the relation between the electroencephalogram and … · muscular jerkings in convulsive states has...

THE RELATION BETWEEN THE ELECTROENCEPHALOGRAM ANDMUSCLE ACTION POTENTIALS IN CERTAIN

CONVULSIVE STATESBY

G. D. DAWSONThe David Lewis Epileptic Colony and Neurological Research Unit, National Hospital, Queen Square

(RECEIVED 25TH FEBRUARY, 1946)

Introduction

STUDY in man of the relation between the occur-rences in the electroencephalogram (EEG) and themuscular jerkings in convulsive states has beenlimited. Possibly this has been due to the difficultyof anticipating the time of onset of seizures suffi-ciently often, and with sufficient accuracy. In twotypes of case this difficulty may be overcome. Thefirst is that in which myoclonic jerkings occur overa period of one or two hours as a prodrome of amajor convulsive seizure and the second that inwhich a wave and spike discharge in the EEG lastsfor 24 to 48 hours and is accompanied by muscularjerkings. It is the purpose of this paper to describethe results of repeated examinations of patientswith these conditions.

Gibbs, Davis and Lennox (1935) showed that inthose patients with epilepsy in whom a dischargeof wave and spike form appeared in the EEG duringa seizure, minor muscular jerkings might occur atthe same rate as that of the repetition of the waveand spike complex. They also reported that in theclonic stage of a major convulsive seizure thejerkings of the muscles occurred in time either withslow waves or groups of spikes in the EEG. Inneither case was fuller analysis of the temporalrelations of the EEG waves and the muscular move-ments described. Jasper and Andrews (1938) byrecording the EEG, muscle action potentials andmovements studied the relations between wave andspike discharges in the EEG and the muscularjerkings during minor epileptic seizures. Theyfound that when wave and spike discharges in theEEG were provoked by overbreathing, a burst ofmuscle action potentials was often associated witheach wave and spike complex. In some subjectsthis correspondence did not occur and from thisthey concluded that in certain circumstances thecontrol of the motor functions concerned might betaken over by subcortical centres. Recording ofthe EEG and muscle action potentials has beenused by Strauss and Landis (1938) to compare thefeatures of spontaneous convulsions and those pro-

duced by cardiazol. They used the method to timethe contractions of flexor and extensor musclegroups but they give few details of their results inthis connection. During the severe muscular jerk-ings which occur in myoclonic epilepsy Grinker,Serota and Stein (1938), who studied a group ofpatients with a hereditary form of this complaint,were able to show that a close correspondenceexisted during the jerkings between groups of spike-like discharges in the EEG and the muscle actionpotentials. They recorded the muscle action poten-tials only from the biceps muscle of the arm and nosimultaneous recordings from antagonistic musclegroups were reported. Their records would notallow accurate measurements of the time relationsof the individual spikes of a group and any par-ticular phase of the muscular jerkings. Confirma-tion of the findings of Grinker et al. is given byJasper (1941) and Gibbs and Gibbs (1941) whoalso describe groups of spike-like discharges in theEEG in time with the myoclonic jerkings, but inneither case are any facts added about the timerelations of the spikes and the jerks.The present investigations have been made on

two types of case with myoclonic symptoms duringepileptic states. The results of simultaneous re-cordings during myoclonic attacks of the EEG andmuscle action potentials in flexor and extensormuscle groups of the limbs will be described.

MaterialThe subjects studied have been selected from

amongst the 450 residents at the David LewisEpileptic Colony. In this number of habitualepileptics five patients have been found in whomsevere myoclonic jerkings are a prominent symptomand one patient in whom mild myoclonic twitchingin the limbs accompanies a discharge of wave andspike form in the EEG that may last up to 48 hours.Of the five patients with severe myoclonic jerkingsit has been possible to predict the onset of seizuresto within a few hours in two patients (337 and 407)and these have been studied repeatedly. The lengthof the seizure in the patient with the wave and

5

Protected by copyright.

on 19 August 2019 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.9.1.5 on 1 January 1946. Dow

nloaded from

http://jnnp.bmj.com/

spike discharge makes recording easy after theseizure has started and in this subject (377) alsorepeated examinations have been made.The two patients with severe myoclonic jerkings

(337 and 407) are strictly comparable as regardssymptoms with those described by Grinker et al.,except that they have no history suggesting a familialincidence of the complaint, and that their conditionis either stationary or only slowly progressive in itsseverity. These patients suffer from jerkings of themuscles which are sufficiently powerful, when theyoccur in the legs, to cause the subject to fall heavilyto the ground; or, when they occur in the arms, tocause the subject, if he is holding some quite heavyobject, to throw this over his head. In neither ofthe patients are the jerks, which are extremelybrief, associated with detectable changes in con-sciousness. Usually the jerks occur at progres-sively shorter intervals over a period of 1 to 2hours until finally one of them passes with littlepause into the tonic stage of a major convulsiveseizure. Occasionally after becoming more fre-quent the jerks occur less often and then cease.When this happens, if the patient goes to sleepbefore the jerks have disappeared, on his re-awakening they may reappear, increase in frequencyas before, and terminate in the usual major seizure.Any excitement or attempted exertion tends toaggravate the jerkings and may precipitate theterminal seizure. Therefore, if an adequate timefor examnination is to be gained, the subject mustrelax as completely as possible when the jerkingsbegin. Although the period between the onset ofthe jerks and the final seizure in this type of case israrely more than 2 hours, it has been possible withthe excellent co-operation of the subjects and theirfriends in the Colony, who helped the subject tothe laboratory immediately the jerks began, toexamine these two patients repeatedly for periodsof 1 to 1 hours before their major seizure.The third patient (377) who has been examined

