EMG analysis of stereotyped voluntary movements in man'EMGanalysis ofstereotypedvoluntary movementsin man (Basmajian andLatif, 1957). Theinitial elbowangle was 900. EMGactivity was

Journal of Neurology, Neurosurgery, and Psychiatry, 1975, 38, 1154-1162.

EMG analysis of stereotyped voluntary movementsin man'

MARK HALLETT, BHAGWAN T. SHAHANI, AND ROBERT R. YOUNG2

From the Department ofNeurology and the Laboratory of Clinical Neurophysiology, Harvard Medical School,and Massachusetts General Hospital, Boston, Massachusetts, U.S.A.

SYNOPSIS EMG activity was recorded in biceps and triceps while subjects voluntarily flexed theirelbows during a visual matching task. With fast flexion, the initial EMG was characterized by atriphasic pattern with a burst of activity first in biceps, then in triceps with a silent period in biceps,and finally in biceps again; these components were analysed quantitatively. Smooth flexion wascharacterized by continuous activity in biceps. Inhibition of tonic activity of triceps in relation to afast flexion occurred in the 50 ms before the initiation of biceps activity. A patient with a severe pan-sensory neuropathy performed normally on these tasks. Physiological mechanisms underlying thesepatterns are analysed; an important conclusion is that the triphasic activity with fast flexion is'centrally programmed'.

The precise timing ofEMG activity in the severalagonist and antagonist muscle groups of thehuman upper limb during voluntary projectedmovement was initially described in 1926 byWachholder and Altenburger, and Wachholder(1928). They found different EMG patterns to beassociated with movements of different velocity.With slow movements there was continuousEMG activity in the agonist, whereas the patternwith rapid movements consisted of a stereotypedalternation of activity between agonist andantagonist. Since then, there have been manycontradictory reports suggesting, for example,that muscular activity in all movements is almostentirely restricted to the agonist (Wilkie, 1950),that the antagonist may be active during theentire movement (Barnett and Harding, 1955),or that the antagonist is active only at the end ofa movement (Bierman and Ralston, 1965).Recent investigators studying EMG activity inprojected movement have, in general, confirmedWachholder and Altenburger's findings (Bas-majian, 1967; Gottlieb et al., 1970; Marsdenet al., 1972; Dijkstra and Denier van der Gon,1 Supported by the Parkinson's Disease Project of the MassachusettsGeneral Hospital and the Allen P. and JosephineB. GreenFoundation.2 Address for reprints: Dr R. R. Young, Massachusetts General Hos-pital, Boston, Massachusetts 02114, U.S.A.(Accepted 18 July 1975.)

1973; Hopfet al., 1973; Angel, 1974), but a com-plete and satisfactory understanding of the EMGpatterns is still lacking. The present study is anattempt to define these patterns quantitativelyfor certain stereotyped movements in a normalpopulation, to learn more about the physio-logical mechanisms involved, and to set stan-dards by which to evaluate the performance ofpatients with various movement disorders.

METHODS

Eighteen normal subjects (15 males and three femalesranging in age from 22 to 44 years) and one patientwere studied. The latter is a 38 year old man with asevere pan-sensory neuropathy. He has little sub-jective appreciation of touch, pain, temperature,position, or vibration, has no deep tendon reflexes,and shows no silent periods in ongoing voluntaryEMG activity to electrical stimulation of the mixednerve. Details of his clinical state have been reportedelsewhere (Adams et al., 1973); his limbs are con-sidered essentially deafferented.

Studies were restricted to elbow flexion. Subjectssat facing an oscilloscope screen with their armresting on a table with a Formica top that wascovered with talcum powder to reduce friction. Thearm was abducted approximately 90° at the shoulderand the forearm was supinated so that biceps andbrachialis become the prime flexors of the elbow

154

guest. Protected by copyright.

on February 17, 2021 by

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.38.12.1154 on 1 Decem

ber 1975. Dow

nloaded from

http://jnnp.bmj.com/

EMG analysis of stereotyped voluntary movements in man

(Basmajian and Latif, 1957). The initial elbow anglewas 900. EMG activity was recorded from the skinover the bellies of biceps and triceps using elec-trodes 3.0 cm apart. A lever strapped to the wristextended vertically from a potentiometer which con-verted movement of the lower end of the lever intoa variable voltage. The output was connected to thevertical axis of one channel of the dual beam oscil-loscope observed by the subject. The vertical axis ofthe second channel of the oscilloscope was con-nected to a signal generator under the experimenter'scontrol. The horizontal axis of both oscilloscopebeams was set on rapid free-run so that the appear-ance on the screen was that of two horizontal lines.

