Occup. Environ. Health 42, 83-90 (1978).

59.

Kalliomäki, P.L., Kalliomäki, K., Korhonen, 0., Koponen, M., Sortti, V., Vaaranen, V.: Int. Arch. Occup. Environ. Health 43, 85-91 (1979).

60.

Freedman, A.P., Robinson, S.E.: conference

IEEE Trans. BME-24, 508-512

Presented at this

175 61. 62.

Freedman, A.P., Robinson, S.E., Johnston, R.F.: Med. In press Freedman, A.P., Robinson, S.E., Green, F.H.Y.: Presented at this conference

J. Occ.

63.

Stroink, G.:

64.

Morgan, W.K.C., Seaton, A.: Occupational Lung Disease, W.B. Saunders Co., Philadelphia 1975

65. 66.

Timbrell, V.: Ann. Occup. Hyg 18, 299-311 (1975). Haematite and Iron Oxide. International Agency for Research on Cancer Monographs 1, 29-39 (1972).

67.

Stern, R.M.: A Chemical, Physical, and Biological Assay of Welding Fumes, Part I (Publication 77.05), The Danish Welding Institute, Copenhagen 1977. Brain, J.D., Valberg, P.A.: Arch. Environ. Health 28, 1-11 (1974).

68.

Personal communication (1978).

69.

Albert, R.E., Lippmann, M.: 37-45 (1972).

Ann. N.Y. Acad. Sei. 200:

70.

Robinson, S.E., Freedman, A.P.: In Frontiers of Engineering in Health Care, IEEE, New York 1979, pp. 183-188

71.

Cohen, D., Arai, S.F., Brain, J.O.: 514-517 (1979).

72. 73.

Halpern, M.: Personal communication (1979) Robinson, S.E., Freedman, A.P.: Presented at this conference

74.

Beischer, P.:

75.

Kauffman, G.E., Michaelson, S.M.: In Sheppard, A.R., Eisenbud, M., Biologic Effects of Electric and Magnetic Fields of Extremely Low Frequency, New York University Press, New York, 1977, pp. 8-24.

76.

Morgan, W.K.C.:

77.

Wikswo, J.P. Jr.: Ph.D. Thesis, Department of Physics, Stanford University 1975, p. 80.

78.

Bothwell, T.H.: In Beeson, P.P., McDermott, W, Wyngaarden, J.B., Textbook of Medicine, Fifteenth edition, W.B. Saunders Co., Philadelphia 1979, p. 1790.

79.

Conley, C.L.: In Beeson, P.P., McDermott, W, Wyngaarden, J.B., Textböok of Medicine, Fifteenth edition, W.B. Saunders Co., Philadelphia 1979, p. 1778.

Science 204,

Ann. N.Y. Acad. Sei. 134 454-458 (1965).

J. Occ. Med. 20, 762-763 (1978).

176 80.

Jim, R.T.S.:

81.

Bauman, J . M . , H a r r i s , J:W:: J. Lab. & C l i n . Med. 7ff,246-257, (19671.

82.

Harris, J.W., Farrell, D.E., Messer, M.J., Tripp, J., Brittenbam, G.M., Danish, E.M., Muir, W.A.: Clin. Res. 26, 504A (1978).

83.

Wikswo, J.P., Jr.:

84.

Peters, M.J., Dunajski, Z., Conference Digest, IC SQUID Fishman, A.P.: In Fishman, Disorders, McGraw-Hill, New

85.

Lancet ii, 1028 (1979).

Op dit, p. 81. Van Der Marel, L.C.: In 1980, pp. 248-252. A.P., Pulmonary Diseases and York 1980, p. 742.

GENERATION OF MAGNETIC FIELDS BY THE HUMAN BODY (THEORY) Robert Plonsey, Ph.D. Department of Biomedical Engineering Case Western Reserve U n i v e r s i t y , Cleveland, Ohio 44106 INTRODUCTION The a s s o c i a t i o n of e l e c t r i c a l currents with b i o l o g i c a l processes dates back to the e a r l i e s t study of e l e c t r i c i t y i t s e l f in the experiments of Galvani and Volta in the late 18th and e a r l y 19th c e n t u r i e s .

Advances in

electrophysiology p a r a l l e l e d those in technology which permitted recording p o t e n t i a l s of decreasing magnitude from t i s s u e regions of diminising s i z e . Einthoven's invention of the s t r i n g galanometer in the early 1900's permitted the detection o f m i l l i v o l t level s i g n a l s at the body surface a r i s i n g from action currents in the heart.

The more recent work of

Hodgkin and Huxley, who elucidated the i n i t i a t i o n and propagation of the nerve impulse through a quantitative model, provided for a generally improved understanding of e l e c t r o p h y s i o l o g i c a l them the 1963 Nobel Prize in Physiology).

processes (and earned for

The e l e c t r o n i c a m p l i f i e r s

e s s e n t i a l to the experimental work of Hodgkin and Huxley was a l s o u t i l i z e d by Hans Burger in the 1930's to detect the e l e c t r i c a l a c t i v i t y in the b r a i n , the amplitudes of which l i e in the range of lO-lOOyv.

To the

above could be added the work of many others involved in the study of the generation of e l e c t r i c i t y by nerve and muscle in the various organ systems of the body. According to the experiments of Oersted a magnetic f i e l d i s n e c e s s a r i l y associated with the flow of e l e c t r i c current.

In p r i n c i p l e , therefore, a

magnetic f i e l d a r i s e s from the e l e c t r i c a l a c t i v i t y of a l l excitable t i s sues, such as those mentioned above.

Such biomagnetic f i e l d s

magnetic f i e l d s of b i o l o g i c a l o r i g i n ) are extremely small.

(i.e.,

The detection

o f even the l a r g e s t such f i e l d , that due to electrocardiographic

currents,

occurred only recently (1963) with the experiments of Baule and McFee[l]. This f i e l d which has a magnitude o f about 5 x 10"^ gauss i s only about 10" 6 that of the e a r t h ' s magnetic f i e l d , underscoring the d i f f i c u l t y in recording such f i e l d s .

The magnetocardiographic f i e l d , though small, i s

nevertheless two orders of magnitude greater than magnetic f i e l d s from neural a c t i v i t y in the brain (EEG).

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

arising

178 While a l t e r n a t i v e approaches are a v a i l a b l e [31] the a v a i l a b i l i t y of superconducting quantum interference devices (SQUID) provide an instrument with both the low system noise and high s e n s i t i v i t y needed for practical biomagnetic detection.

Such devices are capable of measuring the magnetic

f i e l d - a s s o c i a t e d with most currents of b i o l o g i c a l o r i g i n , although special attention to s h i e l d i n g and the use of signal and spatial averaging may be required.

The a v a i l a b i l i t y of the SQUID has opened up new f i e l d s o f study

based on the measurement of biomagnetic f i e l d s for e i t h e r basic or clinical

i n v e s t i g a t i o n of the underlying p h y s i o l o g i c a l system.

A concom-

itant has been the development of appropriate basic theory and physical models to apply in each such case. In t h i s paper we w i l l review the basic theory and i t s a p p l i c a t i o n to very simple p h y s i o l o g i c a l models.

Other papers in t h i s volume consider

s p e c i f i c organ systems in greater depth.

Although biomagnetism has had a

short h i s o t r y there are already a number of excellent review papers which 'develope the basic theory [ 2 - 7 , 3 2 ] .

This present contribution provides

a somewhat greater emphasis on the underlying e l e c t r o p h y s i o l o g i c a l

events

and the physical i n t e r p r e t a t i o n s of t h e i r associated electromagnetic sources.

Our major " r a i s o n d ' e t r e " i s , however, to give t h i s volume

completeness; and to t h i s task we are indebted to the many i n v e s t i g a t o r s on whose work t h i s paper i s b u i l t

[2-7,29],

B i o e l e c t r i c Sources The goal of t h i s section i s to enumerate the b i o l o g i c a l

circumstances

which r e s u l t in the generation of e l e c t r i c a l current (and hence, necess a r i l y , a magnetic f i e l d ) .

We w i l l pursue t h i s d e s c r i p t i v e l y and leave

for a l a t e r section a quantitative a n a l y s i s .

In a l l examples we w i l l

find that excitable c e l l s are characterized by an i n t r a c e l l u l a r

ionic

composition which i s considerably d i f f e r e n t from the e x t r a c e l l u l a r composition.

As a typical example we show, in Table 1, the i o n i c concen-

t r a t i o n s inside and outside a mammalian muscle c e l l ; the great imbalance in concentration of sodium, potassium, and chloride i s evident. The i n t e g r i t y of the b i o l o g i c a l cell i s maintained by a (plasma) membrane which separates the i n t r a c e l l u l a r from e x t r a c e l l u l a r space.

This membrane

i s also s e l e c t i v e l y permeable; that i s , i t allows d i f f e r e n t ion species

179 Ion Species

Intracellular

Na+ K

+

12

145

155

4

Msc. Cations

5

cr

HC0 3 "

12

120

8

27

155

ATable 1.

Extracellular

Ionic Composition of Intracellular and Extracellular Medium for a Mammalian Muscle Cell (mmol/i.). [A" denotes intracellular protein anions.]

differing ease of movement across the membrane. potassium permeability dominates.

At rest, for example, the

One consequence is the establishment

of a resting electrical potential which equilibrates potassium against Its concentration gradient.

That is, since the latter drives potassium from

inside to the outside of a cell, the equilibrating (resting) electric field is inward, the intracellular potential being typically -90mv relative to the extracellular zero reference. The relative membrane permeabilities can be altered by an electrical stimulus (or in certain cases by a physical or chemical stimulus) and the result is typically a phasic change involving the passage o f i o n s across the membrane until the original resting condition is recovered.

The most

common example is the action potential, characterized by a rapid rise in sodium permeability to the point where the sodium ion approaches equilibrium, necessitating a reversal in the transmembrane electric field, (the intracellular potential approaches the sodium Nernst (equilibrium) potential of around +40mv). This is followed by a rise in potassium permeability and finally restoration of resting permeabilities and potential. Since the membrane adjoining the active site is still at rest a flow of current linking the active region with adjoining inactive regions must occur (i.e., in response to the unequal transmembrane potentials).

The

action-currents have a polarity which necessarily depolarize (activate) the adjacent membrane (it constitutes an electrical stimulus) and

180 explains the mechanism for propagation of e x c i t a t i o n . i s observed both in nerve and muscle.

Such phenomena

Since most b i o l o g i c a l t i s s u e i s a

reasonably good conductor the action currents can be conducted throughout the body and t h i s accounts for the presence o f , for example, e l e c t r o cardiographic potentials over the e n t i r e body surface. The synaptic junction provides an example where a chemical substance (transmitter) acts to change the s e l e c t i v e permeability of the target membrane.

The r e s u l t i s a transmembrane i o n i c flux analogous to that

a r i s i n g from e l e c t r i c a l stimulation ( a l l i o n i c permeabilities may be e l e vated, however) and a consequent local depolarization and i n i t i a t i o n of volume conductor currents.

Physical s t i m u l i i

(such as pressure,

light,

heat) are also e f f e c t i v e in i n i t i a t i n g changes in membrane permeability in certain s p e c i a l i z e d receptor c e l l s .

The r e s u l t i s the establishment

of a l o c a l i z e d generator potenital as well as associated currents which may, i f s u f f i c e n t l y l a r g e , i n i t i a t e a propagated action potential

(hence

a signal associated with the presence of l i g h t , heat, pressure, e t c . ) . The aforementioned processes e s t a b l i s h both e l e c t r i c and magnetic f i e l d s each of which can be useful t o o l s in b a s i c , as well as c l i n i c a l ,

studies

of those processes. There are a number of examples of e l e c t r o p h y s i o l o g i c a l become useful in both basic and c l i n i c a l

systems that have

investigations.

The more impor-

tant of these are the following: 1.

Electrocardiography (ECG).

Cardiac muscle i s excitable and

gives r i s e to a c h a r a c t e r i s t i c propagating action p o t e n t i a l . associated action currents r e a d i l y flow throughout the e n t i r e body; the electrocardiogram describes the potential

variation

recorded at the body surface. 2.

Electroencephalography (EEG).

This signal measured over the

scalp a r i s e s from neural e l e c t r i c a l a c t i v i t y in the b r a i n . The EEG o r i g i n a t e s mainly in "slow" synaptic p o t e n t i a l s of c o r t i c a l nerve c e l l s . 3.

Electromyography (EMG).

S t r i a t e d muscle c o n s i s t s of long

The

181 f i b e r s each of which conducts an action potentia in the same way as a nerve f i b e r .

The EMG a r i s e s from the corporate effect of

many propagating action potentials on muscle f i b e r s . 4.

Electroretinogram (ERG). This e l e c t r i c a l signal a r i s e s from receptor c e l l s in the retina and i s an example of the receptor potentials.

The associated magnetic f i e l d has not, as y e t ,

been s u c c e s s f u l l y detected. There are conditions for which b i o e l e c t r i c sources of d.c. currents can a r i s e .

Because of unstable electrode contact potentials as well as

galvanic s k i n p o t e n t i a l s , e l e c t r i c a l s i g n a l s must normally be low-pass f i l t e r e d , consequently only limited studies of d.c. potentials have been performed.

In view of the c a p a b i l i t y of the SQUID to measure d.c. cur-

rents in the absence of direct contact with the preparation new opportuni t i e s have been made p o s s i b l e for such exploration.

D.C. currents a r i s e

in regions of cell i n j u r y presumably because healthy and damaged c e l l s have d i f f e r e n t r e s t i n g potentials so that when interconnected, as in cardiac t i s s u e , a net flow of current r e s u l t s . simply provide a current flow path.)

(The i n j u r y may also

In cardiac t i s s u e ischemia w i l l

produce changes in r e s t i n g potential and p o s s i b l e changes in action potential waveform, and therefore a l s o lead to " i n j u r y " or d.c. currents. The hypothesis has been advanced by Cohen et a l . [ 8 ] that v a r i a t i o n s in e x t r a c e l l u l a r potassium concentration can a r i s e normally, depending on d e t a i l s o f local h i s t o l o g y , and hence lead to s p a t i a l v a r i a t i o n s in s t r i a t e d muscle r e s t i n g potentials which then become a source of steady currents.

Types of E l e c t r i c a l Currents in Physiological

Systems

The quantitative study o f biomagnetism n e c e s s a r i l y s t a r t s with c o n s i d e r a t ion of the nature of the current flow which a r i s e s in p h y s i o l o g i c a l tems.

sys-

In t h i s section we consider a typical excitable cell in a physio-

l o g i c a l volume conductor.

There are then three regions to consider, the

182 i n t r a c e l l u l a r , the membrane, and the e x t r a c e l l u l a r . Both the i n t r a c e l l u l a r and e x t r a c e l l u l a r media are passive and capable of s u s t a i n i n g conduction and displacement c u r r e n t s .

By u s i n g a Fourier

a n a l y s i s formalism the r e l a t i o n s h i p between current d e n s i t y , J , and e l e c t r i c ' f i e l d E i s l i n e a r and given by j = (a+juE) E

(1)

where a i s the c o n d u c t i v i t y and J and E are complex phasors.

Equation (1)

can be rewritten as J = a(l+joje/a)E = a c E and a c i s a complex c o n d u c t i v i t y .

Power spectral a n a l y s t s o ^ b i o e l e c t r i c

s i g n a l s shows l i t t l e energy above 1000 Hz and n e g l i g i b l e energy above 10,000 Hz.