repeatedly is one who, besides having major con-vulsive seizures, is subject to attacks of mildmuscular jerkings which last for a period of up to48 hours and are associated with a continuous waveand spike discharge in the EEG. At the beginningof the attack there is slight clouding of conscious-ness and the twitchings of his muscles are diffuseand irregular. After about twelve hours theclouding of consciousness disappears and the jerk-ings become synchronous in all parts of the bodyand regular at about three per second. If un-treated this state may continue for a further 24hours and then the jerkings and the wave and spikedischarge will cease spontaneously and suddenly.For several days before a spontaneous attack,attacks of jerking may be provoked by over-breathing. For several days after a spontaneousattack the same amount of overbreathing, or agreater anmount, will produce no outburst. Thejerkings of this patient have none of the violentnature of those occurring in the type of casedescribed above and do not monopolise the entire

voluntary power of the muscles concerned. Oncethe initial stage of diffuse twitching has passed offand the jerkings have become regular, the subjectis able to walk around, talk, and dress himself,though all these activities have the jerkings super-imposed on them.

MethodsRecords of both the EEG and muscle action poten-

tials have been made with inkwriting oscillographs andresistance-capacity coupled amplifiers. Four channelshave been available and the usual arrangement has beento record the EEG from the scalp with two of thesechannels and bipolar electrode placements over frontal,motor, and parietal regions of one side of the head; theother two channels being used to record from the flexorand extensor muscles in the thigh or upper arm of theopposite side. On other occasions the EEG has beenrecorded from the motor areas of the two sides and themuscle action potentials from the flexor or extensormuscles of the two sides in the upper arm or thigh.The patients have been examined whilst lying in thesupine position on a blanket covered bed. This isimportant if the friction artefacts are to be avoidedwhich occur during convulsive movements when aleather-covered couch is used.OSCILLOGRAPHS.-The oscillographs used in this work

are copies of those designed by Mr. A. M. Grass ofBoston, who kindly gave permission for them to bemade in this country as the shipping position at thetime this work was started made importation impossible.The writers have a performance strictly comparable withthat of the Grass type with the exception that 20 percent. of the maximum linear amplitude of deflection hasbeen sacrificed to allow a small increase in the naturalperiod of the system; this lies between 70 and 80 c/s,and is critically damped. The importance that isattached to the shape of waves that have componentsboth above and below the resonant frequency of therecording system must be limited. The reason for thisis that although the system may have a good responseto the higher frequency components of a complex wave-form, when these fall above the resonant frequency ofthe writer they may be shifted in phase in relation toany lower frequency components that fall below theresonant frequency. When such mutual phase shifts ofthe components of a complex wave occur the shape ofthe wave will be altered. Though the shape of a waveof the type mentioned may not be recorded accurately,differences in the shapes of waves will remain of signifi-cance with any one recording system. In spite of theserelative inaccuracies at the higher frequencies it has beennecessary to tolerate the inkwriting oscillographs ratherthan to use a more accurate recording system. Such asystem with cathode ray oscillographs, giving foursimultaneous records, has been available and has beenused to monitor and check the performance of thewriters in recording waves with periods of less than 10to 12 milliseconds, which is about their extreme upperlimit of accurate recording. Otherwise this secondrecording system has been of little value in these studieson account of the long intervals that may occur betweenthe myoclonic outbursts which make photographicrecording prohibitively costly. With the inkwritingoscillographs recording paper speeds of 4 and 8 cm.per second have been used during the periods whenjerkings were expected and 2 cm. per second at othertimes. Time scales are indicated in seconds at the topof the records, and also in some cases, at the fasterrecording speeds, by a 50 c/s trace or a 0-1 second markin the centre of the record. Calibration of amplificationis indicated in microvolts (,V.) or millivolts (mV.) atthe right-hand end of the records. The records through-out read from left to right.AMPLIFIERS.-The amplifiers used are of conventional

design and give the same amplification for all frequenciesfrom 1 to 100 c/s, with optional reduction of the

6 G. D. DAWSONP

rotected by copyright. on 19 A

ugust 2019 by guest.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.9.1.5 on 1 January 1946. D

ownloaded from


RELATION BETWEEN E.E.G. AND MUSCLE ACTION POTENTIALS337alAI

FIG. 1.-EEG records between seizures from a patient (337) with myoclonic epilepsy showing (a) rhythms at6 to 7 c/s unaffected by closing the eyes, and (b) groups of spikes and slow waves. In this and allfollowing records the time scale at the top indicates one-second intervals. The + and - signs atthe electrodes in all the head diagrams indicate that a potential difference such that the electrodemarked + is positive with respect to that marked -, will cause an upward deflection in the record.

response to the higher or lower frequencies. The inputconnections are symmetrical with respect to earth andlead to an input circuit of the type described by Toennies(1938). In later experiments this circuit has beenreplaced by one designed on the principle described byBuchthal and Nielsen (1936) and Offner (1937), whichhas advantages when followed by pushpull amplifierssuch as those used to drive the inkwriting oscillographs.With this circuit it is possible to record simultaneouslyfrom scalp and a limb without the electrocardiogramappearing in one or both records; it also prevents themuscle action potentials, which may be as large as 10or 20 mV., from appearing in the EEG record, evenwhen this is small and is being recorded with a highamplification.The standard amplifier setting used for recording the