Weights could be applied to the forearm by beingconnected to the lever by flexible wires strung overpulleys attached to the edge of the table. Weightswere attached to the wires with light-weight electro-magnets so that they could be released at any pre-scribed time. Release was signalled by the interrup-tion of an electrical circuit between the electro-magnet and the suspended weights. All this informa-tion was displayed on a second six-channel oscillo-scope, where a permanent record was made with aPolaroid camera.At the beginning of all experiments, subjects were

asked to match the experimenter's line on theoscilloscope with their line. This was done by vol-untarily adjusting the angle of their elbow and there-fore the angle of the lever attached to their forearmand the potentiometer. At a time not predictable bythe subject, the experimenter's line would suddenlydrop to a new position which could be matched bymeans of elbow flexion. Subjects were told, 'Matchthe line as rapidly and as accurately as possible'. Inpractising this movement, subjects were continuallyurged to 'Try to do it faster'. The data were re-corded only when the movement was well practisedand appeared to be as rapid as possible. In mostexperiments, the movement consisted of approxi-mately 100 angular displacement of the elbow. Asecond set of instructions, usually presented after thefirst, asked the subjects to 'Match the line smoothly,steadily, and accurately. You might have to go moreslowly, but you should begin just as promptly'. Intraining for this task, subjects might be told, 'Thatlast movement wasjerky; try to do it more smoothly'.Again, data were recorded only for practised move-ments.

Durations of the components of the EMG activitywere measured with an accuracy of + 10 ms fromthe photographic record and averaged for each sub-ject. More attention was paid to the timing of thecomponents than to their amplitudes which weremerely approximated by measurement of the max-imum voltage seen within any component.

The results using five different protocols are re-ported in this paper. The first was a fast flexion (FF)without any weights. In the second, designed tostudy antagonist inhibition (Al), the triceps wasmade to contract tonically by attaching a weight tothe lever which, if unopposed, would act to flex theelbow. The subject was then told to match the linerapidly. The third was a smooth flexion (SF) withoutweights. The fourth protocol was arranged to get apassive extension (PE) of the elbow during a fastflexion effort. Approximately equal weights, usuallyabout 2 kg, were attached to both sides of the lever.With equal and opposite forces on the forearm therewas no net pull and consequently no resting activityin either biceps or triceps. Subjects were instructedto match rapidly. Sometime between the displace-ment of the experimenter's line and the onset ofbiceps activity, the weight pulling in the direction ofelbow flexion was electromagnetically released. This

. I

i' I't'l , st00--PIwil,A~~

A

I

BA FAST-FLEXION MOVEMENT

BB KB2

liAAgBICEPS--U - vmx-AiA

TRICEPS ,

FLISTEP _

ARM POSITION

EXION

tMARKER, J \ _

V/ELOCITY

FIG. 1 Electromyographic activity in biceps andtriceps during fast flexion of the elbow. A is a photo-graphic record; the six oscillographic traces arediagrammatically described in B. Components of theEMG activity pictured in B are further defined in thetext. Calibration: 50 ms, 500 t V.

1155

I . . I



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from


Mark Hallett, Bhagwan T. Shahani, and Robert R. Young

TABLE 1NORMAL VALUES (MS) FOR FAST FLEXION EMG PARAMETERS*

Name S-BJ BI-B2 BJ-TJ T1-B2 BJ B2 Ti

GL (22) 130 40 -5 -15 65 > 100 60RM (3) 170 60 0 0 70 50 60TL (2) 265 105 20 10 105 145 60SF (4) 220 85 0 5 80 > 175 80GC (3) 255 45 0 0 80 >95 45JC (7) 180 40 0 5 85 40 30BD (7) 190 100 10 35 60 35 55SH (7) 150 140 25 35 75 65 80BS (7) 175 55 -5 15 65 25 45HF (5) 225 150 0 65 65 10 85MH (9) 210 60 -10 5 75 55 65DD (2) 220 75 -15 25 105 > 100 60CO (4) 190 80 - 15 30 95 > 70 45GT (6) 170 45 -20 10 100 95 50TB (5) 175 65 -5 -10 75 50 75DC (7) 210 45 0 -5 100 > 105 55MF (8) 270 45 -20 -25 65 > 110 85JB (4) 235 80 0 5 90 40 55