Subject to t h i s c o n s t r a i n t , and with t y p i c a l t i s s u e con-

d u c t i v i t i e s in the range o f .05 - .5 mhos/m, y i e l d s a L e / a i r a t i o o f l e s s -5 than 10

f o r a r e l a t i v e p e r m i t t i v i t y of u n i t y .

Direct measurement of

t h i s r a t i o by Schwan and Kay [9] for a v a r i e t y of t i s s u e s shows i t to be much l a r g e r but s t i l l

n e g l i g i b l e ; the conclusion being that p a s s i v e

t i s s u e can be considered to be r e s i s t i v e . apply, a l s o , to the i n t r a c e l l u l a r

We assume t h i s c o n c l u s i o n to

space.

Because of i t s high l i p i d content the membrane i t s e l f i s

characterized

by a high r e l a t i v e p e r m i t t i v i t y and very low c o n d u c t i v i t y . values for a passive membrane are 1 pf/cm as the leakage r e s i s t a n c e .

2

Typical

capacitance and 1000 ohm cm

2

Thus both displacement and conduction

currents are important ( f o r the f i g u r e s given they would be equal at a frequency of 160 Hz). Based on the r e s i s t i v e nature of b i o l o g i c a l

t i s s u e and the low freqencies

contained in b i o e l e c t r i c s i g n a l s i t can be shown that the a s s o c i a t e d e l e c t r i c (and magnetic) f i e l d s must be q u a s i s t a t i c , [ 1 0 ] .

One consequence

i s that the e l e c t r i c f i e l d can be derived from the gradient o f a s c a l a r potential f u n c t i o n , that i s E = -v®

(2)

Since the e l e c t r i c f i e l d in equation (2) i s c o n s e r v a t i v e , and volume currents in the passive i n t r a c e l l u l a r and e x t r a c e l l u l a r space a r e d i s - s i p a t ive a non-conservative f i e l d must a l s o be present and the membrane i s the only remaining region in which i t can l i e :

Such a f i e l d i s a formal

183 representation of the non-electrical source, whose energy is converted into the electrical form.

(This device is similar to the use of "EMF"

as a measure of the electrical equivalent of a chemical battery.)

This

supply of energy, equal to that electrically dissipated, arises from the (chemical) potential energy of the unequal extracellular and intracellular composition, as noted earlier.

The Nernst potential of an ion species

is, in fact, an electrical measure of the diffusional "force" associated with the unequal transmembrane concentration. While over the short run the energy required to sustain an action potential and its associated action currents comes from the aforementioned potential energy, the long-run maintenance of the intra-and extra-cellular composition itself requires a source of energy.

This is supplied from

cellular metabolism and operates through active membrane processes.

Norm-

ally such "pumps" do not need to be considered in the study of the action potential. There is no satisfactory model of the biological membrane which is completely based on physical principles although much in known concerning membrane behavior.

The Hodgkin-Huxley formulation is a semi-empirical,

though quantitative, description of the Squid axon membrane, which probably applies in a general way to other excitable membranes.

In this

description the membrane currents are enumerated as follows [11] J

m

= C (dV /dt) + sg.. v(V -V., ) + 3 g„(V - V J + g„(V -V„) nr nr ' Na m Na' KV m K m i '

x

(3) '

where the first term on the right is the capacitive (displacement) current while the next two are potassium and sodium ionic currents respectively (the last term, the "leakage" current, is small and frequently negligable).

Each ionic current depends on a non-linear conductance and a

voltage which represents the difference between the transmembrane potential and its equilibrium (Nernst) potential.

The ionic currents are

essentially the non-conservative component of the total membrane current referred to above.

We refer to these currents (densities) with the

1

symbol J , the superscript i symbolizing an impressed current.

For the

physiological volume conductor, then, the total current density, J, is given everywhere by J = -aVi+J1

(4)

184 where J 1 = 0 except in the membrane and a i s real (except in the membrane). Since the total current i s n e c e s s a r i l y solenoidal

(for the conservation

of charge) we require v-J = 0

(5)

Note that i f eq. (4) i s substituted into eq. (5) we demonstrate that $ s a t i s f i e s Poissons equation namely

Basic Field Equations - Homogeneous Unbounded Media In a t y p i c a l b i o e l e c t r i c preparation there may be many c e l l s which are simultaneously active.

An idealized model of such a system c o n s i s t s

of an impressed current density function, J 1 , representing a " c o a r s e grained" average of the c e l l u l a r sources, and which l i e s in a uniform conducting region of i n f i n i t e extent. idealized

model l a t e r .

We shall return to and refine t h i s

For the present we wish to apply Maxwell's

equations to the assumed currents in order to obtain basic expressions for the r e s u l t i n g e l e c t r i c a l potential and magnetic f i e l d . field

The l a t t e r

satisfies _

(7)

VXH =-aV4>+j1

(8)

v-H

= 0

In view of eq. ( 7 ) , H can be derived from a vector potential A, so that H = VxA

(9)

S u b s t i t u t i n g eq. (9) into eq. (8) leads to VxH = V X V X A = V(V-A) - V2A =-aV$+j 1

(10)

The function of A i s not completely defined by eq. ( 9 ) , in f a c t , according to the Helmholtz theorem [ 1 2 ] , v-A i s at our d i s p o s a l .

We choose t h i s to

s a t i s f y the Lorentz condition thereby i n s u r i n g the s a t i s f a c t i o n of the continuity of charge, equation (5). V-A =-a = V m / 0 e m

are (34)

190 a i 3 $ i _ a e 3$ e an-

an

(35)

where s u b s c r i p t i and e r e f e r to i n t r a c e l l u l a r and e x t r a c e l l u l a r conditions.

Since

discontinuous.

both the p o t e n t i a l and normal d e r i v a t i v e are However

¡p, as defined by eq. ( 2 7 ) , s a t i s f i e s

(accord-

ing to eq.s (34) and (35)) f j t 1>e

(36)

3e "in

an

and consequently the e f f e c t o f an a c t i v e c e l l

i s f u l l y accounted f o r

by a current double l a y e r l y i n g i n and d i r e c t e d normal to the membrane surface and imbedded in a region o f uniform c o n d u c t i v i t y a g e x t r a c e l l u l a r f i e l d s ) or a., ( f o r i n t r a c e l l u l a r f i e l d s ) . * =

h

SL m

1 4ira

( V M

v(a

- V 1 R

(for

Specifically

1

(37)

ä D - dS m $ - a . $ . ) K ,. m e e i t ' ^ 2

, (38)

I n eq.s (37) and (38) dSffl i s d i r e c t e d outward from the c e l l whose surface membrane i s designated S m .

A l s o , i n eq. (38) a i s the con-

d u c t i v i t y at the f i e l d p o i n t , as noted above.

The equivalent

impressed

current o f the a c t i v e c e l l can be w r i t t e n Kg = ( < > e V a i V "

(39)

where Fi i s outward from the c e l l . 3.

Surfaces A s s o c i a t e d with C o n d u c t i v i t y

Perturbations

Changes can take place i n the volume conductor impedance due to r e s p i r a t i o n , blood flow, o r , over l a r g e r periods of time, tumor growth etc.

I f an external current i s applied to such a r e g i o n then the

e f f e c t of these changes i s manifest i n a changing p o t e n t i a l to changing i n t e r n a l c o n d u c t i v i t i e s .

The potential

f i e l d due

f i e l d due to these

changes (alone) i s given by [17] 1 iïF

V0 a a

l 2

J -V(l/R)dV' 3

(40)

191 where c^ i s the i n i t i a l conductivity and a^ i s the f i n a l

conductivity

d i s t r i b u t i o n ( f o r which J = i s the applied current density) and cl Aa = a^-a-j. An equivalent secondary source i s generated as a consequence of the conductivity change and i s given, according to eq. ( 4 0 ) , by (41)

p

While the applications to magnetic f i e l d detection i s c l e a r , t h i s , of course i s not ttrruullyj an example of a biomagnetic effect since J , i s an a

exogenous current.

For each of the above examples a magnetic f i e l d i s produced and i s a s s o ciated with the primary and secondary current sources.

S p e c i f i c a l l y we

can apply eq. (16) to eqs. (33), ( 3 9 ) , and (41) to obtain " " +

ir

5ixV(l)dV^

J$k(ak"-ak-)n-xV(^)dS k k

1 A a " ..„,,1, ,„, ^ 1 3 x V d V ' •^ K' ' + t ^tt ^m e V V i ^ ' ^ m 4TT a, a, a I i m

(42)

Since b i o e l e c t r i c sources are s p e c i f i c a l l y enumerated in the l a s t express i o n , J 1 ( i n the f i r s t i n t e g r a l ) represents exogenous impressed currents only.

I t i s sometimes convenient to approximate the contribution of the

l a s t expression by an e f f e c t i v e volume source J 1 , as has been suggested earlier.

For a passive fixed medium the l a t t e r two terms in eq. (42) can

be dropped.

In the second expression, the surface integral bounding a

region with conductivity cr-can be transformed to a volume integral the following way.

in

We note the vector i d e n t i t y

v'x(ov'(l/R)

=

v'i>xv'(l/R)

(43)

and by i n t e g r a t i n g both sides over the volume containing a we have | v ' x ( o v ' ( l / R ) ) d V1 = - [ $ v ' ( l / R ) x d S = V

k

h

v' $xv' (l/R)dV'

(44)

\

Consequently the f i r s t two terms in eq. (42) transform to (J^ffV-'aOxv'O/RjdV 1 (45) 4 TTJ This can be confirmed by noting that the volume integral indicated by eq. (45) must be evaluated separately in each subregion of constant conductivity.

The f i r s t term sums over the component volumes to give the

192 total volume integral which is the first integral of eq. (42).

The second

term transforms to a surface integral by eq. (44); if one recognizes that each interface is described in two adjacent surface integrals then the second expression of eq. (42) is obtained.

In view of eq. (4), eq. (45)

could have been obtained directly from the Biot-Savart relation. Geselowitz [18] has described another form into which the second expression of eq. (42) can be transformed.

We obtain this by noting that for

each region where a is constant, the vector identities v'x[(l/R)v'$] = - v'$xv'(l/R) iV'xV 1 (l/R) = - V ' i x V ' d / R ) + V'x$V'(l/R) = 0 as well as the relation /VxBdV=/nxBdS applies.

(46) (47)

This, in turn, leads to

-Jav'fxv'(l/R)dV' = -|a(l/R)V'$xdS = Jai>V' (1 /R)xdS

(48)

The first two terms of eq. (42), with the help of eq. (48), become: H = J^'xVa/RjdV'

+ ^

E|(ak"-ak'Xl/R)V$kxdSk

(49)

Magnetic Field of a Single Fiber [19] For an active cell in a uniform extracellular conducting medium we have shown that the equivalent current source is a double layer which lies in and is directed normal to the-membrane.

This source, in turn, is

inbedded in a region of uniform conductivity of infinite extent which takes on the magnitude a Q , for extracellular fields, and a^ for intracellular fields.

The magnitude of the double layer current is given by

eq. (39) and leads to electrical potential and magnetic fields given by

*

=

JiT"

i S

(Ve-ai*i,fi'v,{1/R)dSm

(50)

m ( restatement of eq. (38)), where a=a e or cr. depending on the location of the field point, and H = (a e i e -a i i 1 OnxV'(1/R)dS i n 1

W.

(51)

S m

In the above expressions the reader is reminded that $ e and $ i are potent-

193 ials just outside and just inside the (membrane) surface.

An alternative

expression for H can be obtained by substituting eq. (39) into eq. (21) H

=

'x[(o p-a.$,)6(n)n]

1

e

S~ 4l7

e

1

1

R

dV

(52)

V where the double layer is considered as a degenerate form of a volume impressed current density through the inclusion of a transverse delta function 6(n).

The result can also be expressed as 1

H

=

f j. m

V ^ e W i ^ %

dS

m

(")

and v t operates only in the (transverse) coordinates of the membrane surface. If one thinks of the double-layer as having a small but finite thickness (within which it is uniform in strength) one notes that no source contribution arises due to the discontinuity of ( a e * e - a i $ -j)

1n

the

direction

normal to the surface since both K^ (in eq. (39)) and the surface normal a have a zero cross product in the surface integral of eq. (17), which evaluates the possible contribution from the aforementioned discontinuity. Further simplifications of eq. (53) are possible for the case that the cell is circular cylindrical, a shape that applies to a significant number excitable cells of interest (i.e., nerve and muscle fibers). of this section applies to such geometry.

The remainder

The surface normal, n in eq.

(53), translates into a p in circular cylindrical coordinates (i.e., the radial direction) while

and

can be expected to be axially symmetric,

hence, dependant only on z (the axial variable).

Thus

and eq. (53) then becomes 1 a iir fiber surface

a 0 3$ Q /3Z-a l . 3$,./3Z e e 1 1 dS b T

(55)

In eq. (55) a^ is the unit vector in the azimuthal direction at the source point dS.

194 For the s i n g l e i s o l a t e d f i b e r i t i s u s u a l l y possible to ignore the extrac e l l u l a r p o t e n t i a l s which are r e l a t i v e l y very small [20].

In t h i s case

= Vm (the transmembrane p o t e n t i a l ) and eq. (55) has the simple form of « " - J 4ir

f.s.

,

- -1J

*r-Tr^

t

15

Ö.-3V/3 Z

f.s.

Since the i n t r a c e l l u l a r axial current density J.

=

while the

e x t r a c e l l u l a r surface current d e n s i t y * J g z = -a e 3i> e /3z we can a l s o express eq. (55) as

h

4

ä

4D

^ - d S

(57)

f.s.

The second term in eq. (57) can u s u a l l y be neglected in the unbounded axon. The r e s u l t expressed in eq. (56) can also be thought to a r i s e from an axial dipole moment per unit volume J ft a z f i l l i n g the axoplasm with no v a r i a t i o n in cross section but having a strength that varies a x i a l l y as -a^ 34>^/3z. This can be v e r i f i e d because i f eq. (17) i s used to evaluate the f i e l d from such a current source (J^ i s both a current density and a dipole moment density) there i s no contribtuion from the volume i t s e l f , ( w h e r e vxJ^(z)a z = 0) but a c o n t r i b u t i o n does a r i s e from the surface integral where

fixJn(z)a, H Z

= - nx(a-3$,./3z)i = - a . a . 3 $ . / 3 z . I I Z

* S t r i c t l y J i z , l i k e J e z , should be the current density at the i n s i d e membrane surface.

However, the i n t r a c e l l u l a r current density i s f a i r l y u n i -

from in contrast to the e x t r a c e l l u l a r current density [ 2 0 ] , so that

J^

i s , to a good approximation, the total i n t r a c e l l u l a r current divided by the f i b e r cross sectional

area.

195 For a f i b e r of radius a, equation (58) i s approximated by A "

a aHTra2 z

1

4tt

30./3Z IR dz a

(59)

where R, i s the distance from the f i e l d point to the f i b e r a x i s (the c u r rent sources J A a z may be thought to be concentrated on the a x i s ) .

I f the

f i e l d point i s at a distance which i s large compared to the extent of the f i b e r undergoing depolarization then R can be assumed constant in t h i s a interval and can be set to correspond to the "center of g r a v i t y " of the d i s t r i b u t i o n "¡».¡(z), namely R^.

In t h i s case the magnetic vector potential

c o n t r i b u t i o n from f i b e r elements undergoing a c t i v a t i o n (recall s . , - ^ ) A

a a^ira^ d = - W ^ V a k

- Vrest)

is

(60)

and depends s o l e l y on the excursion of the action potential.