EEG gives an equal response to all frequencies between2 c/s and 50 c/s. When the muscular jerkings are verysevere low frequency movement artefacts may become aserious nuisance and under these conditions the responseis kept constant from 50 c/s down to 20 c/s and thenattenuated by 30 per cent. at 10 c/s. The records takenwith these last constants cannot be used for direct com-parison of waveforms with those in records taken at thestandard setting on account of the mutual phase shiftsof the components of complex waves that occur; how-ever, changes of the shapes of waves that occur in any

one record remain significant. For recording themuscle action potentials amplifier settings are usedwhich give an equal response to all frequencies between20 c/s and 50 c/s. Below 20 c/s the response is attenu-ated by 30 per cent. at 10 c/s. Above 50 c/s the responseis increased by 10 per cent. at 80 c/s and then fallssteadily and is attenuated by 80 per cent. at 150 c/s.The rise of the response at 80 c/s has been allowed onaccount of the considerable extension of the response tohigher frequencies which it makes possible.ELECTRODES.-The electrodes used to pick up the

potential differences from the scalp have in all casesbeen similar to those described by Walter (1936). Theseconsist of a gauze-covered sponge pad, 1 cm. diameter,soaked in saline solution and supported on a chloride-coated silver tube. They have been held in place on thescalp by a rubber and whalebone cap. On account ofthe violence of the subject's movements during attackssmaller electrodes held in place by collodion might havebeen preferable. However, the time saved in applicationof the saline pad type of electrode has been of greatimportance and they have not in fact been found, whenproperly applied, to cause any significant artefacts.For recording the muscle action potentials large

surface plate electrodes, 2-5 by 3-5 cm. in size, makingcontact through Cambridge electrode paste and held inplace by rubber straps have been used. Two such

7

I %~Vqffn. 5.

I I I I I I a, i I c L-e- "



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


plates separated by 1-5 to 3 cm. have been applied toeach muscle group to be studied. These large contactshave been used in preference to needle electrodes fortwo reasons. First, the needle electrode was introducedby Adrian and Bronk (1929) because of the very re-stricted space from which it picks up potential differences.Since these studies are concerned with the behaviour ofmuscle groups as a whole this restricted pick-up isundesirable. Secondly, the risk of damage to themuscle, or breakage of the needle during the jerks, is sohigh that to use such an electrode has been consideredan unwarranted tax on the co-operativeness of thesubjects. The chief disadvantage of the big surfaceplate electrodes used is the large area from which theypick up potential differences. In practice the localiza-tion of pick up given by these electrodes has been foundto be satisfactory for this work. That is to say, whenflexor muscles in the thigh or upper arm are contractingno potential differences of significant size appear inplate leads on the extensor aspect of the limb at thelevels of amplification used in these studies. The sameapplies with regard to electrodes on the flexor musclesin the upper arm when the extensor muscles are con-tracting. In the thigh the muscle masses are less favour-ably disposed. When the quadriceps extensors arecontracting action potentials may appear in plates overthe hamstring muscles. These potentials are about10 per cent. of the size of the action potentials picked

up by lplates, with a similar separation between them,applied over the contracting muscles themselves. For-tunately the order in which the potentials have beenfound to appear has made it possible to exclude thisdegree of spread from being important.

Results

1. Myoclonic EpilepsyINTERSEIZURE RECORDS.-The two subjects who

have severe myoclonic jerkings as a prodrome totheir major convulsions show both similarities anddifferences in their resting EEG records. Bothsubjects show a bilateral discharge from frontal,parietal and occipital regions at 6 to 7 c/s whichis unaffected by opening or closing the eyes. Inone case (337) this discharge is rhythmical andrarely broken up by slower rhythms (Fig. 1 (a)).In the other case (407) the discharge is very irregularand is broken up by continuous slower dischargesof small amplitude at frequencies down to 4 c/s(Fig. 2 (a)). Both subjects show outbursts of highvoltage slow activity at 2 to 3 c/s, but these arelarger and more common in the second case (407).

407alA/h.

407A1E711.

FIG. 2.- -EEG records between seizures from a patient (407) with myoclonic epilepsy showing (a) greateramount of slow activity at 2 to 4 c/s than in the records of Fig. 1, and (b) effect of overbreathing in thiscase.

8 G. D. DAWSONP



j.com/

J Neurol N


ownloaded from


RELATION BETWEEN E.E.G. AND MUSCLE ACTION POTENTIALS

RI>

U

f-i

-J &.1

UW0'J

-o

I

I

I01(El

L

I

cncn

C-

on_

I_:1S

I

I

I

I

I

I

I

0

C.C

J7-

=5 o

m;0-o

0~

C-

0E )

C-,

-S 0

C-)S

*_C- 0'

pvi

r%

cn

Oi0

I-:

6

II

II

9

11.1I-

I

c

i-

2

7iiiI



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


G. D. DAWSON

)

ULi

8Ul)

06

Q)

10

g

vI

x-,I.-Li.

oZ!..

0 e

C)C) >

.0

0~L

C)

o. -E-

. o

05

00

C)

C:0

C-C

0

0)

0LL 0

to.o

.0

8 Q

. 0C s

0., c)

E) 0.

-C)C.0-0

.0

.E.. Lu'

C rO U

o

3 D )

COCO0 .00=

[.L 0

[.L 0

t EC

(5.s

(3

IIEII

I

AII

I

i

A!ii

r

i

FZ6



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


RELATION BETWEEN E.E.G. AND MUSCLE ACTION POTENTIALSThese slow waves are largest in the frontal regionsand may be accompanied by discharges of singleor multiple spikes (Figs. 1 (b) and 2 (a)). In thequiescent periods muscular jerkings do not accom-pany these spikes. In the first case during over-breathing no effect on the slow waves is seen, butin the second case the slow waves are greatlyincreased in number (Fig. 2 (b)). In neither case hasoverbreathing been found to affect the frequencywith which the groups of the spikes occur. Thedischarges present in these two cases are usuallysymmetrical on the two sides of the head.RECORDS DURING SEIZURES.-During the episodes