Average(18) 200 70 -5 10 80 > 80 60

Range 130 to 40 to -20 to -25 to 60 to 10 to 30 to270 150 25 65 105 > 175 85

* Data are averages for the number of trials shown in parentheses after the subject's initials.

caused a passive extension of the elbow and tensionon the triceps was reduced. When the biceps con-tracted, it was not successful in flexing the elbow orstretching the triceps. In performing this experimentthe weight was dropped at random times in only one-third of the trials. In the fifth experiment (SU:smooth unload), the unloading reflex during smoothflexion was studied. A weight, usually about 2 kg,was attached to the lever pulling in the direction ofelbow extension, requiring tonic activity in thebiceps to maintain elbow flexion of 900. This weightwas also released at random times in only one-thirdof the trials in which the subjects were asked tomatch smoothly.

RESULTS

FAST FLEXION (WITHOUT WEIGHTS) The initialexperiment for most subjects was to flex theelbow trying to superimpose their line upon theexperimenter's line as rapidly as possible (proto-col FF). The initial part of the EMG pattern inall subjects could characteristically be describedin the following manner (Fig. 1, A). There is aburst of action potentials in biceps (B 1) followedby a period of silence in that muscle (BI-B2)during which triceps is active (Tl). Triceps thenbecomes silent and biceps active again (B2).Usually further electrical activity is seen but is

not analysed here because a pattern for the latteractivity is much more difficult to discern thanfor the initial part.

In order to study the initial EMG patternquantitatively we arbitrarily defined certainfurther parameters of the pattern which areillustrated in Fig. 1, B. Bl, Bl-B2, TI, and B2refer to the durations of these components de-fined above. S-Bl refers to the time from thechange in position of the experimenter's line tothe beginning of Bi ('reaction time'). BI-TIis defined as the interval between the end of B1and the beginning of Tl, and Tl-B2 as theinterval between the end of TI and the be-ginning of B2; if there is overlap between B1 andTl, or TI and B2, these intervals will have anegative value.

Analysis of the averaged data for 18 normalcontrols is shown in Table 1. The measurementsof the parameters are moderately variable fromtrial to trial for a single subject and, as can beseen, there is a marked variability from one sub-ject to another which gives rise to broad 'normallimits'. The parameters which seem least var-iable are B1, TI and, to a lesser extent, B1-B2.

ANTAGONIST INHIBITION BEFORE AGONIST ACTIVA-

1156



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



FIG. 2 Inhibition of tonictriceps activity before a fastelbow flexion. This is acomposite offour indi-vidual records from one

subject. The first line is atracing of the bicepsactivity and the fifth lineis a tracing of the tricepsactivity of the same trial.Lines 2 and 6, 3 and 7, and4 and 8 are related simi-larly. The tracings are

arranged so that the verticalline denotes the onset ofBJ. Calibration: 20 ms,200 pLV.

TION In this experiment, the temporal relationbetween the cessation of EMG activity in a pre-viously contracting antagonist and initiation ofEMG activity in the agonist is investigated. Thetriceps is made to contract tonically before a

rapid flexion movement (protocol Al).Triceps activity ceased 0-50 ms before the

initiation of Bl. When triceps remained activeup to the moment that biceps began, there wasnever any overlap of the activities in the twomuscles. In addition to the cessation of tricepsactivity before the onset of Bl, the amplitude ofthe terminal part of the triceps trace was oftendiminished. This phenomenon may best bethought of in a statistical sense-that is, theprobability that EMG activity will be recordedfrom triceps becomes progressively less duringthe 50 ms before agonist activation (Fig. 2).

Four trials are recorded, the EMG traces fromwhich are arranged so that the onset of BIoccurs at the time indicated by the vertical line.In this example, the triceps activity stopped 6,16, 28, and 48 ms before the initiation of B1.