For recovery

a s i m i l a r expression can be obtained except for a change in s i g n and where the distance R^ replaces R^, where R r i s from the "center of g r a v i t y " of the region of recovery to the f i e l d point. For the cardiac action potential the a c t i v a t i o n and recovery components may be temporally separated but for other excitable t i s s u e both c o n t r i b u t ions must be considered together. be expressed as H

=

The magnetic f i e l d from eq. (60) can

2 a • ira vxA

d = - i H V a k

- W v

where v operates at the f i e l d point.

v

n / R

d

)

(6i)

The magnetic f i e l d generated by the

depolarization region alone i s , therefore, d i p o l a r [ 0 ( 1 / R 2 ) ] , while

if

both a c t i v a t i o n and recovery are considered the f i e l d i s n e c e s s a r i l y quadrupolar [ 0 ( 1 / R 3 ) ] .

Magnetic F i e l d of Fiber Bundles - ( I n f i n i t e Medium) The f i e l d generated by a compound action potential can be determined by applying eq. (55) to each f i b e r and superposing the r e s u l t s .

For the

special case where a l l f i b e r s have the same diameter and are excited synchronously no d i s p e r s i o n w i l l occur and the propagating action potential on a l l f i b e r s i s i d e n t i c a l .

In t h i s case the i n t r a c e l l u l a r and i n t e r -

s t i t i a l p o t e n t i a l s are related by the l i n e a r core conductor model [21].

196 S p e c i f i c a l l y i f V m (z) i s the transmembrane potential ( f o r each f i b e r ) , r . i s the axial resistance per unit length for a f i b e r and r g i s the i n t e r stitial

axial resistance per unit length associated with each f i b e r , then oi(z)=^V

(z)

m

(62)

•e(*> = - K T F - V z ) i e

(63)

I f there are N f i b e r s , a total cross-secton of A, and f , f . are the e l i n t e r s t i t i a l and i n t r a c e l l u l a r f r a c t i o n a l areas (fg+f^ = 1) then r

e = i r h e e

r

i

(64)

while =

=

W1 1

(65)

cr.j ira

where a i s the f i b e r radius. When eq. (62) and eq. (63) are substituted into eq. (55) one obtains as the f i e l d for each f i b e r : i

H

-

1

on a • (

0

1

1

3V

m

f i ber surface In eq. (66) R i s the distance from each source element of a p a r t i c u l a r f i b e r to the f i e l d point.

When the distance from the f i e l d point to the

f i b e r bundle i s large the contribution from each f i b e r i s approximately the same and the total f i e l d i s N times that for a representative

fiber.

Note that when the f i b e r packing i s dense (f-j > > ' ( : 0 ) the i n t e r s t i t i a l medium plays an important part in e s t a b l i s h i n g the magnitude o f the sources and can c e r t a i n l y not be neglected.

E x p l i c i t Consideration of Membrane Impressed Current Sources In the d e r i v a t i o n of eq. (51) the membrane region was not e x p l i c i t l y considered since the equivalent current source of eq. (39) was u t i l i z e d and t h i s was based on the assumption that the membrane could be considered as

197 a surface.

In t h i s section we provide an alternate derivation in which the

membrane region i s considered e x p l i c i t l y .

This w i l l permit an examination

of the b a s i s for the approximations leading to eq. (55). we consider a c i r c u l a r c y l i n d r i c a l c e l l .

For d e f i n i t e n e s s

Within the membrane we have a

uniform r a d i a l , (outward) i o n i c current density J ^ o n a p plus currents which are proportional to the e l e c t r i c f i e l d in the membrane (conduction and displacement currents) e s s e n t i a l l y as expressed in eq. (1).

Assuminq

axial symmetry the membrane e l e c t r i c f i e l d has a radial and axial component only.

I f the c o n s t i t u t i v e properties of the medium are represented by

the complex phasor a , as discussed e a r l i e r , then the total membrane current J m can be written. m (67) where E and E , are the radial and axial components o f the membrane mp mz field. As noted before

i s the impressed (primary) current density.

Secondary

sources a r i s e at the membrane-intracellular and membrane-extracellular interface.

Application of eq. (49) y i e l d s H

=

J

ion xv'(l/R)dV+^ z ( o j ' - y M E / R i x d S j

(68)

where Sj i s the interface separating regions of d i f f e r e n t conductivity CTj',aj",E

= - v ' $ i s the e l e c t r i c f i e l d and dS^ i s oriented from the prime to

double prime region.

In our application the volume integral extends only

over the membrane while j = 1,2 denotes the membrane interface with the i n t r a c e l l u l a r and with the e x t r a c e l l u l a r regions.

In evaluating the

surface i n t e g r a l s in eq. (68) only the tangential e l e c t r i c f i e l d at the interface in required; in fact the c o n t i n u i t y of t h i s component f i e l d across each interface has been u t i l i z e d in d e r i v i n g eq. (68); (e.g. E

mz =

" V

3

^ "

The f i r s t term in eq. (68) can be modified by the transformation that c a r i e s eq. (16) into eq. (17).

Since J i Q n i s normal to the upper and

lower surfaces bounding the membrane the r e s u l t , using eq. (17), i s that / 3ionxv'(l/R)dV' = /(VxJ1on)/RdV mem mem

(69)

198 (70)

= ; S ( a j i o n / 3z)/RdV' mem where the volume integral i s taken within the membrane. The final result from formal substitution into eq. (68) i s thus "•irC mem'

(a i -a m )(3î> i / 3z)

3jion/8z

R

%(V

-dV' -

C T

J{V

3 z

>

d S ]

»VERSE X

i ^ - n

i/-

'"1 f

*

0.10 pAm V 0.10 MAm :

PATIENT: S. I5-JAN-76, 22:47:31 SECONDS SAMPLED:24 LOCUS 20 RECORDING 1 POSITION: ANTERIOR

P,

EH

0.20 mV

C

0.10 uAm 2

c

SUMMARY TVPE: MAGNETIC

0.10 pAm2 lml -1 1-1 !—

OK) pAm

MHV PATIENT: 24 28-APR-76, 20:02:38 SECONDS SAMPLED: 20 LOCUS 1 RECORDING 1 POSITION : ANTERIOR

(a)

(b)

Fig. 5. MHV components and dipole loops for a normal subject (a) and MHV and EHV components and dipole loops for a subject with a left bundle branch block. It is seen that the hypothesis of the MHV-EHV perpendicularity is not valid for this abnormal case, (from Barry et al. (21)).

Fetal MCG A research area of interest in cardiomagnetism is fetal magnetocardiography (FMCG). The FMCG was first observed by Kariniemi et al. (31). A typical FMCG recording is shown in Fig. 6. One interest for performing FMCG measurements can be seen in Fig. 6: the FMCG is relatively free from the artifact caused by the maternal heart. Therefore it would be well suited for the clinically important fetal heart rate (FHR) monitoring, which often fails during the third trimester of gestation when it is derived from abdominal fetal electrocardiogram (FECG). Using FMCG to support the abdominal FECG

231

Fig. 6. Simultaneously recorded FMCG (upper trace) and abdominal FECG (lower trace) signals. The fetal QRS complexes are denoted F and the maternal ones M. Note that 10 nG = 1 pT. (from Hukkinen et al. (32)). the FHR pattern readability could be augmented by almost a factor of two (32) . Thus FMCG could be used to complement abdominal FECG in FHR monitoring during the third trimester of gestation, when early detection of pathologic FHR characteristics could lead to successful operational care. The form of the FMCG (or the FECG) has not been of such interest as the instantaneous FHR. Fig. 7 shows an FMCG obtained with the aid of signal averaging (33).

Fig. 7. An FMCG obtained by averaging fetal complexes (33).

232 After the birth the heart of a newborn infant undergoes a change when its lungs assume their normal function. This change is reflected in the neonatal MCG (NMCG), shown in Fig. 8.

1

2

3

NMCG

4

j

—A—.a

A

5

6

.

(

—

— -

—

1

2

3

4

5

6

Fig. 8. A neonatal MCG, recorded from a six days old infant (34). The vertical spacing of the measurement points is 2 cm and the horizontal spacing 3 cm. The point D4 of the grid is just above the heart of the subject. The vertical scale is 10 pT and the length of each recording is 0.4 s. One can compare the NMCG of Fig. 8 with the normal adult MCG morphology shown in Fig. 2. The NMCG differs considerably from the adult MCG, but the mechanisms underlying these changes are not yet fully understood.

233 DC MCG Ischemic conditions and infarctions give rise to ST segment shifts both in the MCG and in the ECG. It has been questioned whether the ST segment shifts result from primary bioelectric activity or whether these shifts are a consequence of a DC injury current flowing during the resting state of the heart when ischemic changes in the resting potentials of myocardial cells are present. These hypotheses are extremely difficult to verify, due to the difficulty of DC ECG measurements. The MCG can be relatively easily recorded down to DC, and an MCG measurement for studying the existence of an injury current has been made by Cohen and Kaufman (35). They studied how the MCG baseline changed when ischemia was artificially caused in dogs by occluding one of the coronary arteries with an implanted cuff. The absolute baseline level was then measured by wheeling the dog to the range of a SQUID magnetometer and out again. A result of such a measurement is shown in Fig. 9. |4x10' 7 gauss

Fig. 9. DC MCG of a dog during occlusion of a coronary artery. At the center of the trace the dog is in the range of the magnetometer while to the left and to the right of the trace it is out of that range. The change in the baseline suggests that the primary effect is an injury current flowing in the heart during the resting state. The injury current is interrupted by the ventricular activation which is seen as the return of the ST segment to the initial baseline, (adapted from Cohen and Kaufman (35)). The result of Fig. 9 strongly suggests that the ST segment shift is a secondary effect caused by the interruption of a primary injury current during the ventricular depolarization.

234 These r e s u l t s apply to early a n d r e v e r s i b l e t i o n s . C o h e n et al.

condi-

(36) in an e a r l i e r e x p e r i m e n t a l study

Miller and Geselowitz

(37,38) in a m o d e l study suggest

at a later p h a s e or u n d e r i r r e v e r s i b l e segment c o m p o n e n t m a y Further experimental

ischemic

and

that

conditions a primary

studies w i t h h u m a n s u b j e c t s are

required

to c o m p l e t e l y v e r i f y the d i f f e r e n t h y p o t h e s e s w i t h v a r i o u s chemic

ST

exist.

is-

conditions.

His-Purkunje

s y s t e m MCG

The m a g n e t i c

s t u d i e s of the H i s - P u r k i n j e

(HP) s y s t e m are a

r e l a t i v e l y n e w r e s e a r c h a r e a in c a r d i o m a g n e t i s m . A first

ob-

s e r v a t i o n of a HP m a g n e t o g r a m was r e p o r t e d by F a r r e l l et al. (39)- F u r t h e r o b s e r v a t i o n s a n d e x p e r i m e n t a l ported elsewhere

in this v o l u m e

s t u d i e s are

(40). The HP a c t i v i t y

MCG is s e e n as " r a m p s " d u r i n g the H - V i n t e r v a l of the cycle

(see Fig. 5, Ref.

re-

in the cardiac

(40)). S u c h a ramp c a n be s h o w n to

arise w h e n a p r o p a g a t i n g a c t i o n p u l s e in a f i b e r is a p p r o a c h i n g the m a g n e t i c d e t e c t o r three

(4l). In the h e a r t t h e r e

separate c o n d u c t i o n p a t h s in the P u r k i n j e

the c o n t r i b u t i o n of

system,

all of t h e m is in p r i n c i p l e

are and

seen in the

HP m a g n e t o g r a m r e c o r d e d o n the t o r s o s u r f a c e . Since

source

l o c a l i z a t i o n f r o m m a g n e t i c f i e l d r e s u l t s h a s b e e n s h o w n to give e x c e l l e n t r e s u l t s in c e r e b r a l e v o k e d r e s p o n s e s

(42),

one c o u l d t h i n k that HP m a g n e t o g r a m s c o u l d in the future be u s e d in c l i n i c s to h e l p in HP s y s t e m

Theoretical

diagnostics.

Studies

The b a s i c p r o b l e m in c a r d i o m a g n e t i s m is to c a l c u l a t e the

mag-

n e t i c f i e l d w h e n the c a r d i a c a c t i v a t i o n p r o c e s s is k n o w n . p r o b l e m is t e r m e d the f o r w a r d p r o b l e m , a n d it w i l l be

dis-

This

235 cussed in the first part of this chapter. Clinically, a more interesting problem is the inverse one, i.e. the determination of the cardiac activation sequence when the cardiac field has been measured. Different approaches to the inverse problem are presented in the second part of this chapter.

The forward problem The original theoretical contributions to the solution of the MCG forward problem have been given by Geselowitz (43), Grynzspan (44) and Grynzspan and Geselowitz (45). A detailed presentation of the theory is given in a review by Wikswo et al. (46) . Also the generation of the magnetic fields by the human body is discussed in detail by Plonsey elsewhere in this volume (4 7). Therefore only the results relevant to this presentation are summarized here. The total current density ;j(r') in a volume conductor can be written as: (1)

j(?') = j1«?') - a(r') V-.M?1) .

Here the cardiac bioelectric sources are represented by an "impressed" current density j 1 ^ 1 ) , and the currents in the passive volume conductor by the second term on the right-hand side of equation (1) (Ohm's law). If we consider the inhomogeneous and bounded volume conductor to be piecewise homogeneous, the equation for the electric potential can be written in the form: (2)

(r) = -

4ira

+

r-r'

r-r'

I j

dS 1

[(a'-a-'J^ir^nif') •

S

j

236 where o is the conductivity of the region where (r) is calculated. n(r') is the unit normal vector of the j-th surface and it is directed from the region with conductivity a' to the region with conductivity a". A similar equation can be written for the magnetic field using the same symbols as in equation (2): o)

4it

" X 3 X < ? '> d 3 r'

+

I j S

(a'-a")

•

Thus with a knowledge of d, obtained from a measurement of A, together with a measurement of the maximum field B m the magnitude of Q can be deduced. Although only the tangential component of Q contributes to the external magnetic field, both tangential and normal components contribute to the electric potential at the surface of the half space.

This potential is an expression of the volume current

that spreads from the current dipole.

The source of the volume

current can be imagined equivalently as being a charge dipole of moment p = ke Q Q/a whose charge separation equals the length of the current dipole.

Here k is the dielectric constant of

the medium, a is the conductivity of the medium, and e Q is the permittivity of free space.

The volume currents flowing

outward from the positive charge and inward toward the negative charge match the impressed current of the dipole itself. The resulting potential V at the surface can be deduced by conventional arguments based on assuming an image dipole to satisfy the boundary condition of no outward flow.

If Q is

inclined from the z-axis by an angle 9' toward the +y-axis we can express the contribution of the tangential component of the dipole as

Vm = "T

Q sin9' - ,2 2-rrcyd

y

(6)

T7~

2 ~ 2, 3/2 (1 + x + y )

This has precisely the same form as Eq. 3 for B z but the pattern is rotated by 90° about the z-axis.

Moreover there

will be a contribution to the potential from the component of the dipole that is normal to the surface: =

N

"

Q cose-

2uad

2

.

1

(1 + X 2 +

y2)3/2

(7)

370 The isopotentials shown in Fig. 5 vary in a complicated way as the angle of inclination 0' is varied.

One remarkable

feature is the dominant influence exerted by the end of the dipole which is closest to the surface.

