Journal of Neutrology, Neurosurgery, and Psychiatry, 1974, 37, 825-834

Isometric contraction of the abductor digitiminimi muscle in man

DAVID BURKE, NEVELL F. SKUSE, AND A. KEITH LETHLEAN

From the Unit of Clinical Neurophysiology, Division of Neurology,Prince Henry Hospital, Little Bay, N.S. W. 2036, Australia

SYNOPSIS Isometric contraction of the abductor digiti minimi muscle (ADM) has been studied insix normal subjects. Twitch contraction times ofADM ranged from 60 to 68 ms and twitch torqueranged from 2-33 to 6-24 x 10-3 Nm. In three subjects torque declined by an average of 310% aftertetanization at 50 Hz for 30 seconds but there was no similar diminution in the evoked muscle actionpotential suggesting that the fatigue arose from intrinsic muscular mechanisms. A marked decline intetanic torque occurred with continued tetanization in two subjects for a total of five minutes, butthis change was accompanied by a decrease in the muscle action potential. In six subjects thresholdstimulation to the ulnar nerve at the wrist and to various sites over the motor point of ADMallowed 55 threshold twitch contractions to be identified after averaging. A unimodal range of con-traction times ranging from 40-100 ms was found and this was confirmed by additional experimentsin two subjects in whom 30 threshold twitch contractions were identified using a needle electrode tostimulate various sites in the motor point. Tetanization at 50 Hz was performed using thresholdstimulus levels. Nine threshold tetanic contractions were evoked in two subjects. In eight tetanictorque progressively fatigued to between 14 and 20% within 60-90 seconds, but, in one tetaniccontraction, torque proved relatively fatigue resistant. These results suggest that there is a homo-geneous group of motor units in ADM (with respect to contraction time) and that this group containswhat are probable fast twitch fatigue sensitive and fatigue resistant motor units. No evidence of a

distinct group of slow twitch units was found.

The documentation of motor unit properties inman remains in its infancy. Isometric twitchcontractions have been elicited in the adductorpollicis muscle (Merton, 1954; Desmedt, 1958;Slomic et al., 1968), in frontalis, first dorsalinterosseous, lateral gastrocnemius, and extensordigitorum brevis (McComas and Thomas, 1968),and in extensor hallucis brevis (Sica andMcComas, 1971) but activation of an entiremuscle provides little data about its constituentmotor units. In biopsy specimens of differenthuman muscles fast and slow twitch fibres havebeen found (Brust and Cosla, 1967; Ebersteinand Goodgold, 1968) but in vivo studies of indi-vidual motor units have been few. In soleus,gastrocnemius, tibialis anterior, and biceps andtriceps brachii, Buchthal and Schmalbruch(1970) demonstrated a correlation between con-traction time and fibre type, and in extensor

825

hallucis brevis (EHB) Sica and McComas (1971)were able to distinguish two groups of motorunits (fast and slow twitch units) by averagingthe mechanical response to threshold stimulationof the motor nerve. Perhaps the most sophisti-cated in vivo technique has been that of Milner-Brown et al. (1973a, b, c) who have determinedthe contractile properties of individual motorunits in the first dorsal interosseous muscle (IDI)of human subjects during voluntary contraction.

This study presents preliminary observationson the isometric properties of the abductordigiti minimi muscle (ADM) in human subjectsand documents the method used for studying thecontractile properties ofADM. The total muscletwitch response, the response to tetanization,and the properties of individual motor unitsactivated by threshold stimulation have beendetermined in six normal subjects. These data

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

David Burke, Nevell F. Skuse, and A. Keith Lethlean

FIG. 1. The frame usedforrecording isometric contrac-tions ofADM. For simplicitythe three padded compressivemetal plates which fix thearm and hand are not shown.These plates are screweddown onto the limb from thesuperstructure by the threescrews illustrated, thepositions of which can bealtered.

will serve as control observations for a subse-quent paper on paramyotonia congenita (Burkeet al., 1974).

