augmentation of the contraction force of human thenar muscles by and during brief discharge trains

ABSTRACT: We investigated the influence of the history of activity on thecontractile properties of abductor pollicis brevis (APB) to define how theforces produced by individual stimuli change within a stimulus train, with aview to clarifying the optimal discharge frequency for force production in brieftrains. Supramaximal electrical stimuli were delivered to the median nerve atthe wrist singly or in trains of 2–5 at various interstimulus intervals (ISIs). Theforce and electromyographic (EMG) responses to trains of n stimuli weredefined by online subtraction of the responses to n � 1 stimuli. The forceattributable to the n th stimulus was normalized to that produced by a singlestimulus. The contraction force produced by 2 stimuli exceeded the forceexpected with linear summation of 2 single twitches by 30–40% at ISIs of2–100 ms. Increasing the number of stimuli resulted in less augmentation ofthe force produced by the last stimulus in the train for ISIs up to 20 ms, butgreater augmentation for ISIs of 50–100 ms. At ISIs of less than 10 ms, thetime to peak force produced by the last stimulus in a 5-pulse train wasdelayed by �100 ms, the peak force produced by that stimulus was lessthan that produced by a single stimulus, and it occurred on the falling phaseof the overall contraction. These properties are best explained by the catch-like property of muscle. This implies that the augmentation of contractionforce due to this property can increase throughout a stimulus train, and is notrestricted to the doublet discharges that have conventionally been studied.We conclude that, with brief discharge trains, maximal forces occur at ISIs of56–75 ms, intervals that are longer than those conventionally associatedwith the catchlike property. Discharge rates of 15–20 HZ appear to beoptimal for force generation by APB during steady contractions.

Muscle Nerve 33: 384–392, 2006

AUGMENTATION OF THE CONTRACTIONFORCE OF HUMAN THENAR MUSCLESBY AND DURING BRIEF DISCHARGE TRAINS

JAMES HOWELLS, BSc, LOUISE TREVILLION, BSc,

STACEY JANKELOWITZ, MB, BCh, PhD, and DAVID BURKE, MD, DSc

Institute of Clinical Neurosciences, University of Sydney and Royal Prince Alfred Hospital,Sydney, NSW 2006, Australia

Accepted 23 September 2005

At the onset of voluntary contractions, motor unitscommonly discharge 2 or 3 impulses at very shortintervals, producing “doublets” or “triplets” with in-terdischarge intervals of 6–10 ms,3,20,28,32 a dischargepattern that should enhance force generation due tothe catchlike property of muscle.6,8,9 Whether or notdoublets or triplets occur, maximal voluntary effortproduces a discharge rate that is greatest early in the

contraction, at 30 Hz or more for the intrinsic mus-cles of the hand, decaying over some tens of secondsto a tonic rate of less than 20 Hz.4 Doublets occurrelatively infrequently during sustained voluntarycontractions, except in patients with neuromusculardiseases,23 and the interval following the doubletdischarge is usually long.15

It has been hypothesized that there is a relation-ship between the contractile properties of muscleand the optimal discharge rate for force produc-tion.21 Recent findings do not support this hypothe-sis, at least for fatigue.11–13 Leaving fatigue aside, theextent to which the contractile properties of humanmuscle are indeed tuned to produce optimal force atspecific interdischarge intervals is not well appreci-ated. In addition, it is not clear the extent to whichthe augmentation of contraction force at relatively

Abbreviations: APB, abductor pollicis brevis; CMAP, compound muscleaction potential; ISI, interstimulus intervalKey words: catchlike property; contraction force; contraction time; musclecontraction; muscle twitchCorrespondence to: D. Burke, Office of Research & Development, Collegeof Health Sciences Medical Foundation Building–K25, University of Sydney,Sydney, NSW 2006, Australia; e-mail: [email protected]

© 2006 Wiley Periodicals, Inc.Published online 24 January 2006 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/mus.20469

384 Mechanical Properties of Muscle MUSCLE & NERVE March 2006

long intervals (50–100 ms) is necessary to compen-sate for the timing of the twitch on the falling phaseof the preceding contraction, rather than closer toits peak, as would occur with shorter intervals.