SMOOTH FLEXION (WITHOUT WEIGHTS) Subjectswere instructed to match the line smoothly(protocol SF). In this circumstance, completelydifferent EMG patterns were seen. Most com-mon was continuous activity in biceps withoutany triceps activity (Fig. 3). At other times, therewas also essentially continuous and simultaneoustriceps activity. Alternating biceps and tricepsactivity were not seen unless the movement wasso quick as to mimic the fast flexion task.

PATTERNS IN 'DEAFFERENTED' MAN This subject

TABLE 2VALUES (MS) FOR FAST FLEXION EMG PARAMETERS IN PATIENT WITH PAN-SENSORY NEUROPATHY

Values S-BJ BI-B2 BJ-Tl T1-B2 BJ B2 Ti

Average 90 70 -5 -10 85 > 105 90Range 130- 20- -40- -135- 50- 20- 20-

350 135 30 30 130 410 170

1157



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



I

FIG. 3 EMG activity during a smooth flexion (SF).The first line is biceps activity, the second is tricepsactivity, the third is the step to be matched, and thefourth is the arm position. Calibration: 50 ms, 20 j V.

A/'9{1_- 4' ~~~~~~~~~~~,W

C

_~~~~

-v--

-1

FIG. 4 EMG of a fast flexion movement (FF) forthe patient with the pan-sensory neuropathy. Tracingsare the same as in Fig. 1. Calibration: 50 ins, 200 ±V.V

B!

-hW~~~~~i

.

1. I

liLlf*M%O^ -6qVYV%A. V

II I

FIG. 5 The effect ofpassive elbow extension on thefast flexion pattern (PE). The tracings in each partof the figure are similar to Fig. 1. A illustrates theEMG pattern with a 2 kg weight on each side of thearm. In B, the weight pulling in the direction ofelbow flexion is dropped at the time noted by thearrow. In C, no voluntary movement is requested; thesubject merely matches the baseline and the weight isdropped as in B. The record shows three superimposedtrials. A small stretch reflex is seen in biceps, buttriceps is quiet. Calibration: 50 ms, 500 ,u V.

-&&-AMA a - ---4NNANA-p- 'WIMVIEWWI - --Im

1158

I,

44;;.Ilwv-- 17%



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



300

200

1 00

Control

-100

FIG. 6 Average changes inEMG components of the fastflexion pattern with passiveextension. EMG componentswere measured and averagedfor six subjects with andwithout passive elbow exten-sion. The grand average for thepercentage change of eachcomponent produced byrelease of the weight isplotted on the diagram.

DurBi AmpBi DurBl-B2 DurTi AmpTi

was asked to do the three testing protocols de-scribed in the preceding sections. All three wereperformed normally. An example of the fastflexion (FF) pattern is shown in Fig. 4. Averagesand ranges of the various parameters for 22trials are shown in Table 2.

FAST FLEXION WITH PASSIVE EXTENSION Thepattern during a fast flexion movement with bothweights attached (Fig. 5) is similar to that seenwithout weights. When the weight pulling in thedirection of elbow flexion is released, TI waspreserved both in duration and amplitude de-spite the fact that the arm was passively ex-tended (Fig. 5, B). To show that this tricepsactivity is not merely part of a reflex patternproduced by the passive extension, the weightwas released at times when the subject was nottrying to make a voluntary movement and thereis no triceps activity under this condition (Fig.5, C).Components of the EMG pattern were com-

Dur B2 AmpB2

pared with and without release of the weight forsix subjects. For Bi, TI, and the amplitude ofTI there was no significant change with release ofthe weight, some subjects showing a small in-crease and some a small decrease. B1-B2 in-creased in four subjects, remained the same forone, and decreased slightly for the last. Theamplitude of Bi increased slightly for all sub-jects and the amplitude of B2 increased to amuch larger extent in all subjects. The averagepercentage change in these parameters producedby release of the weight is illustrated in Fig. 6.

UNLOADING REFLEX DURING SMOOTH FLEXION Un-loading responses were obtained during a smoothflexion (protocol SU). On control trials whenthe weight was not dropped, the rate and ampli-tude of the biceps EMG increased after a cer-tain reaction time. On the experimental trials,the weight was dropped at various times after(and occasionally even before) four subjectswere given the stimulus to move. At no time for

1159



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



any subject did the unloading reflex changesignificantly in its duration.