Moving from the

symmetrical tangential orientation of 9 ' = 90° by reducing 9' by only 15° produces a factor of 2 difference between the magnitudes of the potentials at the two extrema.

A

reduction to 9' = 60° leaves only a hint of the weaker region. Only a very careful and sensitive mapping of the isopotentials could reveal the presence of a tangential component for orientations 8'84%). 1000 (X20)

O

100

x

N

X

DC

b $J O a.

10

10

12

FREQUENCY Hz

Fig. 2: Evidence for correlation between the 9 cm brain spectra of Figure 1 (full curve) and a heart spectra for the same subject (dashed curve). Note scale change for the brain spectrum.

428 This i m p l i e s t h a t the s t r u c t u r e on the l a t t e r above 3 Hz. mainly a r i s e s from the MEG. Below 3 Hz., however, i t i s e v i d e n t t h a t the b r a i n makes the dominant c o n t r i b u t i o n t o the observed power. T h i s i n t e r p r e t a t i o n i s supported by t h e s t r o n g i n c r e a s e noted below 3 Hz. ( s e e F i g u r e 1) as t h e d e t e c t o r approaches the head, comparable with the i n c r e a s e s observed a t Hz. and ^10 Hz. For a l l o t h e r s u b j e c t s , t h e c o n t r i b u t i o n from h e a r t a c t i v i t y i s s m a l l enough t h a t the 3 Hz. and 6 Hz. r e g i o n merge i n t o one l a r g e r "window". The low frequency regime o f the PSD from 2 t o 6 Hz. can be f i t t e d t o a f u n c t i o n o f the form P = Aexp(f m ) where P i s the power s p e c t r a l d e n s i t y , f i s the frequency and A and m are c o n s t a n t s with a r e g r e s s i o n c o e f f i c i e n t i n e x c e s s o f .82 f o r a l l 5 s u b j e c t s , with n i n t h e range from - . 0 8 2 t o - . 3 9 5 and A from 5 . 6 9 t o 7 . 7 7 . Simpler m f e x p r e s s i o n s ( e . g . , P = Ae~ ) gave s u b s t a n t i a l l y p o o r e r f i t s when a l l s u b j e c t s were c o n s i d e r e d . I n summary, although obvious h e a r t c o n t r i b u t i o n s have been i d e n t i f i e d , f o r a l l s u b j e c t s t h e r e are well-marked "windows" i n which the b r a i n s p e c t r a are f a i r l y f r e e o f h e a r t a c t i v i t y . For our s p e c i f i c c a s e , t h e s e a r e below 3 H z . , Hz. and ^10 H z . , as i n d i c a t e d i n F i g u r e 1 , and i n g e n e r a l t h e e n t i r e r e g i o n below 6 Hz. i s r e l a t i v e l y f r e e o f such c o n t a m i n a t i o n .

Discussion Examining the s p e c t r a o f F i g u r e 1 i n a l l t h r e e windows i n d i c a t e s a s t r o n g dependence on d i s t a n c e . As mentioned, t h i s i n turn argues f o r a c o r t i c a l o r i g i n f o r the a c t i v i t y i n each r e g i o n . The 10 Hz. o r s o - c a l l e d " a l p h a " a c t i v i t y i s o f c o u r s e the most conspicuous and c h a r a c t e r i s t i c f e a t u r e o f the human ECG. However, as p a r t o f t h e g e n e r a l i r r e p r o d u c i b i l i t y mentioned p r e v i o u s l y , we have found t h a t even

429 t h i s 10 Hz. a c t i v i t y i s s u b j e c t to c o n s i d e r a b l e v a r i a t i o n . Not only d i d i t not appear a t a l l i n one of our s u b j e c t s , b u t i n two of the o t h e r s i t appeared and d i s a p p e a r e d i n a very c a p r i c i o u s manner. Furthermore, i t s v a r i a b i l i t y could not be c o r r e l a t e d with any obvious p h y s i o l o g i c a l p a r a m e t e r s of our s u b j e c t s . In the c o n t e x t of a l l t h i s i r r e p r o d u c i b i l i t y i t i s t h e r e f o r e of c o n s i d e r a b l e i n t e r e s t t h a t the one completely r e l i a b l e and c h a r a c t e r i s t i c magnetic a c t i v i t y of the human b r a i n a p p e a r s to be t h a t from 2 t o 6 Hz. This was p r e s e n t on every s u b j e c t with a d i s t a n c e dependence s i m i l a r t o t h a t shown in F i g u r e 1. Because the computer r o u t i n e d e s c r i b e d used segments 2 . 5 s e c . in l e n g t h , the r e s u l t s become l e s s r e l i a b l e a t f r e q u e n c i e s below ^2 Hz. The b e h a v i o r in the low frequency window of F i g u r e 1 was t h e r e f o r e checked with another r o u t i n e which used 2 0 . 5 s e c . n o n - o v e r l a p p i n g epochs and a Hanning (8) window. No s i g n i f i c a n t d i f f e r e n c e was o b s e r v e d in the PSD r e s u l t s . Although the b e h a v i o r t h e r e f o r e seems w e l l e s t a b l i s h e d , i t would be premature a t t h i s time t o s p e c u l a t e on i t s o r i g i n . Further work i s f i r s t r e q u i r e d to b e t t e r c h a r a c t e r i z e and map i t , p r e f e r a b l y with a 2nd o r d e r g r a d i o m e t e r s o t h a t the h e a r t a r t i f a c t can be e s s e n t i a l l y e l i m i n a t e d . In c o n c l u s i o n , the problem of MEG power spectrum e s t i m a t i o n has been c o n s i d e r e d . Over 5 hours of MEG d a t a o b t a i n e d from a SQUID g r a d i o m e t e r r e v e a l s s p e c t r a which vary smoothly with f r e q u e n c y . The i n s t r u m e n t a l n o i s e power i s 'v-lOO times s m a l l e r than the MEG power a t 10 H z . , however, the major n o i s e i s not i n s t r u m e n t a l b u t r a t h e r contamination from the MCG. Frequency windows have been i d e n t i f i e d where the MEG spectrum may be viewed r e l a t i v e l y f r e e of such c o n t a m i n a t i o n . Two of the windows, below 3 and 6 H z . , e f f e c t i v e l y o v e r l a p f o r most s u b j e c t s s o t h a t in g e n e r a l the r e g i o n between 2 and 6 Hz. a p p e a r s to c o n t a i n s t r o n g c o r t i c a l a c t i v i t y . T h i s r e g i o n of the PSD i s much more r e p r o d u c i b l e and s u b j e c t independent than any o t h e r f e a t u r e of the normal MCG power

430 spectrum. It is therefore possible that departures from the norm, when they do occur, would contain clinically useful information. Although a clinical system capable of recording this activity could probably be implemented (at very considerable cost) further research work of the sort described here is essential to first validate (or disprove) the interpretation presented.

Ackn ow le dgme n ts I would like to thank Dr. D. Farrell for suggesting this research and for placing the superb instrumentation of the Case Biomagnetic facility at my disposal. Thanks are also due to many members of his group, particularly J. Tripp, M. Dietz, D. Jaffe and J. Patrick for discussions and help with various aspects of this investigation.

References 1. Cohen, D.: Science 175, 664 (1972). 2. Hughes, J. R., et al.: Electroenceph. and Clin. Neurophys. 40, 261-278 (1976). 3. Reite, M., et al.: 4. Zimmerman, J.:

Ibid, 59-66.

Private communication.

5. Farrell, D., Tripp, J., van Dören, C.: to this conference.

Contributions

6. Welch, P. D.: IEEE Trans. Audio. Electroacoust. AU-15, 70-73 ( 1967). 7. Oppenheim, A. V. , Schafer, R. W.: Digital Signal Processing, Prentice Hall, Inc., New Jersey, 241 (1975). 8. Beauchamp, K. G.: Signal Processing Using Analog and Digital Techniques, John Wiley & Sons, New York 356-361 (1973) . 9. Hewlett-Packard:

Application Note 245-1, 12

(1979).

EVOKED MAGNETIC FIELDS REVEAL DIFFERENT VISUAL AREAS IN HUMAN CORTEX

D. Brenner, Y. Okada, E. Maclin, S.J. Williamson, and L. Kaufman Departments of Psychology and Physics, New York University 6 Washington Place, New York, New York 10003, U.S.A.

Introduction The classic method for studying the gross electrical activity of the human brain is to measure the differences in electrical potential between electrodes attached to the scalp. These potential differences develop as a result of volume currents that spread out from their source in active neural tissue throughout the conductive media of the head. This method has now been supplemented by one in which superconducting devices with sufficient sensitivity are employed to detect magnetic fields outside the head (1). These fields are associated with current flowing within the cortex. Theoretical considerations as well as empirical evidence point to the fact that these two methods, the method of measuring potentials and the method of measuring fields, provide complementary information despite the fact that both phenomena probably have common sources. As Okada et a^. (2,3) point out in this conference, the somatic evoked field is similar to the somatic evoked potential when measured at the surface of the cortex. In experiments performed by Goff et al. (4) it was found that the pial response — the response measured between an electrode on the surface of the cortex and linked earlobes — falls off rapidly when the active electrode is moved a short distance from the known site of neural activity in the vicinity of the

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

432 Rolandic fissure. When the somatic evoked potential is measured at the scalp the response is relatively unaffected by moving the active electrode over much larger distances. This difference between the pial and scalp recordings is no doubt due to the fact that the intervening media — cerebral spinal fluids, the dura, and the skull and the skin — smear the volume currents. One consequence of this is that the electrode on the scalp "sees" the superimposed activity of many sources, even some distant from the site of recordings. Moreover, the so-called "inactive" electrode is not truly indifferent and it makes an unknown contribution to the recorded response. The widespread volume and skin currents that produce the scalp recorded evoked response could not be major contributors to the detected magnetic field. The reason for this is that small changes in the position of the pickup coil result in a large variation in response magnitude. This variation is similar to what is seen in the pial recordings but not in the scalp recordings. Apparently, the current density in the region giving rise to the large pial potentials is sufficiently great to produce detectable fields outside the head. We may assume from the similar behavior of the pial and external field recordings that the intervening media are transparent to the field and that the field measurements are in many ways equivalent to the electrical study of the exposed brain. This conclusion is strongly supported by the apparently high degree of spatial resolution afforded by magnetic recordings. For example, stimulating the little finger produces a field pattern about the head similar to that which would be produced by a current dipole oriented orthogonally to the Rolandic fissure. Brenner et al. (5) found that they could localize this equivalent current dipole to within 1 cm. Very, similar field pattern is produced by stimulation of the thumb of the same hand. However, the hypothetical equivalent current dipole that would produce the thumb's field was

433 located 2 cm lower on the head along the Rolandic fissure. Okada et al. (2) report similar resolution for the transient field evoked by median nerve stimulation. Such resolution has not been attained with conventional electrodes attached to the scalp. The auditory evoked field is another source of evidence for the assertion that the magnetic method provides a high degree of spatial resolution. The auditory evoked field is sharply localized in the vicinity of the Sylvian fissure, the site of the auditory projection areas (6). The auditory evoked potential is strongest when the active electrode is loc&ted at the vertex and it is quite strongly represented at many locations about the head. One limitation of magnetic field measurements follows from the fact that they do afford so high a degree of spatial resolution. Unlike evoked potential recordings, the magnetic recordings show no sign of far-field effects such as responses that arise in the brain stem a few milliseconds after sensory stimulation. Such far-field effects are undoubtedly due to the weak volume currents that spread out from distant sources. While these are sufficient to produce detectable potential differences at the scalp, their local brain stem current is too deep inside the head to produce a detectable magnetic field. In addition, the symmetry of the volume currents may well produce a vanishingly small net magnetic field. In view of the resolution that is possible in recording from the somatic and auditory systems, it is of some interest to determine if that resolution makes it possible to obtain useful information about the human visual cortex. The visual cortex is a complex structure composed of several different areas. These areas occupy relatively large portions of the primate brain and the visual field is fully mapped onto several of them. Consequently, given a high degree of spatial resolution in the detecting system it may well be possible

434 ultimately to separately study the activity of these visual areas in response to diverse kinds of visual stimuli. This capability, if possible, may have important applications in clinical neurology. With this motivation in mind, we measured the visually evoked magnetic field in some detail — studying how it varies with the position of the 2.3 cm diameter pickup coil over the scalp as well as with the properties of the stimulus. Our data suggest that it is possible to detect different "visual areas". Maps of the responses obtained at various positions in the posterior portions of the head indicate the presence of more than one source and, moreover, the recorded responses differ as a function of stimulus parameters. The data we obtained differed greatly among subjects but this is to be expected in view of the wide range of individual difference in cortical geometry (7).

Methods The stimuli were sinusoidal gratings generated on the facfe of a CRT and reversed temporally in contrast in a square wave fashion. Three viewing conditions were employed. These were produced by having the subject fixate the center of the 9 deg diameter display when it was filled with a grating pattern or when either the left or right halves of the grating were occluded. In the latter two cases this resulted in stimulation of either the left or right hemiretina and, consequently, the left or right hemispheres of the brain. The spatial frequency of the grating employed here was 5 c/d, ? its contrast was 33%, and its average luminance was 52 cd/m . Since the contrast reversal rate was 13 Hz, the detector (SQUID) output was filtered at 13 Hz before averaging to provide a steady state sinusoidal response which is completely

435 characterized by its amplitude and phase. The phase is measured from the stimulus to the maximum field directed out of the head.

Results and Discussion Figure 1 contains maps of responses obtained from subject RL when he was stimulated by the left, right and full 9 deg visual display. The maps derived from left and right halffield stimulations are nearly symmetrical. Each contains two regions of strong responses located over the appropriate cerebral hemisphere. The orientations of the dark bars in the figure indicates the response phases. (The thin bars repreRVF

FF

*56 40 -104- 32-

J O

1 »8

LVF

r'^r/j

]

28- BJ^ISb^Jlf 80

3^ 8Cr

' I S 1 3 6 / 2J2 "¥^^16^116 *72 JO

o «

ZdC 2^)^80/^0^86.

7Ì Ìb4 // Pl4/yà\ 1^6

« 0 I

-4

L

-2

0

J

*Z

I

»4

I

-4

L _L J

-2

0

+2

I

+4

-4

J

-2

I

0

L

+2

+4

distance from midline ( c m ) evoked f i e l d

:

femtotesla

reversal rate = 13 Hz

isochamp contours = 1 0 0 , 2 0 0 , 3 0 0 , 4 0 0 f T

spatial frequency - 5 c/d

observer : RL

contrast = 3 3 7 « luminance = 5 2 c d / m 2

Figure 1. Visual responses recorded at various points about the scalp of subject RS for stimulation of the right visual field (on left), left visual field (on right), and full visual field (center). Solid lines are isochamps. Thè short lines indicate response phases; thin lines indicate responses at or near the noise level.