The present study was undertaken to documentthe optimal discharge intervals for force productionby the thenar muscles using brief trains of stimuli tothe median nerve. This muscle group was chosenbecause it is reasonably homogeneous, whether stud-ied histochemically (73% type I muscle fibers17) orphysiologically (72% fatigue-resistant motor units29).In addition, it was the muscle used in the studies ofMarsden et al.,21 and in microneurographic studiesof motor unit properties.5,14,29,30 The present studyfocuses on the change in contraction force evokedby different stimuli within a train rather than on theposttetanic changes or the staircase phenomenon.

METHODS

Twenty experiments were performed on fourhealthy volunteers (aged 33–60 years; 2 males), whogave informed written consent for the procedures,which had the approval of our institutional reviewboard and were in accordance with the Declarationof Helsinki. The reproducibility of the findings wasestablished by repeated studies in the same subjects(the authors), rather than by recruiting more sub-jects, because truly supramaximal stimuli are painfulwhen delivered in recurring trains of varying length.Three of the subjects were between 33 and 38 yearsof age. However, the findings were qualitatively sim-ilar in the 60-year-old (male) subject, indicating thatthe phenomena studied are not appreciably age-dependent, at least in a healthy population.

Experimental Posture. Recordings were made withthe right forearm supported, elbow flexed and thehand semipronated, resting on the ulnar border,with the fingers curled around a solid plastic half-cylinder (120-mm diameter) made of ultra-high-density polyethylene, on which a force transducerwas mounted. The interphalangeal joint of thethumb was secured to the force transducer to allowthe recording of twitch contraction force in the di-rection of abduction, the abduction contractioncompressing the thumb further against the trans-ducer. The hand was in a comfortable relaxed posi-tion with wrist extended by 30° and the thumbaligned above but opposing the index finger, aswould occur if the hand was holding a glass, anoptimal posture for APB function.1,2,16,18,24,27 To con-firm that this posture was also appropriate for forceproduction, twitch force was recorded with the the-

nar muscles stretched (much as is illustrated in Fig.1 of ref. 30) and shortened (see Results). Stimulus–response curves for twitch force were recorded withthe wrist and hand in the experimental position, in aposition where the muscle was stretched, and in athird position where the muscle was shortened.Force production was maximal when the thumb wasin the experimental position for all four subjects.During experiments, care was taken not to alter theposition or orientation of the force transducer onthe thumb. Compound muscle action potentials(CMAPs) were recorded using surface electrodes ap-plied over the abductor pollicis brevis (APB) in abelly–tendon arrangement. A ground electrode waspositioned in the palm of the hand. The experi-ments were performed with the test muscles relaxed,as confirmed by the absence of EMG activity.

Stimulation. Electrical stimuli of 0.2-ms durationwere administered from a computer-controlled, bi-phasic, constant-current source via saline-soaked but-ton or gelled Ag–AgCl surface electrodes (UnilectECG electrodes, Maersk Medical, Ltd., Gloucester-shire, UK; 1 cm diameter) to the median nerve, withthe cathode at the wrist crease and the anode 4 cmmore proximal. To minimize subject discomfort, theanode was relocated in some experiments to thedorsal aspect of the wrist so that the median nervewas stimulated through rather than along the wrist.Different stimuli were delivered in a regular se-quence by computer (see later), with the responsesto different stimuli recorded sequentially on differ-ent channels. Supramaximal stimuli were used in allexperiments, and this was confirmed by the findingthat increasing the stimulus intensity by up to 100%at the 2-ms and 2.5-ms interstimulus intervals (ISIs)did not result in an increase in force or CMAPamplitude with either the single or multiple stimuli.

Measurement of Force. Twitch contraction forcewas measured using a 250-N load cell (XTran K2,250N, Applied Measurement, Victoria, Australia).CMAP and twitch force measurements were ampli-fied and digitized using a PC data acquisition boardat sampling rates of 10 and 5 kHz, respectively.Stimulus delivery and data sampling were controlledby software written in qtrac (© Professor HughBostock, Institute of Neurology, London).7 The am-plitudes of the CMAP and the evoked contractionforce were measured from baseline to peak, and thearea (force–time integral) of the twitch contractionforce was also measured.