DISCUSSION

NORMAL EMG PATTERNS ASSOCIATED WITH A FASTMOVEMENT In order to understand how thenervous system controls movement, the detailsof the movement itself must be understood.Even the specific EMG patterns recorded fromagonist and antagonist muscles during a simplerapid flexion movement are rather complex. Atthe beginning of the movement any activity inantagonist groups ceases and the agonistactivity appears in two bursts separated by asilent period; activity in the antagonist is noted,to a close approximation, during this silentperiod. The physiological question is, can thesereproducible patterns be explained entirely as aproduct of segmental reflex interactions? If not,it would be necessary to assume the presence ofsuprasegmental 'centrally programmed' inputdirectly to the motor neurone pools. Answerscan be obtained by looking at the interactions ofH-reflexes, tendon reflexes, muscle stretching orunloading during the course of the movement.EMG activity in a tonically active antagonist

muscle ceases before agonist activation and thus,in the experimental setting, Al is an earlier signof voluntary movement than the onset of agonistactivity itself. This phenomenon has been de-monstrated in other settings, both in man(Hufschmidt and Hufschmidt, 1954; Gottliebet al., 1970) and in monkeys (Evarts, 1974). Be-cause this phenomenon is present in the patientwith the pan-sensory neuropathy, we suggestthat it can be accounted for by a central signaldirectly onto the alpha motor neurone pool.Other studies, with unloading (Angel et al.,1970) and H-reflexes (Pierrot-Deseilligny et al.,1971; Simoyama and Tanaka, 1974), have sug-gested that there is also normally a decrease inthe influence of the gamma loop at this sametime.The presence of the double burst pattern of

the agonist in the deafferented patient suggeststhat it is also centrally programmed. The silentperiod between the bursts must come from aninhibitory, or decrease of excitatory, influenceof this suprasegmental signal. Garland and Angel(1971) showed that BI is not affected by unload-

ing, while B2 is. Using passive muscle stretchrather than unloading (the PE task), our resultsshow a slight increase in amplitude of B1, amild foreshortening of Bl-B2, and a large in-crease in the amplitude and slight increase induration of B2. Thus, segmental afferent inputhas little effect on the agonist early in the fastflexion movement, but increases its influence asthe movement progresses. On the other hand,the amplitude of the H-reflex of the agonist in-creases before, and reaches a peak about, thetime of the onset of BI; it then gradually de-clines throughout the course of the movement(Gottlieb et al., 1970; Pierrot-Deseilligny et al.,1971). It is of interest that polysynaptic reflexesin man also have two components (Shahani andYoung, 1971, 1972) which appear to have simi-lar characteristics to those seen here for vol-untary movements, but the physiological inter-pretation of this similarity is uncertain.Most investigators (Basmajian, 1967; Gottlieb

et al., 1970) assume that Ti is a result of astretch reflex produced by passive extension oftriceps by Bl. Two results here disprove thathypothesis. Tl is present in the deafferentedpatient; and TI is present, and to a large extentunmodified, even when the elbow is passivelyextended (and triceps NOT stretched) during anattempt at fast flexion. It appears, therefore,that Ti, like B1, is part of a centrally pre-pro-grammed pattern which is not particularly sub-ject to segmental influence. Because TI is notexactly co-extensive with B1-B2, and becauseTI is not subject to segmental influence, whileB1-B2 is, TI and B1-B2 are apparently or-ganized separately and not merely reciprocal.The amplitude of the H-reflex in the antagonistis reduced during the course of the movement(Gottlieb et al., 1970), but this reduction istransiently less marked about the time of Ti.For both the agonist and antagonist, there are

at least three, not necessarily independent, con-trolling forces which differ in time course. Firstis the pre-programmed suprasegmental signaldirectly to the alpha motor neurone pool.Second is a similar signal directed to the gammamotor neurone pool, which influences the alphamotor neurones indirectly via the servo-loop.Third is a signal thought, on the basis of H-reflex studies, to control Ia feedback gain(Gottlieb et al., 1970). The time course of these

1160



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



AGONIST

ANTAGONIST

VELOCITY

AGONISTSIGNALS

ANTAGONISTSIGNALS

1111I11111 1111111

' ~ ~~'