METHODS

The data to be presented were obtained from 18experiments in six normal subjects aged 22 to 45years. Two of the authors were studied in greaterdetail because of availability and because intra-muscular recordings and tetanization as used in someof the studies were not readily acceptable to volun-teer subjects.The isometric contraction of ADM was studied

with the right hand rigidly immobilized in a metalframe designed for use in myasthenic patients by DrGeorge Preswick (Fig. 1). Subjects lay supine on acouch with the right arm abducted at the shoulder,extended at the elbow, and immobilized by a leatherstrap fixed around the upper forearm and by threepadded metal plates which were screwed down fromthe overhead supporting frame. Two of these com-pressive plates fixed the limb at wrist and palmlevels, while a third plate immobilized the index,middle, and ring fingers, which were taped togetherand which were additionally isolated from the fifthfinger by an adjustable vertical plate. Flexion move-ments of the fifth finger in the vertical plane wereprevented by a retaining ring over the distal end ofthe middle phalanx. The retaining ring was free tomove laterally along a horizontal slot in the base-

plate of the frame so that abduction movements ofthe fifth finger were unimpeded. Such lateral (orabduction) movement was then restricted by a ten-sion transmitting rod with an 'L' shaped end whichwas fitted over the proximal phalanx at the meta-carpophalangeal joint. The other end was fixed bytwo nuts into a slot in a mild steel bar of length 15cm such that the length of the tension transmittingrod could be altered where necessary. A four-armstrain gauge bridge was constructed by bonding fourPhillips etched foil strain gauges (Type PR 9831) tothe steel bar and this bridge was energized by aTektronix carrier amplitier type 3C66. The length ofthe tension transmitting rod was adjusted to impartslight stretch to ADM, thus 'pretensioning' thestrain gauge bridge. Compliance of the steel bar waslow, such that a 1 kg weight attached to the end ofthe bar produced a displacement of less than 1 mm.Recording were therefore virtually isometric. Thenatural (resonant) frequency of the metal arm of thestrain gauge was 100 Hz in the unloaded state.

Stimuli of 0-2 ms duration were delivered by twoDevices stimulators coupled in series and triggered bya Devices Digitimer. For studies of the total muscletwitch characteristics supramaximal stimuli weredelivered by surface electrodes 2 cm apart taped tothe wrist over the ulnar nerve. Stimulus voltage wasincreased for tetanization to exceed maximal levelsby approximately 50%/, thus hopefully avoiding theeffects of the excitability changes described byBergmans (1970). For studies of threshold motor

826

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

Isometric contraction of the abductor digiti minimi muscle in man

unit twitches, threshold stimuli were also deliveredto the motor point of ADM using a pencil-shapedmonopolar stimulating electrode. For this purpose,this electrode was held in a stereotactic micro-manipulator so that changes in position of the pointof stimulation could be readily produced.The output of the strain gauge bridge was dis-

played on a Tektronix 564 storage oscilloscope andphotographed when necessary with a Polaroidoscilloscope camera. In some experiments theelectromyogram (EMG) of ADM was recorded bysurface electrodes taped over ADM, one directlyover the motor point and the other over the meta-carpophalangeal joint. The recorded muscle actionpotential was amplified by a Tektronix 122 low levelpreamplifier (high and low frequency filters of 1000Hz and 0-2 Hz respectively) and displayed on asecond channel of the storage oscilloscope.For studies of threshold motor units, the strain

gauge output was monitored on the storage oscillo-scope and fed into a ND801 Enhancetron 1024 inwhich between 100 and 500 consecutive contractionswere averaged. The read-out of the Enhancetronwas displayed on a second oscilloscope and photo-graphed with a Polaroid oscilloscope camera. Be-cause arterial pulsation often produced a forcewhich exceeded that produced by contraction ofthreshold motor units, threshold stimulation wasperformed during arterial occlusion produced by asphygmomanometer cuff inflated to 200 mmHg andplaced above the elbow. Efforts were made to mini-mize the duration of circulatory arrest to less thanfive minutes, since Buchthal and Schmalbruch (1970)have demonstrated that prolonged hypoxia elimi-nates slower contraction times within a muscle, thusshifting the peak of the spectrum to fast contractiontimes. Threshold stimuli were therefore delivered atthe rate of 2 Hz. Such low frequency stimulationevokes the 'positive staircase' phenomenon, butwith the technique used the slight inaccuracy pro-duced was constant for any one patient, and thisfactor has therefore been ignored (see Results sec-tion for further justification).