of jerking the EEG takes on an unusual and highlycharacteristic appearance. In this state briefgroups of spikelike discharges occur at a frequencyof 8 to 13 per second. The term spike will be usedto describe these waves since they appear sharp atslow speeds of recording, though at the higherspeeds used for measuring time relations they oftenlose much of their spikelike character. The spikesin the EEG may occur singly or in groups (Fig. 3(a) to (d)). Although the spikes are usually largerin frontal than in precentral leads no case of jerkingoccurring synchronously with the spikes in theEEG has been seen when they have not appearedin the precentral leads. The duration of the EEGspikes is variable from about 25 to 75 msec. andis related to the placing of the electrodes withrespect to the potential gradients on the scalp, andto the place of the spike in its group (Fig. 3). Overthe duration of one burst the shape of the spikeschanges in a way that tends to repeat itself in anyone subject. The first phase of the potential changeis surface negative and this is followed by a secondphase that is surface positive. Sometimes a smallsurface positive change precedes the surface nega-tive one (Fig. 4 (b) and (d)), but this is not constant.Towards the end of the burst the discharge appearsas a short duration surface positive wave alone, ora rectangular surface negative wave. This changeof shape may be due either to a movement of theorigin of the potential difference with respect to theelectrodes or to an actual change in the form ofthe potential difference. Which of these alterna-tives is the correct one cannot be decided in theabsence of fuller studies of the potential distribu-tions in the attacks. Following the spike burstthere is usually a " silent period " of 0-1 to 0 5 sec.during which electrical activity falls to a minimumbefore returning to the same level that it had beforethe spike burst (Figs. 3 (a) and (b), and 4 (a)).

Associated with these discharges there occur themassive muscular jerkings found in this type ofmyoclonic epilepsy. With the amplifications usedno muscle action potential or detectable muscularjerk has been found to accompany a single isolatedspike in the EEG. Two spikes are usually accom-panied by a minimal action potential in flexormuscles alone, whilst when a longer burst of spikesoccurs action potentials most commonly appearfirst in the flexor muscles, follow in the extensormuscles and recruit steadily up to the end of the

spike burst, then cease suddenly. A series ofrecords in Fig. 4 (b), (c), (d), and (e), shows a pro-gressive increase in the number of spikes in fourconsecutive bursts, with a corresponding increasein the number and size of the muscle action poten-tials. The length of the " silent period " or periodof " extinction " after the spike burst also increaseswith the number of spikes in these records. Thesections shown in Fig. 4 (b), (c), (d), and (e), areparts of a continuous record with intervals of8 seconds between them in which no other equallylarge paroxysmal disturbances occur. They suggesta rapid and progressive increase in the cerebralabnormality over a period of 30 seconds. Onceboth the flexor and extensor muscles are con-tracting their action potentials are, within theaccuracy of the measurements possible with thesemethods, apparently synchronous and without anysign of reciprocal innervation.

It was stated above that when a single isolatedspike occurs in the EEG it has not been seen to beaccompanied by a muscle action potential or clini-cally obvious muscular contraction. However, ifthe single spike follows within about 1-5 or 2seconds of a previous burst of spikes, apparentlysome of the facilitation that has been built upduring the burst remains, and a jerk and actionpotential, such as that indicated by the arrow inFig. 5, may occur in the flexor muscles alone. Foran extensor discharge to occur with a single spikeit seems to be necessary that the interval betweenthat spike and a preceding burst shall not be morethan 0-5 seconds. If a break in the spike dischargein the EEG occurs, when the discharge restartsthere has been a greater decay of facilitation forextensor than for flexor mechanisms. The recordsin Fig. 6 show that after an interval of 0 3 secondsthe facilitation for the flexors has fallen little, andthe muscle discharge restarts at the size at which itceased (Fig. 6 (a)), or reduced by 50 per cent.(Fig. 6 (b)). For the extensors, however, thefacilitation has apparently decayed to a greaterextent, as the first of the succeeding discharges ismuch smaller than the last of the preceding on,s.Some facilitation for the extensors remains, sincethe action potentials recruit more rapidly than inthe initial burst. Measurements of the delaybetween the beginning of the EEG disturbance andthe muscle action potential suggest a value for thisnot less than 15 and not more than 40 msec. Thesemeasurements, however, have to be accepted withcaution for several reasons. First it is only inrecords such as those of Figs. 3 (a) or 4 (c) or (d),that it is possible to decide where the EEG dis-turbances begin, and secondly it is difficult to besure that the particular EEG spike used as a refer-ence is associated with a particular muscle actionpotential. Higher recording rates help little, as itis found that when the waves are opened out bythis means they have no salient features from whichmore accurate measurements may be made.One point of interest in the records of the muscle

action potentials during the myoclonic jerkings"is

11P



j.com/

J Neurol N


ownloaded from


G. D. DAWSON

r

2u*4

n>

FIG. 5.-A single spike discharge in the EEG, marked by the arrow in the record, is accompanied by a muscledischarge if it follows within 2 seconds of a previous spike burst, cf. Fig. 4 (b) where a single spikenot following a previous burst is unaccompanied by a muscle discharge.

337e/C1173.EXT. ;4

LEFT THIGH.

FLEX.

a.1Uti.u,A""MUMnnwmnwnlmmnm-wnmrunwn_~~~~~~~.. . _................

C1170.

4mVT

-H@'df~~i'i-O1b. 0.02 SEC.

. _ _- ....