AlphaGamma

/.,!\ Ia FeedbackI. ~~~~~~~Gain

l Il, -

FIG. 7 A diagrammaticsummary of the six controllingsignals for a fast flexionmovement. The fouirth andfifth lines illustrate quali-tatively the controlling signalsfor the agonist and antagonistrespectively. For each musclethere is the signal direct to thealpha motor neurones, thesignal to the gamma motornelurones and the signal con-trolling the Ia feedback gail?.Modifiedfrom Gottlieb et al.(1970). See text for furtherdetails.

signals, so far as is known, is diagrammaticallyrepresented in Fig. 7, which is modified fromGottlieb et al. (1970).Our conclusion is that the initial 'triphasic'

pattern of fast flexion is centrally prepro-grammed; B1 and TI are relatively immune toperturbation by peripheral influences, but B2is at least partially modifiable. Stetson andMcGill (1923) had suggested that ballistic move-ments are too rapid for feedback to have anyinfluence-that is, 'preprogrammed'. An ana-lysis of the velocity of these movements also sug-gested preprogramming (Bouisset and Lestienne,1974). Hopf et al. (1973) have come to the sameconclusion as we have on the basis of interactionof rapid movements with vibration, direct elec-trical stimulation, and stretches of muscle.The normal ranges of duration of certain

features-B1, TI, and 'inhibition' of antagonistactivity before BI-B2 are sufficiently narrowfor them to be useful as standards for evaluationof patients with movement disorders. It shouldbe noted that the values reported here are sim-ilar to previously reported approximate valuesfor similar movements (Dijkstra and Deniervan der Gon, 1973; Hopf et al., 1973; Huf-schmidt and Hufschmidt, 1954; Hufschmidt,1954). The ranges of normal values for the other

parameters of the fast flexion movement aremuch broader. For example, the range of valuesfor duration of B2 is so extensive that it would bedifficult to specify an abnormal performance.One might guess that its variability arises fromthe strong influence of feedback on this com-ponent and the subject's appraisal of whether heis over- or under-shooting the mark.

NORMAL EMG PATTERNS ASSOCIATED WITH ASMOOTH FLEXION MOVEMENT Evidence is mount-ing that the central nervous mechanisms con-trolling rapid, ballistic movements are separatefrom those responsible for slower, smooth' ramp movements' (Kornhuber, 1971; DeLongand Strick, 1974). It seemed useful, then, to tryto establish standards of performance for thislatter kind of movement as well. In all subjects,there was continuous activity of the agonist andin no subject was there a triphasic pattern asnoted above. Thus, it is possible electromyo-graphically to distinguish, in this paradigm,ballistic from ramp movements.

It had already been observed by Marsdenet al. (1972) that unloading a muscle during asteady movement produces a momentary silencein the ongoing EMG, an 'isotonic' unloadingreflex similar to the more traditional 'isometric'

1161



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from



unloading reflex. In addition to confirming that,we have made further observations during boththe period between the stimulus and responseand the first part of the response itself. A similarbrief period of EMG silence was seen in bothinstances. In this respect, fast movements andsmooth movements are different, since segmentalinput does not influence the first part of the fastflexion pattern. Thus, in addition to the EMGdifferences, the controlling signals for fast andsmooth movements appear to be different. Forsmooth movements, the servo-loop seems to becontinuously active and strongly influentialfrom the beginning of the movement.

REFERENCES

Adams, R. D., Shahani, B. T., and Young, R. R. (1973). Asevere pansensory familial neuropathy. Transactions of theAmerican Neurological Association, 98, 13-15.

Angel, R. W. (1974). Electromyography during voluntarymovement: The two-burst pattern. Electroencephalographyand Clinical Neurophysiology, 36, 493-498.

Angel, R. W., Garland, H., and Alston, W. (1970). Inter-action of spinal and supraspinal mechanisms during vol-untary innervation of human muscle. ExperimentalNeurology, 28, 230-242.

Barnett, C. H., and Harding, D. (1955). The activity of antag-onist muscles during voluntary movement. Annals ofPhysical Medicine, 2, 290-293.

Basmajian, J. V. (1967). Muscles Alive. Their FunctionsRevealed by Electromyography, 2nd edn. Williams andWilkins: Baltimore.

Basmajian, J. V., and Latif, A. (1957). Integrated actionsand functions of the chief flexors of the elbow: a detailedelectromyographic analysis. Journal of Bone and JointSurgery, 39-A, 1106-1118.