Experiments were all performed in an air-conditioned laboratory. Skin temperature wasmeasured over the belly of ADM in some experi-ments using an Ellab electronic thermometer andsurface temperatures of 33-35°C were maintained.

BRIDGE CALIBRATION The output of the straingauge bridge was calibrated in arbitrary units andwas found to be linear over the entire working range,such that 1 unit was equivalent to 0-476 g weight or4-66 milliNewtons (mN). More accurately, themeasurements of contractile strength of ADMshould be expressed not as force (in grammes weight

or in mN) but as torque (in Newton-metres) becausethe output was measured at a point on the proximalphalanx 1 cm distal to the insertion of ADM. Thus abridge output of 1 unit may be considered equivalentto a torque of 4-66 x 10-5 Newton-metres (Nm).Since other authors have not expressed contractilestrength in units of torque, the available data onother muscles cannot be compared with those to bepresented for ADM.

RESULTS

MECHANICAL PROPERTIES OF ABDUCTOR DIGITIMINIMI The torque produced by supramaximalcontraction of ADM ranged from 2-33 x 10-3Nm to 6'24 x 10-3 Nm (mean 3-77 x 10-3 Nm).

I-<----- C.T. - . -4-- 1/2 R.T. -

FIG. 2. The isometric twitch contraction of ADM.Twitch torque and the evoked muscle action potentialare shown at two different sweep speeds. Three con-tractions are superimposed on each occasion. Note thesmall F wave, indicated by the arrow. Contractiontime (C.T.) and half relaxation time (< R.T.) are asindicated. Calibration: vertical: 2-33 x 10- Nm,2 mV; horizontal: 50 ms and 20 ms for upper andlower traces respectively.

827

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

David Burke, Nevell F. Skuse, and A. Keith Lethlean

The contraction was always of smooth contour(Fig. 2) provided that the retaining ring fittedfirmly around the finger. The twitch contractiontime, measured from the onset to the peak ofmuscle contraction, was 60-68 ms (mean 63 ms).The half-relaxation time, measured from peakcontraction to the point at which contractiletorque had decayed to 5000 of the peak level,ranged from 60 to 80 ms (mean 71 ms).A small F wave was commonly recorded in the

EMG in response to supramaximal stimulation,the amplitude of this wave being less than 6% ofthe amplitude of the direct motor (M) response.Conceivably, the presence of the F wave couldhave prolonged the relaxation phase slightly, butin the adductor pollicis Slomic et al. (1968) haveshown that this factor is negligible and it hastherefore been ignored.

Repetitive stimulation and tetanization wereperformed in three subjects using supramaximalstimulus levels. Repetitive stimulation at 3 Hzfor two minutes induced the 'positive staircase'phenomenon as described by Desmedt andHainaut (1968), Slomic et al. (1968), andDesmedt et al. (1973). Twitch torque progres-sively increased during stimulation by 33-450.Contraction time decreased by 8-10% withinthe first 10-20 stimuli and thereafter remainedrelatively constant. The rate of development oftwitch torque was therefore greatly increased.There was no concomitant change in the evokedmuscle action potential so that these results areconsistent with those reported by other authors.The frequency at which a fused tetanus was

produced was 25-30 Hz. Peak tetanic force was4-66-6-9 x 10-2 Nm and the twitch-tetanusratio was 0 09. The tetanic contraction was wellmaintained over short periods (up to 10 s) but agradual decline in peak torque occurred withprolonged tetanization (Fig. 3a) so that in threenormal subjects (seven observations) the con-traction level had fallen by an average of 31%after 30 s. There was no concomitant decreasein the evoked muscle action potential to eachstimulus. The amplitude of the negative phase ofthe action potential was virtually unchanged butthe area of the negative phase as measured byplanimetry was actually increased due to widen-ing and slurring of the action potential (Fig. 3b).It may be concluded that the 30%° decline intetanic torque was due to intrinsic muscular

A

B

FIG. 3. Tetanization at 50 Hz for 30 s at supra-maximal voltage. A. The mechanical response de-creases by 30-35°/ within 30 s. B. The evoked muscleaction potential. In each of the upper and lower tracesthree isolated contractions are superimposedbefore andafter tetanization respectively. In the middle trace thechange with tetanization is seen. The potential evokedby the first stimulus of the tetanus is shown and every3 s throughout the tetanus a further potential has beenrecorded. Note the progressive widening of the poten-tial without loss of amplitude of the negative phase.The area of the negative phase increases progressively.Calibrations: vertical: 2 33 x 10-2 Nm for A, 3 mVfor B; horizontal: 10 s for A, 4 ms for B.

mechanisms rather than to fatigue at the myo-neural junction.