436 sent responses not significantly different from the level of the background noise.) The phases of the responses in the two regions from each hemisphere are approximately 180 deg apart. The appearance of the pattern and the reversal of phase suggests a current dipole model. This hypothetical dipole is located in the appropriate hemisphere for the stimulated visual field. To a first approximation, the full field response map is equal to the vector sum of the responses resulting from separate stimulation of the two half-fields. The data are not completely described by two current dipoles, one in each hemisphere, as indicated by a careful analysis of the phases of the responses. It is necessary to consider the role of noise in this analysis. If the response consists 6f a signal from the brain and noise having half the amplitude of the signal, then, in the worst case, when the signal and noise are 90 deg out-of-phase, the resultant phase will differ by ±27 deg. Now, if only one current source is retsponsible for the field in each hemisphere then the phases will be either in or 180° out-of-phase with each other. The data for the left and right visual stimuli contain responses whose phases differ from each other by amounts other than 180 degrees and which can not be explained on the basis of noise alone. For example, the phase at (0,4) in the right field map is 90 degrees out-of-phase relative to the maximum responses at (-2,-2) and (-2,6). The responses at their maxima are 8 to 10 times the noise level and the response at (2,4) 2.5 times the noise level. Therefore, at most noise can only be responsible for 29° of the 90 degree phase difference. Apparently, some other source is contributing to the response. Figure 2 shows maps from a second subject (DB). The same stimuli were used. Instead of trying to construct isochamps (iso field strength lines) as in Figure 1, the left and right maps have been marked off into two groups of isophasic responses about 180 degrees apart.

437 RVF

72n «'-SS -36 ~63

81- 49- 31 58 22'

36.1 67si|40 1 |II2 1 22-

108- 94' 5 8' '31 ,81

45v 63v 9 9^ 40^,54

126' 90— 130' 8IS II7S

4 9x ,3 6 I44v I2IV 243^

l98-369"202 63\ I89x

2 7' 8 5-j 76x I98v 249(1

I6&j 8 5'>576 ^8 5 I03

262x I44'*6I2 \39 -108

103' 58' 135^/63 ¿112

202^ 5 4^,112,207,207 1 ( 279n229s252v I89\/I26l 1 1 1 198^/40 I I98s I08s 63\|

270— 22' \ 12 ^126 —157

49^; I90 *I93 ^63 >84

l

S

„ 10 - 108' 153225 I08 E 2 8 - 8l' 81'ibo'Vs C 0

* 6 s 1

4

8 ^ 10

0 - 2

L VF

FF

121-i26i* 112' Nag V39 S

X

247* 162s, II2\I20^I35

l|06 ^58 \l30 \89 XI98

196\ 72x 81\

/58

>90 '85 '112 "»112

1 4

-4

I -4

-4

I -2

1 0

,27

1 2

-2

0

2

4

distance from midline (cm) evoked f i e l d ' femtotesla

reversal rate - 13Hz

observer' DB

spatial frequency • 5 c/d contrast • 3 3 % luminance • 5 2 cd/m*

Figure 2. Visual responses recorded at various points about the scalp of subject DB.for stimulation of right visual field (on left), left visual field (on right) and full visual field (center). Dashed and dotted lines enclose responses of approximately the same phase. The short lines indicate response phases; thin lines indicate responses at or near the noise level. There are several responses that cannot be placed in either group. Their phases are about 90 degrees apart and some of them are 18 0 degrees out-of-phase with each other. For example, in the left field map, the responses (-4,4), (-2,4) and (0,4) are 180 degrees out-of-phase from the responses at (2,4) and (4,4) while all these responses are approximately 90 degrees out-of-phase from the responses grouped by the dotted and dashed lines. In the right visual field map, the responses at (0,2), (2,2) and (4,2) are 180 degrees out-of-phase relative to the responses at (-2,4), (-4,6), (-4,8), (-2,8) and other positions as well. Again, these responses are about 9 0 degrees out-of-phase from the responses within the isophasic regions.

438 The existence of multiple sources has been confirmed by studying the response phase variation as a function of stimulus parameters. For subject DB Figure 3a shows the phase versus reversal rate data recorded at position (0,4) for full-field stimuli. The full-field response recorded at position (0,4) is dominated by the activity of the left hemisphere. We previously reported that for a given spatial frequency and contrast, the response phase at this same position is proportional to reversal rate (stimulus frequency) over a range of from 8 to 20 Hz (8). Moreover, the slope of the. function (latency) relating phase and stimulus frequency

5 10 19 REVERSAL RATE (Hz)

10

19

REVERSAL RATE (Hi)

20

Figure 3. Phase of response versus stimulus reversal rate for stimuli at various spatial frequency measured at a) a point on the midline of the scalp of DB and 4 cm above the inion and b) on the midline and 9 cm above the inion.

439 ( a = .01).

The latencies for the 1 c/d grating was 71 msec

greater at the higher position; 25 msec greater for the 3 c/d grating at (0,9). These differences in responses have been seen on another subject as well.

Clearly, in these two subjects, two sets of

active cells responding to the stimuli have been detected. It should be noted

(in Figures 1 and 2) that there is an

approximate mirror symmetry of the responses obtained by stimulating the two half-fields.

This symmetry is to be

expected if current flow in the brain is symmetric with respect to the vertical midline. (those 90 degrees out-of-phase)

The remaining responses also

exhibit symmetry.

Since we now know that we can detect activity from at least two sets of cells, we can explain the phase variability seen within the isophasic regions in Figure 2.

Since the detector

is seeing activity from two sets of cells responding 90 degrees out-of-phase, the resultant phase and amplitude are determined by superposition of the fields from these two populations.

If the populations are far enough apart, then

one of them will predominate and the two populations can be studied independently. Figure 5 shows field maps from another subject

(PSR).

His

right visual field map is similar to that of Figure 1 and 2. It contains two main regions with phases 180 degrees apart and other responses with phases indicating other sources. However, responses to left field stimuli show no discernible pattern even though all responses for this subject are based on three one-minute averages.

This difference between the

left and right field responses suggests a large asymmetry.in the geometry of the two hemispheres of this subject.

440 increases monotonically with the spatial frequency of the pattern.

The computed latencies of these responses correlate

highly with & behavioral measure (simple reaction time) to grating presentation.

Similar data have been obtained from

several subjects. Previous full field mapping studies with this subject showed an abrupt change in phase of 90 degrees as the probe was moved up the midline (Figure 4). This was verified with four different spatial frequencies. A similar shift is seen in the more complete full field map in Figure 2. Placing the probe at (0,9) and measuring the phase as a function of spatial and temporal frequencies results in the plots in Figure 3b. The phase versus reversal rate plots obtained for 3, 5 and 8 c/d gratings show an increase in latency with spatial frequency. The latency of the 5 and 8 c/d gratings are statistically the same at both positions (0,9) and (0,4) while they were different for the 1 and 3 c/d gratings

Ol i 2

i

4

i

6

I

8

l_ 10

DISTANCE ABOVE INION (em) Figure 4. a) Amplitude and b) phase of responses for 4 different spatial frequencies at several positions along the midline of subject DB.

441 The responses from this subject are unlike anything seen previously. Measuring near (-3,5), the slopes for 5 different grating stimuli were all about equal, i.e., corresponding to a latency of about 125 usees (Figure 6).

67»

V7

>155 ;

IOr

LVF

FF

RVF

93» 132» -148 >156 -204 -171 N33 >260 : '224 v / % X 4 8 1 3 0 35 123- 15'-152 /I55 96' M 0 » 2 7 2 2 7 l 2 2 3 3 2 2 « 3 7 I 2 9 " « 3 7 7 6 ' ' l 1 3 > £ 6 6 4 3 3 2 9 8 1 «• £ s f 20 -14 92* -!09-446 -417-294 S 4 • 2 99» 63- I44\/I29 ,179 2 6 6 3 5 4 0 1 7 3 4 7 3 * 3 7 4 I 174"* l73lUo5 -367 -164 : 2o j!02 52^ 97» 199^/452 740s 530«. 290^436-507 ¡699*4524» 200y444 -130 fo. •o i 1* N il »74 36x 2l-y60 447« 383» 310s I38| ,91 1463-396-299» 201» 114, - 2 • I I I 1 1 1 1 I ' ' ' 1 1 1

1 8-

0

-6

-4

-2

-6

-4 -2 0 2 4 distance from mldlln« (cm)

-6

6

-4

-2

0

evoked field' fenrtotesla

reversal rate • 13Hz

observer' PS R

spatial frequency - 5 c/d

2

4

6

contrast • 3 3 % luminance • 52 cd/m 1

Figure 5. Visual responses recorded at various points about the scalp of subject PSR for stimulation of the left visual field (on right), right visual field (on left) and full field (center). Dashed and dotted lines enclose regions of approximately the same phase. Short lines indicate response phases; thin lines indicate responses at or near the noise level. The stimuli used here were square wave gratings at 100% con2

trast and an average luminance of 66 cd/m . But the use of square wave gratings of high contrast cannot explain the difference in the responses. In another subject we obtained smooth variations in latency with both sine and square wave stimuli and with high and low contrast. Furthermore, sample data were collected with this subject (PSR) using 33% 2 contrast sine wave gratings at an average luminance of 52 cd/m

and

obtained behavior similar to that obtained with the square

442 wave gratings. At position (2.5,-3) responses were obtained using the sine wave gratings and these did shbw the monotonic increase in slope with spatial frequency. Thus, we also detected two sets of cells responding to the stimuli in subject PSR. One set is similar to a set seen in other subjects. The other set is not, although preliminary data in a new subject suggests that some of his responses

Fiqure 6. Phase of response versus stimulus reversal rate for stimuli at various spatial frequencies recorded at a R a t i o n 3 cm above and 3 cm to the left of the m i o n of subject PSR.

443 are similar.

Conclusions Because of the high spatial resolution possible with neuromagnetic techniques, and because of anatomical differences across subjects, we seem to have detected at least three sets of cells that respond uniquely to contrast reversal gratings. Of course, separate sets of cells are expected bn the basis of extensive work on animals (9). But at this stage it is difficult to compare our responses with the single cell measurements obtained from animals. Evoked potentials from area 17 of the exposed cat cortex show increases in latency with spatial frequency (10). While this is suggestive, it is not sufficient for us to claim that our responses showing similar increases are from area 17 in humans. Although our results pose many more questions than they answer, such as identifying the anatomical regions involved, they confirm our original proposal that it is possible to discretely detect functionally different regions of the visual cortex in man. The remaining questions concerning the anatomical regions associated with our data can only be answered by further study.

References 1. 2.

Williamson, S.J., Kaufman, L.: J. Magn. & Magn. in press. Okada, Y. , Kaufman, L., Brenner, D., Williamson, Superconducting Quantum Interference Devices and Applications, H.D. Hahlbohm and H. Lubbig, Eds. de Gruyter, Berlin, 1980) to be published.

Mat.,

3.

Kaufman, L., Brenner D., Okada, Y. , Williamson, S.J.:

S.J.: Their (Walter

444 in preparation. 4. 5. 6.

Goff, W.R., Williamson, P.D., Van Gilder, J.C., Allison, T., Fisher, T.C.: Progress in Clinical Neurophysiology W. Desmedt, S. Karger, Basel, 126-145 (1980) . Brenner, D. Lipton, J. Kaufman, L., Williamson, S.J.: Science 199, 81-83 (1978) . Aittoniemi, K., Hari, R. Kuusela, M.-L., Katila, T., Varpula, T.: Proc. of the Third National Meeting in Biophysics and Medical Engineering in Finland, Ch. A7 (ISBN 951-763-107-3, Lappeenranta, 1979).

7.

Stensaas, S.S., Eddington, D.K., Dobelle, W.H.: Neurosurg. 4£, 747-754 (1974) .

8.

Williamson, S.J., Kaufman, L., Brenner, D.,: Vis. Res. 18, 107-110 (1978) . Van Essen, D.C.: Ann. Rev. Neurosci. 2, 227-263 (1979).

9.

J.

10. Snyder, A.: unpublished Thesis, Rockefeller University, New York (1977) .

Reprint from: Biomagnetism Editors: S. N.Erné, H.-D. Hahlbohm, H.Lübbig © Walter de Gruyter & Co., Berlin • New York 1981 - Printed in Germany

APPLICATION OF A SQUID TO MEASUREMENT OF SOMATICALLY EVOKED FIELDS:

TRANSIENT RESPONSES TO ELECTRICAL STIMULATION OF THE

MEDIAN NERVE Y.C. Okada, L. Kaufman, D. Brenner and S.J. Williamson Neuromagnetism Laboratory, New York University 6 Washington Place, New York, New York, 10003, U.S.A.

Introduction In a previous investigation of somatic evoked fields (SEFs) (1) , the little finger of one hand was stimulated by a periodic transcutaneous electrical stimulus to measure steady-state magnetic fields produced by cells in the primary projection area of the finger. The detectable fields normal to the skull were found to be well-localized within a small region of the somatosensory area, contralateral to the side of stimulation. Furthermore, the polarity of the fields was found to reverse when they were measured at different positions.along the central fissure, suggesting that the current source of the SEFs can be localized within the somatosensory area. In a similar experiment carried out with the thumb, the iso-amplitude contours of thè SEFs similar to that for the little finger were observed, but they were about 2 cm lower along the somatosensory area, in agreement with the classic map of somatotopic projection (2) in man. The above results showed that the evoked fields generated by a current source or sources can be measured and the source can be localized. In the present study, the transient SEFs as opposèd to steadystate SEFs were measured over the somatosensory area in order to study the time course of responses to stimulation of a somatosensory nerve and possible bhanges in polarity of the components of the transient SEFs over the somatosensory area. The shifts in polarity across the scalp should yield

© 1981 Walter de Gruyter & Co., Berlin - New York Biomagnetism

446 information about the location of the current source(s) generating fields with varying latencies. Also, transient SEFs could be compared with somatic evoked potentials (SEPs) to examine the usefulness of SEFs and to see if the two types of data when considered together yield more insight about the nature of the underlying neural process. The transient SEFs were obtained by electrically stimulating either the left or right median nerve and recording from either ipsilateral or contralateral hemisphere. The detection system consisted of a second-order gradiometer coupled to a superconducting quantum interference device (SQUID). The gradiometer is well-balanced to reject uniform fields and fields with uniform gradients and to detect local spatial variations in the field in magnetically unshielded laboratory environments (3).

Method Stimulation: A periodic 1.9 Hz train of electrical pulses, 1 msec in duration and 4 mA in amplitude, was generated by a Grass stimulator and applied to the left or to the right wrist with a pair of circular, brass electrodes (2.5 cm in diameter) placed at the base of the thumb and on the other side of the wrist. Stimulation of the median nerve was verifield by the subject's reported sensation along the thumb and the first three fingers but not along the little finger of the stimulated hand. The current amplitude was near threshold for thumb twitch. To measure the SEFs, the bottom, pick-up coil of the gradiometer was placed 1 cm away from the scalp with the axis of the gradiometer (2.3 cm in diameter) normal to the surface of the scalp. The SQUID output was amplified, applied to a comb-filter to reject 60 Hz and its harmonics, band-passed between 1 and 100 Hz and then was processed by a PDP 11/34

447 computer to recover the average transient SEFs. Each session consisted of 2-minute long trials during each of which one median nerve was electrically stimulated while the subject lay still on a bed. He was given no extraneous task to perform. The data presented here are from one of four subjects. All were young male adults. Their results are essentially the same as those presented here.