Mechanical Properties of Muscle MUSCLE & NERVE March 2006 385

Study Protocols. In the first protocol supramaximalstimuli were applied alternately as single stimuli or inpairs, each stimulus (stimulus pair) recurring onceper second. The ISI was increased from 2 ms to 200ms in a logarithmic progression over 18 intervals.The CMAP and force responses to the second stim-ulus of the doublets were measured after onlinesubtraction of the response to the preceding singlestimulus (as in Fig. 1B).

The second protocol was designed to determinethe time-course of the augmentation of contractionforce by trains of stimuli. The test supramaximalstimulus was conditioned by trains of 2 or 5 supra-maximal stimuli. The stimulus sequences were deliv-ered regularly at intervals of 1.5 s. The intervalsbetween stimuli in the conditioning train were 5, 10,or 50 ms. These intervals were selected because dou-blets were found elsewhere to produce peak force atISIs between 5 and 15 ms,30 whereas the augmenta-tion during the trains of protocol 3 (below) oc-curred at 50–100 ms. The test stimuli were delivered2.5, 5, 7.5, 10, 20, 50, 100, 200, 400, and 600 ms afterthe final stimulus of the conditioning train. The testresponse was measured online after subtracting theresponse to the immediately preceding set of stimuli.These conditioning–test intervals were chosen to de-termine the interval at which the muscle contractionwas no longer influenced by previous activity.

The third protocol was designed to measure thecontraction force produced by the last stimulus of abrief train of supramaximal stimuli as the number ofstimuli was increased from 1 to 5. The stimuli weredelivered in a regularly alternating sequence at in-tervals of 1.3 s. The ISI was increased in 16 steps,from 2 ms, as in the first protocol, but not beyond100 ms because of the greater number of stimuli.The CMAP and twitch force produced by the laststimulus was defined by the online subtraction of theresponse to the immediately preceding train: forexample, the response to the last stimulus of the5-pulse train was obtained by subtracting the re-sponse to the 4-pulse train.

Controls. The effects of the staircase phenomenonwere minimized (although probably not eliminated)by normalizing each trace to the twitch produced byan unconditioned supraxamimal stimulus at eachconditioning–test interval. In addition, data wererecorded for each subject, both increasing the ISIfrom 2 ms to 200 ms and, in a separate experiment,decreasing the ISI from 200 ms to 2 ms. Throughoutall experiments the skin temperature at the rightwrist was monitored and maintained at or above32°C using a thermistor-controlled heating pad,

placed under the forearm, and a polar fleece gloveto wick away perspiration and avoid evaporative cool-ing. Experiments were repeated for each subject,and the data for each subject averaged. Group dataare presented as the mean of the subject averages �standard error. The significance of differences wasassessed using Student’s t-test or regression analysis.

RESULTS

The twitch force produced by an unconditionedsupramaximal stimulus increased in size by 21.9 �5.4% over the first 5 minutes of each experiment,and then decreased by 7.1 � 3.3% over the next 10minutes. The enhancement of the control twitch wasaccompanied by a decrease in contraction time from98.9 � 1.4 ms to 93.1 � 0.6 ms. These changesrepresent the “staircase phenomenon,” the effects ofwhich were minimized by recording the “control”unconditioned twitch for each stimulus cycle andnormalizing the data to this contraction. It was alsohoped that this control would minimize the effectsof fatigue. The fact that the first set of experimentswas repeated in reverse sequence in the same sub-jects and the normalized response curves were iden-tical provides some support for the validity of thiscontrol.

All experiments were performed on relaxed mus-cle, with the hand positioned so that the thenarmuscles were at or near the optimal length for forceproduction.1,2,16,18,24,27 Accordingly, passive abduc-tion of the thumb from the test position (shorteningthe thenar muscles) decreased twitch force by 30 �6.3% (n � 4 subjects). Passive stretch decreasedtwitch force by a comparable extent (29.2 � 2.7%,n � 4 subjects).

Time-Course of Twitch Augmentation in Doublets.