In two subjects tetanization was continued forfive minutes. Tetanic contraction was again wellmaintained for the initial 10 s, decreased to 65-700/o by 30 s, but decreased further to 25% overthe ensuing 90 s. Thereafter a plateau level wasmaintained for the final three minutes with onlyslight further decrease (Fig. 4). As measured byplanimetry, the area of the negative phase of theevoked muscle action potential was again largerafter tetanization for 30 s due to slurring of the

828

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

Isometric contraction of the abductor digiti minimi muscle inman8

FIG. 4. Tetanic torque durinig prolonged tetanizationat 50 Hz at supramaximal voltage for five minutes:each oscilloscope sweep lasts 50 s. The sweep was

triggered four times durinig tetanization to record thestart, the finish, and two different stages duringtetanization, 60 s being allowed to elapse betweensuccessive sweeps. Maximal fatigue occurs earlybetween 30 anid 90 s and thereafter only slight furtherfatigue occurred. Calibrationis: vertical: 233 x 10-2

Nm; horizonital: 10 s.

potential, but after tetanization for 60 s theamplitude of the negative phase decreasedgreatly so that the area of the negative phase was

4400 of its initial value. A slight further decreaseoccurred with continued stimulation. Thus onlya part of the fatigue produced by prolongedtetanization can be attributed with certainty tointrinsic muscular mechanisms.

MOTOR UNIT COUNTS An attempt was made toestimate the number of motor units in ADM bydetermining the average EMG increment pro-duced by progressively larger stimulus levels, as

originally described for extensor digitorumbrevis (EDB) by McComas et al. (1971) and as

recently applied to the thenar muscles by Brown(1972, 1973). Such studies suggested that ADMcontained between 100 and 150 motor units.However, even with careful placement of surfaceelectrodes, it proved impossible to record selec-tively from ADM, although the pick-up from theother hypothenar muscles could be reduced toless than one-third the voltage recorded fromADM. The motor axons of lowest threshold inthe ulnar nerve at the wrist were found to inner-vate hypothenar muscles other than ADM in atleast two subjects, so that low voltage EMGactivity was picked up by the electrodes over

ADM but no muscle twitch was recorded by thestrain gauge.

In Brown's experiments the surface electrodeswere positioned over the abductor pollicis brevisand the calculated motor unit counts were con-sidered to be estimates of motor units in thethenar muscle group. In our experiments, whilesurface electrodes over ADM picked up theactivity of other hypothenar muscles (flexordigiti quinti brevis and opponens digiti quinti),such activity always appeared somewhat attenu-ated. EMG increments due to recruitment ofmotor units to these muscles would appearfalsely small, and, as a result, discrepancies inmotor unit counts could arise. Moreover, it can-not be assumed that the action potentials evokedby activation of motor units from the threehypothenar muscles would sum algebraically, asrequired by the technique of McComas et al.(1971), who pointed out that their technique wasnot reliable for muscles other than extensor digi-torum brevis of the foot.

Estimation of the number of motor units inADM was also attempted by dividing the torquerecorded on supramaximal stimulation by theaverage size of torque increments. With supra-liminal stimuli a number of units commonlypresented with similar electrical thresholds, andonly with near-threshold stimuli could individualunits be activated. Such estimates suggestedbetween 50 and 100 motor units for ADM butthis figure is likely to be an underestimate becausedifficulty was experienced in distinguishingindividual torque increments. However, as shownin the next section, single motor units produceda torque of generally less than 4 66 x 10-I Nmand never more than 9 32x 10-5 Nm, and divi-sion of the total muscle twitch torque by thesevalues provides a motor unit estimate of similarmagnitude to the calculated estimate.