Results and Discussion Transient SEFs recorded from the somatosensory region of the right hemisphere during the stimulation of the left wrist are shown in Fig. 1. Each record is an arithmetic average of responses to 480 stimulus pulses. The location of the pickup coil for each record is shown in the right: The two numbers associated with each record represent the distances in cm above the auditory meatus and behind the nasion respectively. The path along which the pick-up coil was moved from one of its positions to the others was determined by first finding the locations at various heights from the ear canal that produced maximum responses as the pick-up coil was moved transversally. Transient SEFs along the transverse paths showed no obvious polarity reversals. The resulting track seems to run along the posterior side of the Rolandic fissure, according to its classic projection onto the skull (4). The first notable feature is the orderliness of the data. As the pick-up coil was moved in small steps along the fissure, the waveform of transient SEFs changed gradually until the probe was near the null point 12 cm above the meatus and 16.5 cm posterior to the nasion, then its polarity reversed rather abruptly within a span of 2 cm, and as the pick-up coil was moved farther away from the null point the waveform again changed gradually, exhibiting a mirror-image symmetry about the null location. The data demonstrate that highly reliable

448 i—i—i

r

T I M E

(msec)

Fig. 1. The transient SEFs recorded from the right hemisphere of subject CS during the stimulation of the left median nerve. The pair of numbers at the right of each record indicate the position of the pick-up coil in the area above the right and back of the nasion.

transient SEFs can be measured with a second-order gradiometer. The waveforms of the transient SEFs consist of several major components.

The first has two small components with latencies

449 of 25 and 45 msec. Their latencies are highly reproducible across the various records in the figure and also across replications. The second major component has a peak latency between 75 and 125 msec. The peak latency of the third component is between 175 and 250 msec. The polarities of the major components reverse with longitudinal changes in the position of the pick-up coil, in agreement with the result described earlier for steady-state SEFs. The polarity of the first major component is positive below the null location and negative above it; thus, the fields with latencies between 25 and 45 msec emerge from the head below the null point and re-enter above. Similarly, it can be seen that the second component exits above the null point and returns below it, while the third component emerges from below the null point and returns at locations above. The polarity reversal of the major components at the same location suggests that they were all generated by a current source or sources located beneath the null location. According to the classic map of somatotopic representation, the null point is located near the primary projection area of the median nerve (2). Thus, the current sources in the small region seem to have been isolated with thermeasurement of evoked fields. The extent of SEFs seems to depend on size of the cortical projection of the somatosensory fibers. The amplitudes of the various components diminish noticeably over a distance of a few centimeters, but are more widespread than when the little finger was stimulated in the experiment described in the introduction. This discrepancy may be explained by the fact that in the present study the median nerve was stimulated at the wrist. Since the nerve innervates a large area of the hand and wrist, in addition to parts of the forearm, one would expect the fields generated in the cortex to be stronger and more widely spread over the somatosensory area

450 than the fields generated by stimulation of the branch of the ulnar nerve innervating the little finger. Thus, the transient SEFs in Fig. 1. may be thought of as due to multiple current sources in the cortex.activated by various branches of the median nerve. On the basis of the above characterization of the current sources, one would expect that the current sources in the two hemispheres to be symmetrically oriented and hence the polarity of the field from one hemisphere to be opposite that of the field from the other hemisphere. The transient SEFs from the left hemisphere of our subject, shown in Fig. 2b, demonstrate this expectation to be approximately correct. The records were obtained at locations between 8 and 17 cm above the left meatus with spacing of 1 cm. As in Fig. 1., the SEFs exhibit mirror symmetry about the null location which is a half centimeter above the corresponding null point in the right hemisphere (12.5 cm above the left meatus). But, the polarities of the major components are opposite in the two hemispheres at the homologous locations. There are, however, some differences in the waveforms of the transient SEFs obtained from the two hemispheres. One salient difference is in the first major component, that is, in the complex occurring between 25 and 45 msec after stimulation. The component from the right hemisphere (Fig. 1) has two sharply defined peaks while that from the left hemisphere (Fig. 2b) has one prominent peak and another less obvious peak. The other major components also differ in their shapes somewhat across the hemispheres. The differences in the waveform might be due to differences in the amount of contribution of the multiple sources to the overall shape of the SEFs and suggest the possibility of uncovering the multiple sources with more analytical experimental procedures.

451 (a)

0 50100

200

(b)

300 400 500 T I M E

0 50100 (msec)

200

300 4 0 0

500

Fig. 2. The transient SEFs recorded from the left hemisphere of subject CS. (a) Responses obtained during the ipsilateral stimulation of the median nerve. (b) Responses obtained during the contralateral stimulation of the median nerve. The locations of the pick-up coil were between 8 and 17 cm above the meatus spaced 1 cm apart. The results so far have shown that the magnetic fields produced in a small region of the brain may be isolated and detected by the use of second-order gradiometers.

The useful-

ness of the field measurement technique would be bolstered if it could reject magnetic fields produced by volume currents from far-away sources —

for example, volume currents

452 originating from the primary projection area of a median nerve on one side of the head spreading through the brain mass, the subdural fluids, the skull and the scalp. In order to see if such fields can be indeed rejected by the present field detection system, SEFs were measured from the left hemisphere while stimulating the left wrist. The locations of the probe were identical to those used with contralateral stimulation: SEFs were recorded at each of 11 locations by first stimulating the contralateral wrist and then stimulating the ipsilateral wrist. The result is shown in Fig. 2a. Clearly, the ipsilateral response are near the noise level and are markedly different from the contralateral responses of Fig. 2b. However, one should note that a weak response was detected 17 cm above the meatus (the vertex is 18 cm above the meatus in this subject) and also a consistent pattern of ipsilateral responses between 8 and 11 cm above the meatus. The response at the higher position near the vertex is probably due to fields generated in the contralateral hemisphere as can be seen by comparing the polarity of the top record with the top record of Fig. 1. This spill-over is however quite small. The more interesting responses, perhaps, seen in the lower tracings of Fig. 2a are probably not due to volume currents because of the fact that some of the strong early contralateral components (e.g., the component with 25 msec latency) are not present and because of the limited region within which these responses are detected. Rather, these ipsilateral responses may be due either to signals traversing direct ipsilateral neural pathways or to transcollosal signals that activate neural tissue in the ipsilateral hemisphere. The SEFs recorded ipsilaterally then show that the fields generated by volume currents from the contralateral side of the head may be greatly attenuated. Further insight into the nature of the source of the SEFs can be obtained by comparing it with recordings made with a small

453 electrode placed directly on the pial surface of the exposed brain of a human subject. Goff, Williamson, Van Gilder, Allison and Fisher (5) reported that the pial response reverses in polarity when the recording electrode was moved from the post- to the pre-central gyrus, but not when the electrode was moved parallel to the post-central gyrus. Their result is orthogonal to ours. Figure 3 shows SEFs recorded from

0

100200 300400

T I M E ( msec )

Fig. 3. A comparison of transient SEFs and pial SEPs. The transient SEFs were recorded from 11 and 13 cm above the right ear of subject CS; they are the same as in Fig. 1. The pial SEPs were recorded by Goff et al. (5) from the preand post-central gyrus near the primary projection area of the left median nerve for patient # 2 •

454 locations 13 and 11 cm above the meatus on the right side of the head. These are the same records shown in Fig. 1. Also shown are pial SEPs recorded by Goff et al. (5). The SEPs were collected by stimulating the left median nerve with a train of 50 0 usee pulses, 3 mA above thumb-twitch threshold. The recordings were made from positions in the brain near where motor responses of the left hand, wrist and digits could be elicited by electrical stimulation. The conditions of sensory stimulation are comparable to ours since our lower stimulus intensity is offset by the longer pulse duration we employed. The patterns of polarity reversals in the SEPs and the SEFs immediately suggest that both measures were produced by the same current source or sources. This idea can be examined by comparing the waveforms and polarities of the SEPs and SEFs. The SEPs recorded from the post-central gyrus contain three major components whose latencies are comparable to those of the SEFs. The first component has two peaks with latencies between 25 and 50 msec, the second has peak latencies of about 75 msec and the third has peak latencies between 150 and 250 msec. In the SEPs recorded from the pre-central gyrus, these components may be seen with their polarities reversed and in addition a fourth component with peak latencies of about 300 msec can be seen. This last component is either highly attenuated or is absent altogether in the SEFs. Thus, there is a remarkable similarity of the pial SEP and the SEF waveforms, even though some differences exist. In addition to similarity of the waveform, the pattern of polarity reversals in the two types of measures Strongly suggests that there is a large amount of commonality between the underlying sources for the SEF and pial SEP. The polarity of the first component of the pial recording shifts from + to - as the electrode is moved posteriorly across the central sulcus. This would mean that current is flowing from anterior to posterior. The corresponding component of the SEFs, on the

455 other hand, indicate that the fields emerge from the head below the null point and re-enter above the null point. This polarity reversal with longitudinal displacement of the probe is precisely what one would predict from the direction of current flow indicated by the pial recordings. The remaining correlated components of the pial SEP and the SEF are in the same relation to each other. Since the pial electrode senses extracellular current flow near the primary projection areas, the above comparison indicates that dense extracellular currents, such as the movement of ions in the intercellular space along apical dendrites of pyramidal cells, may be responsible for the SEF. In the light of its theoretical importance, this finding must be examined more closely, perhaps with animal models. The results obtained in the present study have several theoretical and practical implications. The main result, demonstrating isolation and detection of the magnetic fields generated by a well-localized source or sources, implies that a small population of cells in the brain may be studied with the present non-invasive technique. The similarity of the SEF and the pial SEP furthermore implies that the cellular activity which can be studied with SEFs might be common to that revealed by evoked potentials recorded from the exposed surface of the human brain. The rejection of volume currents produced by far-away sources may be used to help localize the sources of various components of scalp-recorded SEPs and also evoked potentials obtained in other sense modalities. For example, part of the late components of the scalp-recorded SEPs with latencies greater than 100 msec might originate in the projection area of the nerve being stimulated, as Goff et al. (5) have identified using pial SEPs. This so-called somatic late response (SLR) is presumed by Goff and others to be ordinarily masked by the vertex potential in the scalp-recorded SEPs. The vertex potential, which probably is a composite of potentials produced by volume currents originating in

456 different parts of the brain, overlap in time with the SLR. Thus, it is difficult to isolate the two when the volume current smeared by the various media between the brain and the detecting electrode cannot be removed, as is the case with scalp-recorded SEPs. Since the SEF seems to circumvent the problem of diffused volume currents, it could complement the use of scalp-recorded SEPs.

References 1. 2.

Brenner, D., Lipton, J., Kaufman, L, Williamson, S.J: Science 19^, 81 (1978). Penfield, W., Rasmussen, T.: The cerebral cortex of man. Macmillan, New York (1950).

3.

Williamson, S.J., Kaufman, L. , Brenner, D.: Biomagnetism, in B.B. Schwartz and S. Foner, Eds., Superconductor Applications: SQUIDs and Machines. Plenum, New York (1977).

4.

Gray, H.: Anatomy, Descriptive and Surgical, printed as a revised American edition, Bonty Books, New York (1977). Goff, W.R., Williamson, P.R., Van Gilder, J.C., Allison, T., Fisher, T.C.: Neural origins of long latency evoked potentials recorded from the depth and cortical surface of the brain in man, in J.E. Desmedt, Ed., Clinical uses of cerebral, brain stem, and spinal somatosensory evoked potentials. Prog. Clin. Neurophysiol. Vol. 7, Karger, Basel, 1980.

5.

SENSITIVITY DISTRIBUTION IN MAGNETOENCEPHALOGRAPHY

J. Malmivuo Biomedical Engineering Laboratory, Department of Electrical Engineering, Tampere University of Technology SF-33101 Tampere 10, Finland

This paper describes the sensitivity distribution of such a MEG lead system, where the magnetometer coil axis is directed to the center of the skull. Fig. 1. The sensitivity distribution is calculated on the basis of the lead field theory. The calculation method is presented in detail elsewhere /l/.

Method The sensitivity distribution is calculated according to the lead field theory. In this method a hypothetical current is

Fig. 1.

MEG lead system where the magnetometer coil axis is directed to the center of the skull.

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

458 fed to the magnetometer coil. This current induces a current field into the medium (brain region) which is called the lead field. According to the reciprocity theorem the distribution of this lead field is identical with the sensitivity distribution of the magnetometer. In infinite medium the lead field is calculated as follows: -

J =

r2

7

2

^ h 2 + ( n + r 2 ) 2 ) ((1- ^-)K(k)-E(k)) 2

(1)

where h

= distance from the coil in the direction of the symmetry axis ri = coil radius r 2 = radial distance from the symmetry axis

k

=

4r i r 2 h2+(n+r2)

The equation is normalized by defining the reciprocal current and the conductivity as follows:

is-i dt

(2)

a = 1 It is possible to show /l/ that the direction and the magnitude distributions in a cylindrically symmetric case obey the following rules: Vln.Q.c.tlo n

The magnetometer sensitivity lines, the lead field current flow lines, which indicate the direction of the sensitivitv, are circles, concentric with the magnetometer axis. Mag nZtu-dz

The lead field current density, which equals the magnitude of

459 the sensitivity, is zero on the symmetry axis and increases as a function of the radial distance from the symmetry axis. The exact form of this relationship is discussed later. To obtain the cylindrical symmetry needed for accurate calculations, the skull is modelled with a spherical model according to Rush and Driscoll /2/. In this model the outer radii of the brain, the skull and the skin are 80 mm, 85 mm and 92 mm, respectively. If we are interested only in the geometric configuration of the lead field and its relative sensitivity distribution, the conductance of the brain need not be defined. The conductance of the skull and skin do not play a role in this cylindrically symmetric lead system calculation.

Results The lead field is calculated for a MEG lead, where the detector coil is at 150 mm distance from the center of the skull. The result is shown in Fig. 2, which presents the magnitude of the sensitivity (=lead field current density) J as a function of the radial distance from the symmetry axis r. The axial distance from the magnetometer coil h is a parameter. The contour of the spherical model brain region is presented with vertical lines. Note that the direction of the sensitivity (the lead field current flow line) is everywhere tangential to the symmetry axis. The characteristic features of the MEG lead system sensitivity distribution are as follows: 1) The maximum sensitivity is in the plane, normal to the symmetry axis and a few centimeters below the level where the axis coincides with the skull.

In this area

the maximum sensitivity locates in the region on and near the surface of the brain.

460

Fig. 2.

Sensitivity distribution of the MEG lead system. J = lead field current density (=sensitivity) r = radial distance from the symmetry axis h = distance in the direction of the symmetry axis.

2) The sensitivity is zero on the symmetry axis. 3) The sensitivity is always directed in the tangential direction. 4) If EEG electrodes are located on the symmetry axis of the MEG lead system, the EEG and the MEG sensitivities are normal to each other throughout the brain region.

Discussion The characteristic features of the MEG lead system were described.

When comparing these with those of EEG lead systems,

461 we may note, that MEG sees the electric activity of the brain from a very different point of view than EEG. This fact may help in the localisation of the centres of activities in the brain. The spatial sensitivity distribution of MEG is, however, so wide, that MEG alone does not seem to be any better in locating the origin of the brain activities, than the EEG.

References 1. J. Malmivuo: "On the detection of the magnetic heart vector - An application of the reciprocity". Acta Pol. Scan. EL 39, 1976. 2. S. Rush and D. Driscoll: "EEG electrode sensitivity - An application of reciprocity", IEEE Trans., BME-16, 1969, pp. 157-167.

MAGNETORETINOGRAM AND MAGNETO-OCULOGRAM IN MAN

K. Aittoniemi, T. Varpula

M.-L. Järvinen,

T. Katila,

R. Maniewski1, and

Department of Technical Physics, Helsinki University of nology, SF-02150 Espoo 15, Finland

Tech-

Introduction The magnetic fields produced by the retinal currents of human eyes were studied. Two different procedures were utilized: The stimulation of the retina with light gives rise to a fast change of the retinal current resulting in a measurable change in the magnetic field outside the eye. This signal is called the magnetoretinogram (MRG). Its first observation was reported recently (1,2). When the subject moves his eyes a DC change of the magnetic field can be observed. This phenomenon, termed magneto-oculogram (MOG), was reported previously (3) . In the first part of this paper the measurement and results of the MRG are presented. In the second part the MOG, its spatial distribution and modelling are discussed.