The contraction force produced by a test stimuluswas enhanced significantly when conditioned by anidentical stimulus (P � 0.0006; Fig. 1D). The forceproduced by the test stimulus was defined (Fig. 1B)by subtraction of the response to an unconditionedstimulus. The distortion of the force profile and theprolongation in the latency to peak force are clearerin Figure 1C, where the test stimuli have beenaligned. At short conditioning–test intervals, thetime to peak force (equivalent to twitch contractiontime) was prolonged by �50 ms (see later). Theaugmentation of force was, on average, 31.1 � 0.8%for intervals between 5 ms and 75 ms (Fig. 1D) anddid not differ significantly across these intervals(R2 � 0.1087; P � 0.3409). This greater forcecannot be explained by the corresponding


changes in the CMAP, which was depressed to 56%at the 2-ms interval and recovered slowly over 10ms. The findings were essentially similar whetherthe amplitude or the area of the test twitch wasmeasured (Fig. 1D). These changes in contractionforce cannot be explained by changes in the ex-citability of motor axons at the site of stimulationbecause the stimuli were supramaximal (and be-cause their pattern is quite different from thepattern of recovery of axonal excitability followinga single discharge).19,31

Time-Course of Changes in Twitch Force Following

Stimulus Trains. The force produced by a test stim-ulus was measured after conditioning trains contain-ing 2 or 5 supramaximal stimuli (Fig. 2A and B,respectively), using conditioning–test intervals of2–600 ms. In the conditioning trains, the intervalbetween the stimuli was 5 ms, 10 ms, or 50 ms. Withboth conditioning trains, whatever the ISI, the twitchproduced by the test stimulus was maximal at inter-

vals of 50–100 ms. With trains of 5 stimuli at 50-msintervals, the force produced by the test stimulus wasgreater at all conditioning–test intervals, althoughmaximal at 50–100 ms. An implication of this find-ing is that an extra discharge interpolated in a trainat 20 Hz can produce a significant enhancement ofcontraction force, much as has been described forthe catchlike property of muscle. It is notable that, inall instances, the augmentation of the test twitchlasted less than 400 ms.

Augmentation of Response to the Last Stimulus in

Trains of Different Length. The force produced bythe last stimulus in trains of 2, 3, 4, and 5 supramaxi-mal stimuli is shown in Figure 3A for different inter-stimulus intervals. The curve for the second of 2supramaximal stimuli was relatively flat; however,with more stimuli, the force became significantly lessthan that of control at short intervals and signifi-cantly greater than control at long intervals (P �0.015 at 5 ms and P � 0.0005 at 56 ms for the

FIGURE 1. Contraction forces produced by doublets in a single subject. (A) Force profiles for twitches produced by two supramaximalstimuli separated by ISIs of 2 ms to 100 ms. (B) Component of force produced by the second stimulus (obtained by online subtractionof the response to a single stimulus from the response to the doublet). Note the greater contraction force produced by conditioned stimuli(the dotted line indicates the peak force produced by a single supramaximal pulse). (C) Component of the force profile produced by thesecond stimulus of the doublet, aligned to the second stimulus. Note the decrease in latency to peak as the ISI increases. There wasincomplete elimination of the first twitch contraction for the 75-ms interval (B) and, as a result, the relevant profile in C is more prominentthan those for the adjacent ISIs. (D) Amplitude (open symbols) and area (filled symbols) of CMAP (circles) and twitch force (squares)produced by the second stimulus of a doublet. This equates to a recovery cycle where both the test and conditioning stimuli aresupramaximal. Data are mean � SEM (n � 4).


5-stimulus train). A clear peak emerged at the 56-msinterval. Except for intervals of �5 ms, the change incontraction force due to the last stimulus in the traincould not be explained by a change in the CMAP.

As shown in Figure 3A, the augmentation of theforce produced by the last stimulus in the trainincreased with the number of stimuli at intervals of32–75 ms. This indicates that the augmentation can-not be attributed to the elasticity of the noncontrac-tile elements, because the effects of muscle stiffnessshould be the maximal for the second stimulus andless for subsequent stimuli in the train.