THRESHOLD MOTOR UNITS With liminal stimulito the ulnar nerve at the wrist or to the motorpoint of ADM, motor unit twitch contractionscould be identified which behaved in an 'all-or-none' manner and which therefore conform tothe criteria of single motor unit contractions.The surface anatomy of the motor point wasmapped in six subjects using monopolar per-cutaneous stimulation. The motor point ap-peared to be shaped like a narrow-based isosceles

a

829

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

David Burke, Nevell F. Skuse, and A. Keith Lethlean

triangle or like the figure 'Y', oriented longi-tudinally with the base of the triangle or the openarms of the U pointing distally. Considering thedistortion produced by percutaneous stimula-tion, these results are in reasonable agreementwith the anatomical studies of motor pointanatomy detailed by Desmedt (1958) who foundan elongated 'U'-shaped motor point.

Threshold contractions, whether induced bypercutaneous stimulation at the wrist or over themotor point, produced a torque of usually lessthan 4-66 x 10-5 Nm and always less than9-32 x 10-5 Nm. To determine the time course ofsingle motor unit twitches, 100-500 consecutivethreshold stimuli at 2 Hz were summed by afixed programme averaging computer duringarterial occlusion. Care had to be taken at eachstimulus site to ensure that a stimulus whichappeared threshold for only one motor unitduring single stimuli did not activate more thanone unit during repetitive stimulation at 2 Hz. Attimes the stimulus level had to be set so that thethreshold unit was not always activated. It wastherefore not possible to determine the twitchtorque of threshold units accurately by feeding acalibration pulse into the averaging input (asdone by Sica and McComas, 1971), since withsuch 'all-or-none' behaviour the pulse wouldhave been summed at times when no contractionwas recorded.Low frequency repetitive stimulation even at

threshold levels induced the 'positive staircase'phenomenon in the activated motor units, withresultant increase in twitch torque and de-creases in contraction and half-relaxation times.Changes in twitch torque are relatively un-important to the present study, since absolutevalues for twitch torque cannot be given. De-creases in twitch times are more relevant butthese changes occur quite rapidly, generally inthe first 10-20 stimuli, after which a plateau levelis maintained (Desmedt et al., 1973; see above).Care was taken in the present experiments tostart repetitive stimulation 2040 s before aver-aging, and during this period stimulus levelswere determined and computer balance levelswere adjusted. These procedures had to becarried out before each averaging run so that theplateau level of twitch times was reached on eachoccasion before averaging commenced. Thevalues obtained are therefore all slightly lower

A

FIG. 5. Threshold motor unit twitches. A. Surfacestimulation at the motor point. Upper trace: averaged200 sweeps, contraction time 46 ms. Lower trace:averaged 250 sweeps, contraction time 76 ms. B.Intramuscular stimulation at the motor point. Uppertrace: averaged 200 sweeps, contraction time 54 ms.Lower trace: averaged 200 sweeps, contraction time74 ms. In both A and B duration of trace is 250 ms.

than those which would have been recorded withsingle stimuli.A total of 55 threshold contractions was re-

corded in six subjects, and the averaged twitchesof two threshold units are shown in Fig. 5a. Thetwitch contraction times of these 55 thresholdcontractions are shown in Fig. 6a. A unimodaldistribution of these times is seen, unlike the bi-modal distribution reported for EHB by Sica andMcComas (1971). The failure to find a bimodaldistribution of contraction times (1) could re-sult from masking of such a distribution due topooling of data from different subjects; (2) couldbe due to discrepancies in technique such thatsome units were overrepresented, being oflowest threshold at a number of stimulationsites; or (3) could represent the true distributionof twitch times within this muscle. The firstpossibility seems unlikely, since a bimodal dis-tribution was not apparent in the data for indi-vidual subjects. The second possibility alsoseems unlikely, since with a discrete stimulatingelectrode point a number of well-separatedstimulation sites can be found on the relativelylarge surface area of the motor point, and since aspectrum of contraction times was recorded ineach subject. Nevertheless it cannot be deniedthat some units may have been represented morethan once in the histogram.

In an attempt to validate the unimodal dis-

830

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

Isometric contraction of the abductor digiti minimi muscle in man

I

I..