I Magnetoretinogram Recording of the electric counterpart of the MRG, the electro— retinogram (ERG) is a commonly used noninvasive clinical method to observe possible retina injuries. The ERG is detected either by skin electrodes placed near eyes or by contact lenses on the eye. The general form of the ERG when the light stimulus is a short (^1 ms) flash is shown in Fig. 1. ^Permanent address: Institute of Biocybernetics and Biomedical Engineering, KRN 55, Warsaw, Poland

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

464 Fig. 1. Electroretinogram and the nomenclature used for different deflections for a short light stimulus. Amplitude of the a-wave is about 50 yV for a strong flash measured by skin electrodes.

Method The experimental setup of the MRG measurements is shown schematically in Fig. 2. The component of the magnetic field approximately perpendicular to the skull was measured near to the eye with a first order differential SQUID magnetometer. The 2

area of flux detection of the magnetometer is 3 cm and its base length is 14 cm. The noise level of the measurement is about 20 fT//ifz. After amplification the signal was recorded on magnetic tape. A flash-light served as the source of light stimuli. The light pulses had a duration of about 1 ms and an 2 intensity of 80 pW/cm . The flash-light was located in a

DETECTOR

Fig. 2.

T

The experimental setup of MRG measurements.

465 double-wall y-metal shield in order to reject magnetic interferences. For the same purpose, the flash-light was triggered by another flash tnrough an optical fiber. The measurements were done in a magnetically quiet wooden cottage without magnetic shielding. Therefore, the main amplifiers, power sources, etc. were located at a distance of 10 m from the measurement site. The recorded real-time signal and trigger pulses were replayed later on in the laboratory. The signal-to-noise ratio was improved by time averaging with a minicomputer. Responses containing obvious artifacts were cancelled in averaging.

Results Figs. 3 and 4 present MRG results obtained from one subject. The bandwidth is 0.1 - 30 Hz. The time delay due to the filtering is 15 ms. The uppermost traces are ERGs for comparison. The magnetic signals, corresponding to the a-wave of the ERGs are marked by crosses (x). The maximum amplitude of the MRG signal is only about 0.1 pT, hence it is about three orders of magnitude smaller than, for instance, the magnetocardiogram. The large deflection in the signals between 100 and 200 ms is caused by eye motion. The MRG in Fig. 3 b is measured above the right eye and in Fig. 4 b above the left eye in an equivalent position. It should be noticed that the polarities of the magnetic signals are opposite in these two measurements. Tnis phenomenon can be explained by the symmetry of the currents in the head. These observations have been confirmed also from the results of other subjects. In Fig. 4 c, the result of the magnetic control measurement is snown in which the flash was covered but otherwise the setup was unchanged. The small deflections in the control signal at 150 ms are probably magnetic audioresponses of the brain} the subject could hear a click when the flash-light was triggered (4).

466

a)

ERG

b)

MRG

c)

100 T -

200 ms

300

0

100

200 T -

CONTROL

300

ms

Fig. 3. Simultaneous measurements of the ERG and MRG from a normal subject. a) A reference measurement of the ERG with skin electrodes from the rignt eye. The curve is an average of 14 samples.

Fig. 4 a) The ERG measured from the left eye. An average of 14 samples. b) The MRG measured above the left eye corresponding to the position of the measurement of Fig. 3 b . An average of 130 samples.

b) The MRG measured above the right eye. An average of 124 signals.

c) The magnetic control measurement when the flash-light is covered. An average of 127 samples.

Discussion In comparison with the electrical method the measurement of the MRG offers many advantages: There is no physical contact between the subject and the transducer, and the problems of contact potentials and reference point are avoided. Moreover, the

467 MRG obviously yields the ERG.

additional

information

to that given by

There are also serious difficulties involved in MRG measurements. The extremely weak signal is hardly detectable in noisy laboratory environments without proper magnetic shielding. The other biological activities in the body produce signals higher than the MRG. For example, the heart signal was occasionally I - 2 pT in magnitude in the real-time MRG signal. The heart signal can, however, be avoided by synchronizing the stimulus rate to the heart rate. In any case, a time-averaging technique is unavoidable which is not the case in corresponding electric measurements. At the present state of magnetometry it is also very difficult to perform a simultaneous multipositional mapping which can be done easily with skin electrodes. Although the MRG has probably no clinical use today, it is expected that in the future using more sensitive magnetometers and more efficient disturbance rejection the MRG may become a potential clinical method.

II Magneto-oculogram The electro-oculogram is usually measured by means of electrodes placed on the eyelids. At present, the EOG is not of great importance as a clinical tool. It is used mostly in research studies. Tne aim of the present MOG measurements is to map the distribution of the magnetic field produced by human eyes. Two normal subjects were studied in this work.

Method The setup of the MOG measurements is shown in Fig. 5. The magnetic field component was measured in frontal, sagittal and

468 transversal planes in the front, right and top side of the head respectively. The magnetometer employed and the measurement system was the same as described above. During the measurement the subject is lying and moves his eyes over an angle of 55° between ' two fixation points once in every two seconds. About ten up-down and right-left deflections were recorded in 35 positions of each measurement plane. The luminance of the experimental room was kept constant. The signal of the magnetometer was filtered from DC to 3 Hz and then plotted with a chart recorder. Fig. 5. Tne scheme of the MOG measurement on the frontal plane. The component of the field perpendicular to that plane is mapped. Corresponding mapping were performed also in two other planes. The recording system was situated at a distance of 10 m from the subject.

RECORDING SYSTEM

MAIN AM PL.

MAGNETOMETER-

CD

Results An example of the real-time MOG is shown in Fig. 6. The peakto-peak noise level was less than 0.2 pT within the measurement band. The noise due to the magnetometer was not, however, the limiting factor of the sensitivity. The magnetic signal of the heart exceeded 0.5 pT even in this narrow band in some measurement points, especially on the lower left hemisphere of the head. Therefore, about ten shifts from each position were

Fig. 6. The magneto-oculogram measured in the position shown in Fig. 5. The DC shift AB of the magnetic field results frgm the up-down deflections (55 )of the eyes.

&B w

469 manually

averaged.

In this way

the signals,

which

included

serious artifacts, could be disregarded. The measured distribution plane is shown in Fig. 7 a .

of the

AB amplitudes on the frontal

In this case the subject has moved

his eyes from right to left over an angle of with respect to the center point. measurement points in

Fig. 7 are

55° symmetrically

The distances 4 cm.

cates that the field is directed inwards.

between

the

The open circle indiThat is defined also

as a positive direction of the field. The diameters of the circles are proportional to the field strength. The maximum signal of about

4 pT

was found above the subject's forehead

between

the eyes. MEASURED

CALCULATED

0 2pT O

pos.

• negleft-right

4 cm

o Fig. 7 a. The measured distribution of the MOG field on the frontal plane for horizontal movement of the eyes.

O

O

O

o

Fig. 7 b. The calculated field distribution due to two equal current dipoles positioned horizontally in the eyes pointing to the left. The contribution of the volume currents has been ignored.

470 Discussion In the simplest model for the electric activity of the eye, a current dipole across the retina accounts for the standing potential between the anterior and posterior pole of the eye. In the retinogram measurement, the magnitude of the current dipole is changed by stimulation. In the oculogram measurement, the current dipole is rotated which gives rise to measurable electric and magnetic DC changes. It can be shown that the normal component of the field due to a current dipole in a homogeneous sphere is not contributed by the volume currents in the conductor. (This was first correctly shown by Grynszpan in his thesis (5)). These assumptions, of course, are not strictly valid for a current dipole in the eye. First of all, the eyes are not electrically homogeneous and they are surrounded by tissues of different electrical properties such as bone and blood. Also at positions far from the eyes the measured field component was not normal, although the error is small. In spite of these limitations, let us assume in the following that the eye can be electrically represented by a current dipole in the retina, and that the contribution of volume currents can be ignored. Thus, when the subject moves his eyes symmetrically from left to right, the distribution of the MOG on the frontal plane is equivalent to the magnetic field produced by two horizontal dipoles in the eyes. It can be calculated from the results in Fig. 7 a, that the value of this equivalent horizontal current dipole is 0.3 yAm. It is also approximately the value of the component of the standing-dipole pointing forwards in the eye of this subject. Fig. 7 b depicts the distribution of the field component due to these equivalent dipoles. On the frontal plane the calculated field agrees rather well with the measured one. The fit is not as good in all components of the field. Especially on the sagittal plane, close to the eye, the contribu-

471 tion

of the volume currents

lustrated) .

A detailed

seems to be considerable (not il-

mathematical

model

for the

needed to account for all features of the MOG field.

eyes is For elec-

tric potential such a model of the eye has been made (6). According to our results are widely distributed.

the magnetic fields

of ocular origin

Therefore, blinking and eye motion may

cause serious artifacts in magnetic brain studies. in

measurements

triggers

of

eye motion

evoked

As a first approximation magnetic

Especially,

the stimulation

or blinking the signals

difficult to differentiate disturbing

responses

of

from the phenomenon

often

which can be to be studied.

one can evaluate the magnitude of the

ocular

signal

using

the current dipole

model.

Summary The measurement of the magnetoretinogram (MRG) and magnetooculogram (MOG) was discussed. The temporal behaviour of the MRG was observed to be similar to the electroretinogram (ERG). The magnitude of the MRG is extremely small, only 0.1 pT measured close to the eye. The mapping results

of the magneto-oculograms

The results can be explained qualitatively rent dipole model.

The

ocular signal

were presented.

using a simple cur-

must be taken into

ac-

count as possible source of artifacts in magnetoencephalographic studies because its field is widely distributed.

References 1. Aittoniemi, K., Katila, T., Kuusela, M.-L., and Varpula, T.: Proceedings of the Annual Conference of the Finnish Physical Society, abstract 8:13, Jyvaskyla, Finland (1979).

472 2. Aittoniemi, K., Katila, T., Kuusela, M.-L., Varpula, T.: Proceedings of the XII International Conference on Medical and Biological Engineering,Ch 94.6,Jerusalem, Israel (1979). 3. Karp, P.J., Katila, T., MMkipSa, P., and Saar, P.: Proceedings of the XI International Conference on Medical and Biological Engineering, p. 504, Ottawa, Canada (1976). 4. Aittoniemi, K., Hari, R., JSrvinen, M.-L., Katila, T., Varpula, T.: Localization of neural generators underlying auditory evoked magnetic fields of the human brain. In these proceedings. 5. Grynszpan, F.: Relationship between the surface electromagnetic fields and the electrical activity of the heart, (dissertation), University of Pennsylvania, USA (1971). 6. Doslak, M.J., Plonsey, R. , Thomas, C.W.: IEEE Transactions on Biomedical Eng., BME-27, 2, 88 (1980).

SUSCEPTOMETRY Magnetopneumography

PRACTICAL MAGNETOPNEUMOGRAPHY USING FLUXGATE MAGNETOMETERS

12 3 1 1 K. Aittoniemi ' , K. Kalliomäki , T. Katila , and T. Varpula ^Department of Technical Physics, Helsinki University of Technology, 02150 Espoo 15, 2

Outokumpu Oy, 0220 0 Espoo 20, "^University of Oulu, 90570 Oulu 57, FINLAND

Introduction Magnetopneumography

offers a suitable and non-invasive

method

to detect occupational dusts in the lungs. Since Cohen in 1973 reported measurements of magnetic contamination in human lungs, several groups have made investigations in this 12 3 field ' ' • For example in Finland more than 300 subjects have been measured since 1975, mainly with a SQUID magnetometer. This instrument utilizes cryogenic temperatures, which presently are obtained using liquid helium. However cryoliquids are inconvenient in routine work and a well trained personel is needed for the operation. In this paper we describe a computer controlled instrumentation for magnetopneumographic measurements based on the use of fluxgate magnetometers. More than 100 subjects ( mild steel and stainless steel welders, stainless steel grinders, workers of steel works, and trade school students) have been measured with this instrument. The determination measurements nately

of the magnetization

is called the magnetic inverse problem.

this problem

generally

has no

paper presents also one possible' way and

with the aid of

describe the

to

Unfortu-

unique solution. determine

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

This

the amount

distribution of the lung contamination

magnetic measurements.

field

from

476 Theory The magnetic detection

of contaminating particles in the lungs

is usually based on remanent magnetization The lungs are first magnetized netic field

of the contaminant.

in an external homogeneous mag-

and then the remanent field

caused by the magnet-

ized particles is measured with a sensitive magnetometer. Outside field field.

a sphere

enclosing

all sources

the multipole expansion The field

of a static magnetic

is a unique description

is a weighted sum of

dipolar,

of the

quadrupolar,

octupolar, etc. fields: 1

00

n H(r)= — (r )H e (r-r )+Bnm —o (r ) H° —(r-r LI LJ (A ' —mn —o )}, ' 4.ir n=o m=o nm —o —mn — —o

(1) v

where A ^ O ^ ) and

B ^ t r ^ ) are multipole strengths,

even and odd unit field functions,

(r) are

and r Q is the origin of the

expansion. The z-components of the unit fields can be expressed with

the aid of

associated

Legendre

functions (in spherical

coordinates) as follows: (2) H^°z(r,9,4>) = (n-m+1) r~ n ~ 2

(m) P™ +1 (cos8) .

These functions form a complete and orthogonal set on a sphere. Hence the multipole strengths can be uniquely determined by measuring only one component of the magnetic field. When measuring the field in one or two planes the unit field functions (2) are no more orthogonal. Nevertheless they are independent so that the multipole coefficients A n m ' B n m c a n b e uniquely determined even in this case. After magnetization in an external field, H Q e z , the contaminant particles

show

a

magnetizing field.

net magnetization

in the direction

When the magnetizing field

of the

and the density

of the contaminant particles p_, (r) are small the magnetization .4 can be expressed as follows:

477 (3) M(r) = a p c (r) H Q e z where

ez

is a unit vector along the z-axis

bration factor contaminant measured

particles.

for

netization

depending on

is

the

This

coefficient

different types orientated

magnetic

and

must be

of contaminants.

along

the

a is a cali-

properties

z-axis

of the

separately

When the magthe

multipole

strengths are (4) A nm = aH B nm

e m (n-m) I (n+m-1) /

sin

(m) P^_1(cos6) r n _ 1 p c (r) d 3 r,

where eo =1 and em =2 for m^O. In addition Ann =B nn =Bno =0. Thus for various poles, dipole (n=l), quadrupole (n=2), etc. there exist 2n-l coefficients to be calculated. The dipolar strength can be reduced to (5) A 1 0 = aH 0

P c (£) d 3 r

Hence the dipole moment is proportional to the total amount of the lung contamination. The other multipole moments describe the distribution of the contamination. Theoretically in contrast to the other coefficients the dipole strength does not depend on the origin of the expansion.

Evaluation of the dipole moment The coordinate frame and measuring grid used in the multipole calculations is shown in Fig. 1. First the multipole model was tested mathematically using a dipole approximation to the measured field. The calculated dipole moments are shown as a function of the origin of the multipole expansion in Fig. 2. Both the dipolar model and the multipole model up to the octupole term have been used. It is to be noted that the dipole moment given by the multipole fitting is in practice correct, when the error in coordinates is less than 5 cm. The dipole model gives

478 poor results especially in the z-direction.