Latency Prolongation. The time to the peak of forceproduced by the last stimulus in a train was pro-longed by the preceding contraction, more so atintervals of 4–10 ms (where it reached 100 ms, seeFigs. 1C and 4A) than at intervals of 50–100 ms.However, the latency to the maximal force producedby the train was shorter (Fig. 4B), because, at longerinterstimulus intervals, the peak of force due to thelast stimulus in the train fell on the decay phase offorce due to the preceding stimuli. Thus, while theincrease in force was greatest at 56–75 ms, the over-all contraction was not strongest with these interdis-charge intervals. Because of the greater latency pro-longations at short ISIs, there was a relatively narrowrange of latencies from the onset of the train to peakforce, as illustrated for doublets in Figures 1A, B and4B.

Overall Contraction Strength. The aforementioneddata have focused on the response to the last stimu-lus in a train. The strength of the contraction pro-duced by trains of different length at different ISIswas measured as the peak force produced by thetrain (Fig. 3B) and the force–time integral (i.e., the

FIGURE 2. Potentiation of the response to a supramaximal test stimulus delivered at 10 intervals, between 2.5 ms and 600 ms, after aconditioning train. The conditioning trains consisted of 2 (A) or 5 (B) supramaximal stimuli with interstimulus intervals of 5 ms (opensquares), 10 ms (shaded squares), or 50 ms (filled squares). Normalized group data (n � 4; mean � SEM).

FIGURE 3. Forces produced by trains of 2–5 responses at dif-ferent ISIs. (A) Changes in amplitude of the twitch force due tothe last stimulus of trains of 2, 3, 4, or 5 supramaximal stimuli.Normalized group data (mean � SEM; n � 4 subjects; 2 exper-iments for each subject). (B) Peak force of the contraction pro-duced by trains of stimuli at different ISIs. Note that peak force issimilar at intervals of 4–30 ms, but less at 56 and 75 ms, wheretwitch potentiation is maximal (mean � SEM; n � 4 subjects; 2experiments for each subject).


area under the force profile, representing the workdone by the muscle). At each interval, the profile ofpeak force was flat but with a clear fall-off at 56–75ms, such that the force produced by the 5-stimulustrain at the 75-ms ISI was less than that produced bya 4-stimulus train at ISIs up to 32 ms. Despite this,the total work done by the muscle was maximal atthe longer intervals, because the individual “twitch”responses were not fused. The potentiated contrac-tion at these intervals contributed to sustainingforce, but it did not produce a greater maximalforce.

Comparison with “Unaffected” Twitch Contractions.

To determine the extent to which changes in theforce produced by individual stimuli contributed tothe force profile, the recorded data for single, dou-ble, and quintuplet discharges were compared with asimulation in which a “control” single twitch wasappropriately delayed and summed to produce theforce profile that would have been recorded hadthere been no history effects of preceding contrac-

tions (Figs. 5 and 6). To do this, the control singletwitches for a sequence were averaged and, using acomputer, one or four identical force profiles weredisplaced from an identical first force profile by theintervals used in the real experiment. The forceprofiles were then summed to give the simulatedforce profile.

This revealed a complex pattern because therewas slowing of the time-course of contraction forcein addition to changes in contraction force. At short

FIGURE 4. Changes in latency. (A) Latency from the last stimu-lus in a train to the peak of force produced by that stimulus(mean � SEM; n � 4 subjects; 2 experiments for each subject).(B) Latency from the first stimulus of a train to the peak of forceproduced by the train (open symbols) or the peak of force pro-duced by the last stimulus in the train (filled symbols; note thatopen-diamond data points overlie filled-diamond data points)(mean; n � 4 subjects; 2 experiments in each subject). Note thatfor both the doublet and the quintuplet, the latency to peak forceis similar for a wide range of interstimulus intervals.

FIGURE 5. Recorded (unbroken lines) and theoretical (brokenlines) contraction profiles. Data for a single subject showing theforce produced by single stimuli and by doublets at ISIs of 10 ms(A) and 75 ms (B). The response to the second stimulus wasobtained by online subtraction (thin unbroken line). The “theoret-ical twitches” were modeled by offsetting and summing two iden-tical unconditioned twitches. Note that, in A, the recorded twitchcontraction (thick unbroken line) was of lower amplitude andslower time course than the “theoretical twitch” based on sum-ming two identical twitches appropriately offset. There is a pro-longation in latency to peak of force produced by the secondstimulus (vertical dotted lines). In B, the recorded contractionforce is greater than the “theoretical” force. The amplitude of theoverall contraction is less in B than A, but the force–time integralis increased. There is potentiation of the force produced by thesecond stimulus (compare thin unbroken line with the brokenline), but the prolongation of the latency to peak is small com-pared with A (vertical dotted lines).