FIG. 6.-Following a break in the EEG discharge the muscle action potentials restart with the size of those inthe flexor groups reduced less than that of those in extensor groups.

their extreme simplicity and brevity, and theapparent absence of after-discharge. The methodof recording used might possibly introduce as an

artefact an apparent lengthening of the potentialdifferences; it is highly unlikely to shorten theirapparent duration. The recorded action poten-tials have neither more phases nor greater durationthan those evoked by eliciting a reflex jerk with a

tendon tap. It does not seem to be possible toproduce anything as brief as this by voluntarymovements of even the shortest duration. Thisseems to indicate some highly synchronized form ofcentrifugal discharge in the myoclonic state. Afterthis stage of the seizure, characterized by the inter-mittent episodes of jerking, has been in progressfor about one or one-and-a-half hours one of the

jerks passes, with a pause of usually not more than2 or 3 seconds, into the tonic stage of a majortonic-clonic convulsion (Fig. 7). It is notable thatalthough the tonic stage of the seizure is welldeveloped by the end of the second part of therecord, which is continuous with the first part,there is only a low voltage fast discharge in theEEG and the rate of this, 45 to 50 c/s, suggests thatit is of muscular rather than cerebral origin. Theabsence of the 20 to 25 c/s discharges usually asso-

ciated with the start of a Grand Mal seizure (Gibbs,Gibbs and Lennox, 1937) has been noted on severaloccasions in these subjects.

EFFECTS OF SENSORY STIMULATION.-In the periodof prodromal muscular jerkings before the majorseizure two stages may be distinguished clinically.

337 b/B/II.

12

r

-- -- -I



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



1)

t

0. < uflw a- 1A- _

13

O0

C >

o CCt

0

0 0C(

0

O.-C

2) =tw

*rCn

0)

C

o 0

0=

. W

O:E

C-

6

*Cd

0n

.t

01

.. C#d

00u C,vCd

D.0

Y

ed

i""Idr.0N

1.

.l<L



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


In the first stage, if a knee jerk is elicited the initial,normal response is followed by another jerk. Theinterval between the two jerks varies between 80and 180 msec. and the second jerk is usually con-fined to the muscles stimulated, but may occasion-ally occur in the antagonists (Fig. 8 (a) to (e)). Inthis stage no correlate of the jerk is found in theEEG. The action potentials show that the initialjerk, marked KJ in the records, is simple and brieflike that in a normal tendon reflex contraction.The action potential of the secondary jerk, markedby the arrow on the records, has a different charac-ter: it is multiphasic, increases in size slowly up toa maximum, then declines slowly and shows a con-siderable dispersion in time. The record showingthe shortest latency for the secondary actionpotential (Fig. 8 (d)) was made on a different dayand with different amplification from that showingthe longest latency (Fig. 8 (a)). The difference in

the amplifications used does not account for thedifference in the latencies as, although some of theearlier potentials in the secondary discharge inFig. 8 (a) may not be visible on account of the loweramplification used for that record, the dischargereaches its maximum 80 msec. later there than inthe record of Fig. 8 (d). It does not seem to bepossible to correlate the latency or the size of thesecondary discharge with the size of the initial jerkaction potential. The smaller of the two figuresfor the latency of the secondary jerk (80 msec.) issignificantly less than a normal reaction time;therefore, unless the reaction time is considerablyshortened in this condition, it is unlikely that thesecondary response is a voluntary one to a stimulusthat has assumed a startling nature. The secondaryresponse has many of the characters, in a greatlyexaggerated degree, of the rebound that occursnormally at the end of the inhibitory silent period

337./C/86.EXT.

LEFT THIGH. KJ +

FLEX -

a.WMNWM

t A A - -.A. A.-Aa..

337dIAI38. 28rtd..X^ # WD

b.VMVVVW MANWW

C.WWVWPANVW\MVWWAMNW

1r.m.I

s4\Jw4vvY4~frA~~ I

_EC0.02 SEC.

FIG. 8.-The records show the rebound, marked by arrow, following elicitation of a knee jerk, marked KJ,during the periods between spontaneous myoclonic jerkings. No correlate of the secondary discharge inthe muscles occurs in the EEG. The signal trace between the EEG records shows that the jerks areevoked, not spontaneous, and does not indicate the exact time of stimulation. Calibrations for (a) and (b)are on (b), and those for (c), (d), and (e) on (e).

14 G. D. DAWSON

on --wTK-fto PI



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



ft f ~~~~~~~~~~~~~~~~~~~~~~w I

337d/AJ25EXT. --jLEFT THICKH. 5lVI_FL_1I

FIG. 9.-Records from the muscles of the leg at higher amplifications show an abnormally large amount ofresting fasciculation between the episodes of myoclonic jerking. No correlate of the fasciculation appears inthe EEG.

following a tendon reflex. In the normal personthe rebound jerk has an action potential that doesnot exceed one-tenth of that of the primary jerk,whereas in this stage of the myoclonic state it isoften as big as, and sometimes greater than, that ofthe primary jerk. Also in the normal person alarge primary jerk is usually associated with a longerperiod of inhibition than that which follows a smallprimary jerk.

In the records shown in Fig. 8 a comparativelylarge potential difference is picked up in the flexorsurface electrodes at the time when the extensoraction potential of the initial reflex jerk occurs.Some of this potential difference may be due to thedirect electrical leakage mentioned in the section onmethods. That the major part of the potentialdifference is not due to leakage is suggested by theabsence of any large flexor surface potential duringthe presence of the secondary extensor jerk poten-tial, even though this is larger than the primarydischarge on two of the occasions shown (Fig. 8(d) and (e)). If this flexor surface potential differ-ence is arising in flexor muscles, and it seems thatthis possibility needs to be considered, it must eitherbe excited by the same afferent volley that leads tothe extensor discharge, or by the afferent volleyexcited by the stretching of the flexor muscles at thetime the extensor muscles contract. Measurementsshow a small but relatively constant delay of 5 to10 msec. between the beginning of the extensoraction potential and that of the flexor muscle.This apparent delay is probably not due to any lackof alignment in the oscillographs, errors due to thiscause in these records being less than 5 msec. atthe recording speed used. This order of delaysuggests that the second possibility is the less likelyof the two. Without more accurate timing methodsit is unwise to draw any conclusions from thisfinding but, since it is clear that when the myoclonicjerkings occur there is apparently a failure of

reciprocal innervation (Figs. 3 and 4), it may besuggested that a similar failure can also occur tosome extent in this earlier stage when spinal reflexexcitability is increased. That the failure ofreciprocal innervation in the myoclonic jerk is notjust an artefact produced by electrical leakage ofextensor surface potential differences to the flexorsurface electrodes is clear, because the flexor surfacepotential differences commonly appear before thosefrom the extensor aspect of the limb, and also theymay increase in size when the extensor potentialsare not altering (Fig. 5).