ANTERIOR

POSTERIOR

SIDE

SIDE

Fig. 1. The coordinate system and measurement points used in multipole calculations.The dimensions are given in millimeters.

Az/cm Ax/cm Fig. 2. The calculated dipole moments when the origin of the multipole fitting is shifted in z- and x-directions. The curves a) show the result of dipole model and the curves c) the result of multipole model.

479 The multipole model is a suitable method to process the data of the lung contamination measurements. A set of moments describing the magnetization is obtained: the moment of the lowest order (the dipole strength) gives the total amount of the contamination and the higher order coefficients depict its distribution.

Measurements The magnetic fields to be measured in magnetopneumography are weak DC fields. The natural diamagnetism of human body in the magnetic field of the earth produces a weak signal,whose amplitude is 20 — 50 pT. The ferrimagnetic dust particles in the lungs give a response typically 0.1 - 100 nT on the chest depending on the magnetic properties of the dust particles and on the strength of the magnetizing field. In addition, one has to consider the external magnetic noise, which is mainly due to urban civilization.In a magnetically quiet place in an ordinary business building the amplitude of the noise field is about 50 nT/\ZHT and that of the field gradient 30 0 pT/m/VHz- over a band of 0.1 to 1 Hz. Due to high noise level a differential magnetometer has to be used. To obtain the maximum sensitivity when using a fluxgate magnetometer, a first order gradiometer with a relatively long base length (10 - 15 cm) has to be used. -4 The imbalance of the gradiometer should be less than 5-10 For noise rejection the second order gradiometer would be more efficient. However, if the base length is short, this device reduces also the measurement signal and makes its interpretation more complicated. It has been shown above that by measuring only one component of the magnetic vector field all magnetically measurable tion of the source is obtained.

informa-

This procedure is suitable for

automation and has good reproducibility.

480 In the following trolled a support

5

for magnetizing

a computer.

The

fluxgate

of

a realized computer

for clinical use

is shown

for a subject,

mercial

description

instrumentation

strumental setup bed

a

in

Fig. 3.

coils

of a cabin

instrument

and for

The in-

The device consists of detectors,

the

is based on

magnetometers

is given.

con-

situated

of a moving

electronics, the use of in

the

and of two com-

front

and

Fig. 3. The instrumental setup .

back sides of the subject to be studied. the gradiometers

used in the instruments

In the measurement band, from

0.1 to

The noise spectrum of is shown in

Fig. 4.

1 Hz, the noise is white

the noise level being 45 pT//Hz. The base length of the gradiometers is

14 cm and the distance between the posterior and an-

terior tranducers is about

33 cm.

During the measurement

the

481

CL 102 LU (/I O O a: UJ o •

\

10

< a: o 100

10

0,1

FREQUENCY

1000

(Hz)

Fig. 4. The noise spectrum of the fluxgate gradiometers. measurement'band is from DC to 80 Hz.

subject is moved forth

between

with the aid of the

detectors

an

electric motor

The

back and

along five transectional lines.

After magnetization the disappearance (the relaxation) of the remanent magnetic field is followed by performing sequential measurements in the middle line. To improve the signal-to-noise ratio several (10-15) measurements are done. Five minutes after the magnetization the magnetic field of the lungs is mapped in all five lines. In

Fig. 5 results of the relaxation measurement of a stainless

steel grinder are shown. Using the measured relaxation rate and the average field of mapping, the magnetic moment of the

lungs

is calculated and reduced to a time instant of one minute after

482

the magnetization. This moment is proportional to the total amount of contamination in the lungs. The proportionality factor depends on the magnetic properties of the dust particles and it is measured separately in vitro.

^

r

1,5 -

1,0-

0,5B (t) = B^I-q log101)

o,oL

1

10

t [min]

Fig. 5. The relaxation curve of a stainless steel grinder after magnetization in a homogeneous magnetic field of 40 kA/m. The dots indicate the average magnetic field of the lungs. Each point is an average of two separate measurements. In this case the magnetic moment was 15 yAm at one minute after the magnetization. The relaxation rate parameter a = 0.41 corresponding to a decay of 41 percent between one and ten minutes after the magnetization. It turns out that the magnetic field in the vicinity of the lungs is multipolar. A pure dipole model gives a poor description of the measured field. This can be seen in Fig. 6, which shows the measured and calculated field values of a foundry worker in one transectional line on the posterior side of the lungs.

483 The instrument was placed in the ground floor of an office building. There the sensitivity of the measuring system was limited by the noise of the magnetometers. The minimum de2 tectable magnetic moment was about 1.5 uAm corresponding to 0.5 mg

of magnetite

been used.

when a magnetizing

The reproducebility

field of 40 kA/m

has

of the measurement results was

5 % and the measuring time fifteen minutes per subject.

Conclusion An automatic

system suitable

dust in the lungs that

for routine mapping

has been developed.

magnetopneumography

is

a valuable

hygienic conditions in metal industry.

of magnetic

Field tests have shown tool

when studying

The measurement results

are reliable enough e.g. for follow-up studies.

20

15

H C N

CD

10

5

0 RIGHT

-10

-5

0 x / c m

+S

»10 L E F T

Fig. 6. The remanent magnetic field of a foundry worker. The black dots are measured field values and the continuous line shows the result of the multipole fit. The broken line is the result of the dipole fit. The2magnetizing field was 12 kA/m. The dipolar moment is 0.30 mAm .

484 References 1. Cohen, D.: Science 180, 745 (1973). 2. Kalliomäki, P.-L., Karp, P., Katila, T., Mäkipää, P., Saar, P., and Tossavainen, A.: Scand. J. Work Environ, and Health 4, 232-238 (1976) . 3. Freedman, A. P., Robinson, S. E., and Johnston, R. F.: Amer. Rev. resp. dis. 117, 233 (1978). 4. K. Aittoniemi, P. Karp. T. Katila, M.-L. Kuusela,and T. Varpula: in Proceedings of the Third National Meeting on Biophysics and Medical Engineering in Finland, Ch. A2 (ISBN 951-763-107-3, Technical University of Lappeenranta 1979). 5. LUNGCO, instrument for magnetopneumographic (Outokumpu Oy, 02200 Espoo 20, Finland).

measurements,

6. Kalliomäki, P.-L., Korhonen, 0. , Mattson, T., Vaaranen, V., Kalliomäki, K., and Koponen, M.: Proc.of XIX inter. Congress on Occupational Health, Dubrovnik Yugoslavia,1978(in press).

MAGNETIC FIELD MEASUREMENTS ON LUNGS: A COMPARISON BETWEEN TWO METHODS

G. Stroink, D. Dahn Department of Physics, Dalhousie University Halifax, Nova Scotia, Canada B3H 3J5

Introduction Two distinct methods are now in use to detect the magnetic signals generated by magnetic particles in the lungs of human subjects (1,2,3). In the first method, the subject stands between large Helmholtz coils which magnetise the magnetic particles in the lungs. The chest is then scanned with a probe containing a flux transformer connected to a SQUID based detection system. The second method uses a handheld magnet to magnetise the subject. Only one relatively small area is magnetised, after which this area is scanned with the flux transformer (3). In order to understand both methods better, we wrote a set of computer programs and calculated the effect of changing some parameters, such as chest dimension, distance of the probe to the chest, on the signals measured. Particular attention was paid to the response of a second order gradiometer.

The Lung Model As an approximation the lungs are represented by two ellipsoids, centred at (x,y,z) = (±7,0,0) cm, with semi-axes of 5, 17 and 9 cm in the x, y and z directions respectively. The z-axis is perpendicular to the chest, pointing outwards; the y-axis points from toe to head. All of the ellipsoids for

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

486 which y < -4 cm are removed to account for the way the lungs end at the diaphragm.

The dimensions of the ellipsoids were

chosen to give them, as closely as possible, the same size and shape as the standard lung used by Cohen et al (4) in their computer simulation of magnetic fields of lungs.

In the

discussion below we will concentrate on the field profiles obtained in the plane perpendicular to y and containing the centres of the ellipsoids.

In that plane the lung approaches

the skin of the back to about 2 cm.

The points where the

lungs most closely approach the skin are labelled (±7,0,-11) cm.

Uniform Magnetisation The magnetic field of a pair of lungs, uniformly magnetised in the z-direction, was calculated by adding the fields of dipoles with a moment of 10"3 A-m 2 , located on a 1 cm cubic lattice within the ellipsoids. So M = 103 A/m in the zdirection. For the back, the field profiles so obtained agree to within 5% with the results obtained by Cohen, who used a more realistic shape of the lung.

Local Magnetisation The field of our handheld electromagnet, with an iron core of crossection 3.8 x 3.8 cm 2 , is about 78 mT at the face of the iron core; from there the field decreases to about 10 mT 7 cm away from this face.

The surfaces of constant field of

the magnetic core were approximated by hemispheres with their centres at the axis of the magnet.

We then used our ellipsoid

standard lung model to determine what section of those surfaces were within the lung. appropriate magnetisation.

Each section was given the

To calculate the magnetic field at

487 any location outside the lung we integrated numerically over the distribution of the remanent magnetisation.

Discussion of the Results Plots were obtained of the flux through our second order gradiometer probe (area 3.7 cm 2 , baseline 3.8 cm, lowest coil 1.5 cm from the bottom of the probe) when the lowest coil of this probe is scanned along a line given by y = 0, z = -12.5 cm.

It was found that magnetising the lung uniformly with

35 mT produces two peaks about three times larger than the peak obtained when the lung is magnetised locally with the face of the magnet, as described previously, held against the skin at (7,0,-11) cm. In both cases the signals do not change by more than 10% when, with a fixed distance between lung and probe, the size of the lung is changed in the x, y or z direction by a few centimetres. The distance between lung and probe is important, however, Our calculations on the response of a second order gradiometer to the uniformly magnetised standard lung shows that the response drops by a factor of 1.7 when the lowest coil is moved from (7,0,-12.5) cm to (7,0,-14) cm. The field itself, as measured with a single loop, drops by a factor of 1.4 over this distance. If the lung is locally magnetised, in the way described above, then the response of the gradiometer drops by a factor of 1.9 when moved over this distance of 1.5 cm. A more dramatic drop in response will be observed when the distance between lung and skin is varied. In that case the handheld magnet as well as the probe have to move further away from the lung. The response of the gradiometer drops by a factor of 2.7 when the distance between lung and skin is changed from the standard 2 cm to 3.5 cm.

488

Measurements with the second order gradiometer on a plaster lung model and on a welder, confirm the results described above. The results demonstrate the need for accurate positioning of the probe and for a tool to measure accurately the distance between lung and skin. Attempts to measure this distance with ultrasound (by E. R. Smith of this university) have failed.

References 1. 2. 3. 4.

Cohen, D.: Science 180, 745 (1973). Kalliomake, P. L., et. al.: Scand. J. Work Environ, and Health 4 , 232 (1976). Robinson, G. E., Freedman, A. P.: IEEE Conference, "Frontiers of Engineering in Health Care", October 1979. Cohen, D.: "Report of the Low Field Group", M.I.T., February 1978.

EVALUATION OF MAGNETOPNEUMOGRAPHY FOR ASSESSING THORACIC ACCUMULATION OF WELDING FUME PARTICULATE AND LUNG DUST CLEARANCE A.P. Freedman Division of Pulmonary Diseases Hahnemann Medical College and Hospital 230 North Broad Street Philadelphia, PA, USA S.E. Robinson New York University Institute of 550 First Avenue New York, NY, USA

Environmental Medicine

Introduction Magnetopneumography is a non-invasive technique for detecting the presence and location of retained dust particles in the thorax. This is accomplished by directly measuring the remanent magnetic field due to the ferrimagnetic fraction of the dust. Since its initial description by Cohen (1), magnetopneumography has been used to assess the occupationally acquired lung dust content of welders, foundry workers, asbestos workers, and coal miners (2-5). Similarly, the kinetics of dust clearance have been studied in man and animals without the need of a radioactive tracer (6,7), As a prelude to a prospective study of the health effects of welding fume accumulation, we used magnetopneumography to ascertain the burden of welding fume in a representative group of electric-arc welders. We also used an animal model to evaluate the technical limitations of magnetopneumography for measuring dust accumulation and clearance.

© 1981 Walter de Gruyter & Co., Berlin • New York Biomagnetism

490 Method Twenty-five steel arc welders engaged in ship building were studied. Their welding experience ranged from 2 to 40 years. Five machinists from the same shipyard, sixteen former asbestos insulators from the same geographic area, and twentyfour rural subjects served as control groups. All measurements were taken in the fasting state and after showering. Using the uniform-field method (2) thoracic magnetization normal to the frontal plane of the chest was performed for 15 seconds at a field of 52 milliTesla. This was generaged by a pair of coils of 35 cm radius. Measurements of remanent field were made with a second derivative SOUID gradiometer having resolutions to less than 1 pT. The field normal to the anterior chest was measured with the probe touching the skin at 3 5 locations on a 5 x 5 cm grid. No magnetic shielding was utilized. Environmental background fields at a distance from the chest were measured at each location to serve as a baseline. If average thoracic remanent fields were less than 200 picoTesla, the contribution of biologic background fields due to internal ion currents and the diamagnetic susceptability of body tissue were accounted for by repeating the 35 point scan after magnetizing with opposite polarity. The remanent field at each location is then equal to half the difference between the two measurements, as it is the only field component to reverse polarity. Confounding extra-thoracic magnetic sources (i.e. steel pins) were degaussed with a tape eraser prior to the measurements. Average thoracic remanent field was calculated by averaging all remanent fields of positive polarity - i.e. with the same polarity as th'e magnetizing field. Probe sensitivity for the thorax was found to be 5.4 ng magnetite/cm3/pT in phantom lung models with a uniform magnetite distribution.

491 In a separate study, clearance curves were measured in two guinea pigs after inhalation of a polydisperse magnetite aerosol.

Serial sets of measurements were made at 6 anterior

chest locations using the technique of localized-field magnetopneumography described by Robinson to increase accuracy and give higher spatial resolution (8).

Results and Discussion Averaged thoracic remanent fields in the 2 5 welders ranged from 63 pT to 22,000 pT with a mean of 3056 + 4867 pT.(Figure 1) Superficial magnetic erasure ruled out significant contribution from skin contaminant.

Machinists all had lower levels

with values ranging from 2 to 55 pT.

Averaged thoracic

remanent fields for the 16 asbestos insulators were 6.4 + 5.6 pT (mean ± SD) and those of the 24 rural subjects were 4.7 + 1.9 pT.

Estimated thoracic content of ferrimagnetic

material in the welders ranged from 2 to 77 7 mg.

These

values did not correlate well with either smoking history or current pulmonary function.

Remanent fields tended to be

higher in welders with radiologic evidence of hemosiderosis. Values correlated well with total years welding (p

0.01)

provided those workers no longer actively welding, and thus having a net clearance of dust, were excluded from these calculations.

Some welders with greater than expected

remanent fields had significant ferrimagnetic dust exposure from other occupations.

Visual assessment of the remanent

field map showed the expected central enhancement from lymphatic clearance patterns. (Figure 2) The clearance curves of the guinea pigs had long-term (alveolar) half times of 52 and 61 days, in close agreement with reported radiotracer studies (9-12).

More significantly,

492

10,000:

a w

io £

1,000-

UJ

H

zUl z