0

E

z

2

4

6

20 40 60 80

(B)

FIG. 6. Contraction times of threshold motor units.(A) Histogram of 55 threshold twitch contractions insix subjects using surface stimulation. (B) Histogramof an additional 30 threshold twitch contractions intwo subjects using intramuscular stimulation.

tribution of contraction times found with percu-taneous stimulation, a bipolar concentric needleelectrode (Disa 9013K0512) was used as stimu-lating electrode in two subjects. This needleelectrode was introduced subcutaneously, tra-versing the motor point in a number of paralleltracks, with threshold stimulation performedat a number of sites along each track. In this way15 threshold contractions were recorded in eachof the two subjects, as illustrated in Fig. 5b. Thecontraction times of these units are graphed inhistogram form in Fig. 6b, and a unimodaldistribution is again seen with a spectrum ofcontraction times very similar to that obtainedwith percutaneous stimulation.The response of threshold motor units to pro-

longed tetanization at 50 Hz was studied in twosubjects. In both subjects liminal stimulation ofthe ulnar nerve at the wrist and similar stimula-tion of the motor point via the bipolar needleelectrode were used. If the threshold unit at a

particular site was first identified using single

FIG. 7. Threshold tetanization at the motor pointn] using intramuscular stimulation. Ten seconds after

the oscilloscope sweep starts a small increment in100 120 ms stimulus voltage produces a threshold tetanic contrac-

tion which fatigues rapidly to 20%0 by 33 s. In earlierU1 oscilloscope sweeps, which are not illustrated, three

small step increments at 10 s intervals did not producecontraction. Calibrations: vertical: 4-66 x 10-5 Nm;horizontal: 5 s.

shock stimulation, it was frequently found thattetanic stimulation at the same voltage level re-

cruited additional units, often after some delay,during the course of an established tetanic con-

traction. Commonly the activation of the addi-tional unit was at first inconstant so that an

oscillating contraction appeared on top of thepreviously established tetanic level. As recruit-ment became consistent, the torque produced bythe tetanic contraction rose to be level with thepeak of the cyclical oscillations. Similarly, failureof activation of a unit was seen as an abrupt stepdown in tetanic torque or as a similar step pre-ceded by cyclical oscillations. Thus the recruit-ment of an additional unit or the failure ofactivation of a motor unit produced characteris-tic changes in torque which could be readilyrecognized. The technique adopted consisted ofcontinuous stimulation at 50 Hz, initially at sub-threshold levels, with small step-wise incrementsin stimulus voltage every 10 s until a thresholdtetanic contraction was produced. In this way, itwas possible to maintain the threshold tetaniccontraction for some minutes without recruitingadditional units. Nine threshold tetanic contrac-tions have been studied. In eight, the level oftetanic torque fatigued within 60-90 s to 1420%of the peak level (Fig. 7). In one, torque de-creased to 7500 by 60 s and to 5000 after fiveminutes.

(A)9

7

5

3 II

831

I

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

David Burke, Nevell F. Skuse, and A. Keith Lethleani

DISCUSSION

As demonstrated in this paper, ADM can befunctionally isolated from other muscles of thehand so that its isometric responses can bereadily studied, but it is not possible to correlateaccurately mechanical properties with the surfaceEMG response, since this latter does not reflectactivity of ADM alone. The results obtainedappear quite compatible with those reported forother intrinsic muscles of the hand. Stimulationat the motor point at threshold appears to be a

reasonable method for obtaining single motorunit twitches provided that adequate care is takento minimize the difficulties detailed earlier. Itmust be conceded that electrical stimulation is anartificial method of activation compared withvoluntary contraction, as has been used byMilner-Brown et al. (1973a, b, c), but the twotechniques are essentially complementary, sinceeach provides data not obtainable with theother.

In the cat a number of physiological para-

meters-namely, twitch contraction time, tetanicfusion frequency, and sensitivity to fatigue duringprolonged tetanization-have been used todifferentiate motor unit types and to correlatephysiological type with histochemical profile(Burke et al., 1971). This paper represents an

attempt to extend such studies to man. The re-

sponse of ADM to prolonged tetanization sug-

gests that at least some of the motor units ofADM are not resistant to fatigue. The form ofprolonged tetanus used was presumably a more

exhausting form of activation than that used inthe cat by Burke et al. (1971) who tetanizedindividual motor units at 40 Hz for only 330 msout of every second. Fatigue-resistant motorunits were defined by these authors as thosewhich maintained their tetanic tension in excess

of 7500 of the initial level after tetanization fortwo minutes. Fast-twitch, fatigue-resistant unitsultimately fatigued, but only with furthertetanization. Slow-twitch units did not fatigueeven with prolonged activation, while fast-twitch, fatigue-sensitive units fatigued rapidlyafter activation for approximately 60 s. That thedegree of fatigue reported for ADM in thepresent study was not wholly attributable tofailure of neuromuscular transmission is indi-cated by the initial lack of decrement in the

evoked muscle action potential. Presumably,such mechanical fatigue as was recorded at thisstage resulted from intrinsic muscular mechan-isms, either from failure of excitation-couplingor of the contractile mechanism itself.