If recordings are made with a high amplificationfrom either flexor or extensor muscles at this stageof the attack an abnormally large amount of restingfasciculation is found (Fig. 9). This fasciculationis irregular and shows no tendency to becomeorganized into bigger jerkings. No correlate of thefasciculation is seen in the EEG. This may possiblybe another manifestation of the increased motorneurone excitation shown by the increased reflexexcitability and rebound mentioned above.

In the second stage of the prodromal state thatmay be distinguished clinically, the reflex responseto a tendon tap is followed by a generalizedmyoclonic jerk of the same type as those occurringspontaneously. These jerks may also be provokedby sudden unexpected noises, but a tendon tapseems to be a particularly adequate stimulus.During this stage a clear correlate of the jerk isfound in the EEG (Fig. 10 (a) to (c), which showsthe effect of eliciting a triceps jerk in the left arm.The spike potentials that occur in the EEG are thesame in general characters as those occurringspontaneously. The delay between the start of thetendon reflex action potential, marked TJ in therecord, and the first significant wave in the EEGvaries, in the records shown, between 130 and150 msec.; a similar range of values has been foundin other experiments. In the record of Fig. 10 (c)

15

NIVIWO-V



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


the first jerk elicited was unaccompanied by any

cerebral spikes. A spontaneous spike on bothsides follows, suggesting a spontaneous rise inexcitation, and the next tendon tap is followed by a

series of spikes and muscular jerks after an intervalof 140 msec. The records of Fig. 10 seem to showthat the cerebral discharge starts in larger size on

the side of the head contralateral to the limb stimu-lated than that ipsilateral to it. This result has not

been found with sufficient constancy to be accountedsignificant, but two records taken on the same

occasion under the same conditions, when a stimulushas been applied by coincidence after a spontaneousdischarge has already begun (Fig. 11 (a) and (b)),show the cerebral discharge starting in a more

nearly symmetrical fashion. This patient on some

occasions had a tendency to asymmetry of the dis-charges and before this may be taken as evidencethat the tendon tap evokes a discharge in the contra-lateral cerebral cortex first, further controls are

needed. Both clinically and from the point ofview of the EEG this stage is suggestive of the

condition of an animal to whose cortex a subcon-vulsive quantity of a convulsant drug has beenapplied.

2. Myoclonus with Wave and Spike DischargeINTERSEIZURE RECORDS.-The subject who has

attacks of mild muscular jerkings that last for 24 to48 hours shows in his EEG between these episodesa main rhythm that varies between 10 and 16 c/s,that is symmetrical on the two sides, largest inoccipital leads, and that disappears on opening theeyes. In addition there is low voltage activity at4 c/s from all parts of the head. Overbreathing pro-

duces a paroxysmal discharge at 6 to 8 c/s all over

the head, followed by an outburst of waves andspikes of regular form lasting up to 12 seconds.This outburst is accompanied by mild clonicjerkings of all the limbs and a failure to respond toorders.RECORDS DURING SEIZURES.-At the start of the

prolonged periods of jerking the EEG shows a con-

tinuous discharge of slow waves and irregular

4O7dI/A74.

R. TRICEPS Dr.

L Tr.BSr.

a. TJ

3 ~~~~~~~~~~~~~~~~~~~~~I my

|W"

Iv

C.TJ~~~~~~AAAA MPMN

407d/AI33..1 fi

FiG. 10.-Records from the triceps and biceps muscles of the arm to show myoclonic jerkings with spikes inthe EEG following the elicitation of a triceps jerk, marked TJ. This occurs at a later stage of the myoclonicstate than the effect shown in Fig. 8.

16 G. D. DAWSON

V-T--il

b. .-ri 0.02 ssr-.. . A A I A .1- 1. A! II PA IRA AAA A.. AI AAA AA i AAA ik AA, A b A.! AA Ih AA& AAA Ai% AI AAA AI AA



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from



407dIAI155.

--~~~~~~~~1- ---lR. TRICEPS Br.

L.-T,. &.--I

a.~MjW'v0,W'vWThWNA^ vWTh\M '^&WNtW'WPVWMAAMiV'

L.MOTOR.

407d/A1205-

-~~~~-A

b. .fllt5R\f00A.$#Nr0J^JNuaMA,j?#tMWMJAvwWAmIbwO

0.01 SEC.

,S~~~~~~~~1rFIG. 11.-Similar record to that in Fig. 10, but showing the triceps jerks elicited by coincidence after a spon-

taneous jerk has already begun. The EEG record is more symmetrical than in the records of Fig. 10. wherethe EEG discharge follows the triceps jerk.

I377fi/1. 'r- I . , , ._.

a. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~FIG. 12.-Record from the case with prolonged wave and spike discharge with myoclonus. (a) At the beginning

of the attack. (b) Twenty-four hours later when the discharge has become more regular. Calibrations andelectrode placements are the same for the two records.

C

17P



j.com/

J Neurol N


ownloaded from


G. D. DA WSON

ZC

4._

0

to,

~ oCt t

X

I._

con

UO

o JC._

on

c-)Ccn

LX C)

0 0"4o >.