Bierman, W., and Ralston, H. J. (1965). Electromyographicstudy during passive and active flexion and extension ofthe knee of the normal human subject. Archives ofPhysicalMedicine and Rehabilitation, 46, 71-75.

Bouisset, S., and Lestienne, F. (1974). The organization of asimple voluntary movement as analysed from its kine-matic properties. Brain Research, 71, 451-457.

DeLong, M. R., and Strick, P. L. (1974). Relation of basalganglia, cerebellum, and motor cortex units to ramp andballistic limb movements. Brain Research, 71, 327-335.

Dijkstra, S. J., and Denier van der Gon, J. J. (1973). Ananalog computer study of fast, isolated movements.Kybernetik, 12, 102-1 10.

Evarts, E. V. (1974). Sensorimotor cortex activity associatedwith movements triggered by visual as compared tosomesthetic inputs. In The Neurosciences. Third StudyProgram, pp. 327-337. Edited by F. 0. Schmitt and F. G.Worden. MIT Press: Cambridge, Mass.

Garland, H., and Angel, R. W. (1971). Spinal and supraspinalfactors in voluntary movement. Experimental Neurology,33, 343-350.

Gottlieb, G. L., Agarwal, G. C., and Stark, L. (1970).Interactions between voluntary and postural mechanismsof the human motor system. Journal of Neurophysiology,33, 365-381.

Hopf, H. C., Lowitzsch, K., and Schlegel, H. J. (1973).Central versus proprioceptive influences in brisk voluntarymovements. In New Developments in Electromyographyand Clinical Neurophysiology, vol. 3, pp. 273-276. Editedby J. E. Desmedt. Karger: Basel.

Hufschmidt, H. J. (1954). Die rasche Willkurkontraktion.Beitrag zur elektromyographichen Analyse der mensch-lichen Willkurmotorik. Zeitschrift fur Biologie, 107, 1-24.

Hufschmidt, H. J., and Hufschmidt, T. (1954). Antagonistinhibition as the earliest sign of a sensory-motor reaction.Nature, 174, 607.

Kornhuber, H. H. (1971). Motor functions of cerebellum andbasal ganglia: The cerebello-cortical saccadic (ballistic)clock, the cerebellonuclear hold regulator, and the basalganglia ramp (voluntary speed smooth movement) gen-erator. Kybernetik, 8, 157-162.

Marsden, C. D., Merton, P. A., and Morton, H. B. (1972).Servo action in human voluntary movement. Nature, 238,140-143.

Pierrot-Deseilligny, E., Lacert, P., and Cathala, H. P. (1971).Amplitude et variabilite des reflexes monosynaptiquesavant un mouvement volontaire. Physiology and Behavior,7, 495-508.

Shahani, B. T., and Young, R. R. (1971). Human flexorreflexes. Journal of Neurology, Neurosurgery, and Psy-chiatry, 34, 616-627.

Shahani, B. T., and Young, R. R. (1972). Human orbicularisoculi reflexes. Neurology (Minneap.), 22, 149-154.

Simoyama, M., and Tanaka, R. (1974). Reciprocal la in-hibition at the onset of voluntary movements in man.Brain Research, 82, 334-337.

Stetson, R. H., and McGill, J. A. (1923). Mechanisms of thedifferent types of movements. Psychological Monographs,32, 18-45.

Wachholder, K. (1928). Willkiirliche Haltunghund Bewegunginsbesondere im Lichte electrophysiologischer Untersuch-ungen. Ergebnisse der Physiologie, 26, 568-775.

Wachholder, K., and Altenburger, H. (1926). Beitrage zurPhysiologie der willkiirlichen Bewegung. 10. Einzelbe-wegungen. Pflugers Archiv fur die gesamte Physiologie desMenschen und der Tiere, 214, 642-661.

Wilkie, D. R. (1950). The relation between force and velocityin human muscle. Journal ofPhlysiology, 110, 249-280.

1162



http://jnnp.bmj.com

/J N

eurol Neurosurg P


ber 1975. Dow

nloaded from


Documents

EMG analysis of stereotyped voluntary movements in man'EMGanalysis ofstereotypedvoluntary movementsin man (Basmajian andLatif, 1957). Theinitial elbowangle was 900. EMGactivity was