The unimodal distribution of threshold motorunit contraction times suggests that humanADMcontains a homogeneous population of motorunits. On the basis of contraction times, Sica andMcComas (1971) have suggested that humanEHB contains two populations of motor units,but it would be unlikely if these findings in amuscle of the foot could be applied to an intrinsicmuscle of the hand which must perform differentmotor functions. Certainly in various upper andlower limb muscles of man, Buchthal andSchmalbruch (1970) have reported unimodaldistributions of contractions times, there beingno evidence of different muscle fibre groups withdistinctive contraction times, and more recentlyMilner-Brown et al. (1973b) reported a normaldistribution of contraction times for IDI, therebeing no clear distinction between fast and slow-twitch units. However using the myofibrillarATP-ase reaction, which in the cat reflects thephysiological difference between fast and slowtwitch fibres, Johnson et al. (1973) have dis-tinguished two populations of muscle fibres inhuman ADM, each comprising approximately5000 of the sample. The apparent discrepancybetween histochemical properties reported byJohnson et al. (1973) and the physiologicalproperties documented in the present study isunexplained. Nevertheless, Johnson et al. (1973)also reported two histochemically distinguishablegroups of muscle fibres in those muscles studiedphysiologically by Buchthal and Schmalbruch(1970) and by Milner-Brown et al. (1973b) andfound by these authors to have unimodal dis-tributions of contraction times. The range ofcontraction times reported here for ADM, 40-100 ms, agrees well with the 30-100 ms rangereported by Milner-Brown et al. (1973b) for IDI,another ulnar-innervated intrinsic muscle of thehand.The results of tetanic stimulation using thresh-

old stimulus levels suggest that at least some ofthe motor units in ADM conform with the fast-contracting, rapidly-fatiguing group described inthe cat by Burke et al. (1971), but it also seems

832

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

Isometric contraction of the abductor digiti minimi muscle in man

likely that some units are fatigue-resistant. Theeffect of recruitment or drop-out of additionalunits was abrupt and readily recognized so thatthe progressive decline in tetanic contractioncannot be attributed to electrode displacementwith resulting failure to activate a particularmotor unit. It can be argued that thresholdtetanic stimulation may activate more than onemotor unit, but this objection can be raised toall studies involving threshold stimulation, andeven were this so the progressive decrease intetanic torque indicates that the involved motorunit or units was or were undergoing gradualcontractile failure.Thus the results of this study indicate that the

motor unit population of ADM is homo-geneous with respect to contraction time, andcomprises fatigue-sensitive units and some unitswhich, while relatively fatigue-resistant, doundergo some contractile failure with prolongedactivation. These parameters are consistent withtype FF and type FR motor units as described byBurke et al. (1971). Is there any necessity topostulate the presence in ADM of slow contrac-ting non-fatiguing motor units (type S of Burkeet al.)? At face value, the longer contractiontimes reported here would be consistent with thepresence of slow-twitch units especially if humancontraction times are directly comparable withthose of the cat. But such is not the case. In thecat contractile force is measured directly, butvalues obtained in intact man reflect the inertiaof non-contractile tissues. The torque recordeddepends on the development of force at a jointwhich is moved by the appropriate muscle and sorepresents only an indirect approximation ofcontractile force. Thus a systematic distortion ofthe contraction time is introduced in the humansubject by the recording technique. By relatingthe range of threshold motor unit contractiontimes to the total muscle contraction time it canbe appreciated that at least half the population ofthreshold units have contraction times longerthan those of the total muscle. It seems unlikely,therefore, that motor units with slower contrac-tion times have been systematically neglected bythe technique of identification and recording. Itmay therefore be concluded that if slow-twitchunits exist in ADM they form one end of anormal spectrum of motor units and do not con-stitute a functionally distinct group.