C.)-

oI. Q-

18P



j.com/

J Neurol N


ownloaded from



spikes (Fig. 12 (a)) over all parts of the head. Thisstage is characterized by diffuse and irregularmuscular twitching in all parts of the body andsome clouding of consciousness. After about fiveor six hours the EEG shows a more regular dis-charge of wave and spike form (Fig. 12 (b)), andthe jerking becomes regular and synchronous indifferent parts of the body. In this stage there isno great reduction in consciousness and, thoughthe subject may be euphoric and frequently whistlestunes, he is able to answer questions intelligently.When this type of discharge has become establishedit may persist, with little change except some increasein regularity, for 24 hours at 2-5 to 3 c/s, symmetricalon the two sides and with amplitude up to 800 ,uV.The muscular jerkings in this condition are

clinically very slight compared with those occurringas prodromata of a major seizure, described in thelast section. Records of muscle action potentialsalso show considerable differences between the twotypes. The discharges in this condition are stillbrief and may occasionally be as short as those inthe myoclonic jerkings, but usually they are longer,have more phases, and an altogether less simpleform (Fig. 13 (a) and (b)). In addition the muscledischarge may remain localized to the flexor muscles,where it first appears, with only occasional dis-charges in the extensors. The time relations of theEEG spike and the muscle action potential appearto be more variable in this condition than in themore severe myoclonic state. In that state themuscle action potentials have never been found toprecede the cortical wave and usually are delayedafter it by not less than 15 msec. In the wave andspike seizure the muscle action potential does notseem to be so constantly delayed and may actuallyprecede the earliest recognizable phase of the EEGspike (Fig. 13 (b) at the point marked by the arrow.)This suggests that in the wave and spike seizure theefferent discharge giving rise to the muscular activityis not a direct correlate of the EEG spike, but that

both are manifestations of some process thatreaches its peak at the time they occur. That thepresence of a detectable cortical spike is not neces-sary for the occurrence of a muscle discharge in thewave and spike seizure is also suggested by the laterpart of the record in Fig. 13 (b). In this case itcannot be ruled out, however, that the apparentabsence of the cortical spike may be due to theelectrode arrangement used, and a movement ofthe area of greatest cortical activity with respect tothe electrodes; this is suggested by the slow declineand return ofthe EEG spike at that part ofthe record.The difference in size of these muscle action

potentials when compared with those in themyoclonic prodrome is striking. In the latterstate, with the electrode arrangements used, thepotential differences may be as much as 10 mV.Under the same recording conditions the potentialsduring the wave and spike seizure rarely exceed200 ,uV. The size of the spikes in the EEG issimilar to that in the myoclonic prodrome, theirduration is of the same order, and the same areasof cortex are discharging abnormally. Some otherfactor than these must therefore be sought toaccount for the small size of the muscle dischargesin this wave and spike seizure, and their absence oreven smaller size, in most seizures of this type. Itmay be that one factor is the longer interval betweenthe efferent discharges in the wave and spike seizure.Since these discharges seem to reach their peak atabout the time the spike occurs in the EEG theydo not approach within less than about 0 3 of asecond of one another. In the myoclonic prodromeit was seen that during such a period some decaywill occur of the facilitation associated with a pre-ceding spike or spikes. Spikes appearing at a rateof three per second, and not preceded by a burst ata faster rate, may not be associated with the buildingup of sufficient excitatory state to cause a largerdischarge in the muscles. That this factor maypossibly be an important one is suggested by the

377e/ 814.F.XT. 1

RIGHt THIGH.fiov

250pV

FIG. 14.-Record of the muscular jerkings accompanying wave and spike discharge evoked by overbreathing.Compare the small size of the muscle action potentials with those in the severe myoclonic jerkings (Fig. 4 (a)).The small deflections in the flexor record before and after the seizure are an artefact due to the electrocardiogram.

19P



j.com/

J Neurol N


ownloaded from


G. D. DAWSONrecord shown in Fig. 13 (b). Here the sole dis-charge in the extensors takes place after threewaves and spikes have occurred with an intervalbetween them 20 per cent. shorter than that betweenthose preceding them. Thus the differences in thetwo conditions may be due solely to the differentamounts of facilitation occurring, and it is notnecessary to postulate that the efferent cerebraldischarges in them are different in either nature ormagnitude.

In the periods following spontaneous seizures,overbreathing may sometimes evoke wave andspike discharges. In the later parts of the inter-seizure period these evoked outbursts are associatedwith muscular jerkings similar to those occurringin the spontaneous attacks, and the muscle actionpotentials are synchronous with the spikes in theEEG (Fig. 14), and here as in the myoclonic pro-drome the discharge begins in the flexor muscles.In the earlier part of the interseizure period thenumber of wave and spike complexes evoked by a

given amount of overbreathing is less than at alater time. Often only a single wave and spike orslow wave occurs (Fig. 15 (a)) and no muscle dis-charge is seen. If the muscle or group of musclesbeing tested is now stretched slightly it is oftenpossible to pick up a simple rhythmic dischargefrom what is apparently a single motor unit.Under these conditions when a single wave andspike or slow wave occurs in the EEG the rhythmof the single unit is interrupted and a discharge isinterpolated (Fig. 15 (b) and (c)). In the latterpart of the figure it seems that the rhythm is " reset "

from the point where the interpolated dischargeoccurs. In this case the delay between the wave inthe EEG and the muscle discharge is longer andless constant than it is when the regular repetitivespikes occur in the EEG. Probably the delaydepends on the level of excitation at the motorneurone level at the time the efferent cerebraldischarge reaches it.

DiscussionThe results of the investigations described above

in the patients with severe myoclonic jerkings con-firm the findings of Grinker et al. (1938), and inthe patient with wave and spike discharge withmyoclonus those of Jasper and Andrews, (1938).The present cases of severe myoclonic jerkings showspikelike discharges in the EEG apparently exactlycomparable with those described by Grinker et al.After discussing their findings Grinker et al. decidedthat the disturbances in myoclonic epilepsy probablyhad a cortical rather than a subcortical origin.The evidence available from the present studiesdoes not make it profitable to speculate further onthis point, for the disturbances in the EEG inmyoclonic epilepsy are so large and so widespreadthat it seems highly unlikely that of cortex or basalstructures, pyramidal or other efferent systems,any one will be active in isolation. In spite of theprobable diffuse natuire of the disturbances duringthe seizure there are indications for further study.