The authors would like to acknowledge the construc-tive advice given by Professor J. W. Lance. Illustra-tions were photographed by the Department ofMedical Illustration, University of New SouthWales.

REFERENCES

Bergmans, J. (1970). The Physiology of Single Human NerveFibres. Vander: Louvain.

Brust, M., and Cosla, H. W. (1967). Contractility of isolatedhuman skeletal muscle. Archives of Physical Medicine andRehabilitation, 48, 543-555.

Buchthal, F., and Schmalbruch, H. (1970). Contractiontimes and fibre types in intact human muscle. ActaPhysiologica Scandinavica, 79, 435-452.

Brown, W. F. (1972). A method for estimating the number ofmotor units in thenar muscles and the changes in motorunit count with ageing. Journal ofNeurology, Neurosurgery,and Psychiatry, 35, 845-852.

Brown, W. F. (1973). Thenar motor unit count estimates inthe carpal tunnel syndrome. Journal of Neurology, Neuro-surgery, and Psychiatry, 36, 194-198.

Burke, D., Skuse, N. F., and Lethlean, A. K. (1974). Thecontractile properties of the abductor digiti minimimuscle in paramyotonia congenita. Journial of Neurology,Neurosurgery, and Psychiatry. (In press.)

Burke, R. E., Levine, D. N., Zajac, F. E., 111, Tsairis, P., andEngel, W. K. (1971). Mammalian motor units: physio-logical-histochemical correlation in three types in catgastrocnemius. Science, 174, 709-712.

Desmedt, J. E. (1958). Methodes d'etude de la fonction neuro-musculaire chez l'homme. Acta Neurologica et PsychiatricaBelgica, 58, 977-1017.

Desmedt, J. E., Emeryk, B., Hainaut, K., Reinhold, H., andBorenstein, S. (1973). Muscular dystrophy and myastheniagravis. In New Developments in Electromyography andClinical Neurophysiology, Vol. 1. pp. 380-399. Edited byJ. E. Desmedt, Karger: Basel.

Desmedt, J. E., and Hainaut, K. (1968). Kinetics of myo-filament activation in potentiated contraction: staircasephenomenon in human skeletal muscle. Nature, 217, 529-532.

Eberstein, A., and Goodgold, J. (1968). Slow and fast twitchfibers in human skeletal muscle. American Journal ofPhysiology, 215, 535-541.

Johnson, M. A., Polgar, J., Weightman, D., and Appleton,D. (1973). Data on the distribution of fibre types in thirty-six human muscles. An autopsy study. Journal ofthe Neuro-logical Sciences, 18, 111-129.

McComas, A. J., Fawcett, P. R. W., Campbell, M. J., andSica, R. E. P. (1971). Electrophysiological estimation ofthe number of motor units within a human muscle.Journal of Neurology, Neurosurgery, and Psychiatry, 34,121-131.

McComas, A. J., and Thomas, H. C. (1968). Fast and slowtwitch muscles in man. Journal ofthe Neurological Sciences,7, 301-307.

Merton, P. A. (1954). Voluntary strength and fatigue. Journalof Physiology, 123, 553-564.

Milner-Brown, H. S., Stein, R. B., and Yemm, R. (1973a).The contractile properties of human motor units duringvoluntary isometric contractions. Journal of Physiology,228, 285-306.

Milner-Brown, H. S., Stein, R. B., and Yemm, R. (1973b).The orderly recruitment of human motor units during

833

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from

834 David Burke, Nevell F. Skuse, and A. Keith Lethleani

voluntary isometric contractions. Journal of Physiology, twitch units in a human muscle. Journal of Neurology,230,359-370. Neurosurgery, andPsychiatry, 34, 113-120.Milner-Brown, H. S., Stein, R. B., and Yemm, R. (1973c).Changes in firing rate of human motor units during Slomic, A., Rosenfalck, A., and Buchthal, F. (1968). Elec-linearly changing voluntary contractions. Journal of trical and mechanical responses of normal and myasthenicPhysiology, 230, 371-390. muscles with particular reference to the staircase phenom-Sica, R. E. P., and McComas, A. J. (1971). Fast and slow enon. Brain Research, 10, v-xii, 1-78.

Protected by copyright.

on February 28, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.37.7.825 on 1 July 1974. Dow

nloaded from