There are clearly changes in the tendon reflexeswhich indicate an increase in spinal excitability, atany rate for extensor reflex mechanisms. Sincethese changes are unaccompanied by any obvious-changes in the EEG, some steadily maintaineddischarge affecting the spinal centres, or the removalof such a discharge, must be of importance in theperiods between jerkings preceding the major con-vulsion. How far these changes in spinal excita-bility determine the pattern of the jerkings whenthese do occur is not clear, though it is obviousthat the part played by them is considerable. Thegreater accessibility of the flexor mechanisms to theabnormal discharges in the two cases described,shown by the muscle action potentials, may betaken to suggest that although the pyramidalsystem is almost certainly not the only one dis-charging abnormally in these cases, it has thedominating influence at the spinal level.

It must be emphasized that the finding thatmuscular jerkings accompany spike discharges inthe EEG does not apply to all types of myoclonicepilepsy. Many of the minor myoclonic pheno-mena that occur so frequently in epilepsy (RussellReynolds (1861) has estimated that 75 per cent.of all epileptics have them in some form) seemto show no correlates in the EEG. One subjectat The David Lewis Colony showed a rhythmictwitching of the face, arm, and leg of one side ofthe body only, which could not be associated withany disturbance in the EEG. Unfortunately thejerkings in this patient ceased before full investiga-tions could be made.The findings in the patient with wave and spike

discharge with mild mycolonus show a rather closercorrelation of the spike of the complex with theburst of muscle action potentials than in similarcases described by Jasper. Like his cases, thispatient also shows that the spike of the cerebraldischarge and the muscle discharge may not bedirectly related but are probably both correlates ofsome third process. It seems that the presence ofefferent cerebral discharges, in themselves sub-liminal for excitation at the spinal level, may beinferred from their effects on a purely spinal reflexmechanism, such as the rhythmic discharge of astretched single motor unit. This suggests that ifstudies are carried out at the spinal reflex level, inconditions other than epilepsy, a closer correlationmay be found between cerebral rhythms and peri-pheral effects than has so far been shown by studiesat the lower level of the muscles themselves.

Summary1. The EEGs oftwo subjects with severe myoclonic

seizures and one subject with mild myoclonic jerk-ings accompanying wave and spike discharge in theEEG are described.

2. Repeated simultaneous recordings of theEEGs and muscle action potentials have been madein these subjects during myoclonic episodes.

3. All subjects show some form of spike dis-charge, with or without a slow wave, in the

20P



j.com/

J Neurol N


ownloaded from



0

V

LU.,

I

1C:r

V

to._

Q

C)a

-o

C)

CZ

C')

C.t

~C)

-D

~C)

0 E

C)

.0

DC)

Ct

Ct

U .'o 0C)-

6z

t

IL

Ld 4GoNN

C6

GON

)

0

0)

GoN ".:IxX.

21

I

I



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


22 G. D. DA

inter-seizure periods. In the seizures the patientswith the severe myoclonic jerkings show dischargesof spikes in their EEGs with a period of 30 to 75msec. at a rate of 8 to 13 per second.

4. There is an accurate correlation between theoccurrence of the spikelike discharges in the EEGand the action potentials of the muscular jerkingsin the severe cases and a less accurate correspondencein the mild case.

5. The muscle action potentials show recruitmentand facilitation having a different time course ofdecay for flexor and extensor muscles.

6. In the periods between the spontaneous jerk-ings changes in spinal reflex excitability may bedemonstrated.

7. Important differences in the characters of themuscular jerkings in the severe myoclonic statesand the milder state accompanying wave and spikedischarge are described.

8. These findings are discussed and their bearingon the origin and mediation of the abnormal dis-charges in these conditions are considered.The author would like to express his thanks to the

Committee of The David Lewis Colony who made this

[WSONwork possible, to the Medical Research Councilfor facilities to complete it, and to Dr. Handley,Director of the Colony, for his advice and constantencouragement.

REFERENCESAdrian, E. D., and Bronk, D. W. (1929). J. Physiol.,

67, 119.Buchthal, F., and Nielsen,'J. 0. (1936). Skand. Arch.

Physiol., 74, 202.Gibbs, F. A., Davis, H., and Lennox, W. G. (1935).

Arch. Neurol. Psychiat. Chicago, 34, 1133.Gibbs, E. L., and Lennox, W. G. (1937). Brain,60, 377.and Gibbs, E. L. (1941). Atlas of Electro-

encephalography, Cambridge, Mass.Grinker, R., Serota, H., and Stein, S. (1938). Arch.

Neurol. Psychiat. Chicago, 40, 968.Jasper, H. H., and Andrews, H. L. (1938). J. Neuro-

physiol., 1, 87.(1941). In Epilepsy and Cerebral Localisation,Penfield and Erickson, Balliere, Tyndall & Cox,Loiidon.

Offner, F. (1937). Rev. Sci. Instr., 8, 20.Reynolds, Russell (1861). Epilepsy, Churchill, London.Strauss, H., and Landis, C. (1938). Proc. Soc. exp.

Biol. N.Y., 38, 369.Toennies, J. F. (1938). Rev. Sci. Instr., 9, 95.Walter, W. G. (1936). Lancet, 231, 305.



http://jnnp.bmj.com

/J N

eurol Neurosurg P


nloaded from


the relation between the electroencephalogram and … · muscular jerkings in convulsive states has...

Documents