intervals (�10 ms), peak force was greater with thesimulation, and the time to peak force was shorterthan in the recorded contraction. This was so evenwith the doublet (Fig. 5A). At ISIs greater than18–32 ms, peak force was greater in the recordeddata because the slower time-course of the aug-mented contractions led to greater summation offorce at these intervals. Figures 5 and 6 illustrate, forthe 10- and 75-ms intervals in one subject, the ob-served force profile for trains of 2 and 5 stimuli andthe simulated force profile based on summation of 2or 5 “control” twitches (i.e., the profile that wouldhave occurred if there were no history effects on

force generation). The force due to the first and thesecond/fifth of these “control” twitches is contrastedwith the force contributed by the second/fifth stim-ulus in the “potentiated” recording. This emphasizesthe greater size and slower time course of the re-corded contraction and illustrates that the peak offorce due to the last stimulus in the train occurredon the decay phase of the overall force profile.

DISCUSSION

The present study demonstrates for the thenar mus-cles of human subjects that the history of activity andthe pattern of stimulation have complex effects onboth the amplitude and latency of the force pro-duced by the last stimulus of a brief train. Severalnew findings for human thenar muscles are pre-sented. First, for ISIs less than 50 ms, contractionforce peaked at much the same latency after theonset of the stimulus train because of greater slowingof the contraction at short ISIs. Second, when therewere only two discharges (a “doublet”), the forceaugmentation was similar across a range of ISIs,without a clear peak at 5–15 ms, as might have beenexpected from the literature.5,9,10,30 Third, maximalaugmentation occurred at intervals of 56–75 mswhen there were more than 2 stimuli, and wasgreater the longer the stimulus train. Finally, at theseintervals, the peak force produced by the last stimu-lus in trains of 2–5 stimuli occurred after the peakforce produced by the train; that is, it prolonged theactive state and thereby contributed more to sustain-ing force than to achieving the force maximum.

Mechanisms of Changes in Contraction Force. Theseries elasticity may be important in modifying con-traction force.25,26 With the removal of slack in theseries elastic component of muscle, greater force willbe generated by the second of two twitch contrac-tions, and the apparent augmentation of force overand above that produced by a single twitch will bemaximal when the second stimulus is delivered at ornear the peak of the contraction due to the firststimulus.9,22 However, the twitch contraction pro-duced by the first stimulus should have been suffi-cient to increase muscle stiffness such that potentia-tion would be maximal for the response to thesecond stimulus, with little if any further enhance-ment with additional stimuli. This was not the case,as shown in Figure 3A.

The phenomenon documented in the presentstudy cannot be equated with the staircase phenom-enon or with posttetanic potentiation because: (1)the peaks of the force contributions produced by

FIGURE 6. Recorded (unbroken lines) and theoretical (brokenlines) contraction profiles generated by single and trains of 5supramaximal stimuli. Data for a single subject (same subject asin Fig. 5) showing the force produced by single stimuli and bytrains of 5 stimuli at ISIs of 10 ms (A) and 75 ms (B). As in Figure5, the “theoretical twitches” were modeled by offsetting and sum-ming identical unconditioned twitches. The response to the finalstimulus (i.e., fifth of 5 stimuli) was obtained by online subtraction(thin unbroken line). In A, the amplitude of the force produced bythe fifth stimulus (thin unbroken line) is less than that of anappropriately delayed but otherwise unmodified twitch contrac-tion and there is a prolongation in latency to peak by �100 ms(vertical dotted lines). In B, the recorded contraction force isgreater than the “theoretical” force. There is clear potentiation ofthe force produced by the fifth stimulus (thin unbroken line), butlittle distortion of its time-course (vertical dotted lines).


successive stimuli occurred at greatly increased laten-cies; and (2) the time-course of recovery from thosephenomena was in minutes, whereas the recovery inthe present study took 200–400 ms (Fig. 2). Theterm “catchlike” property refers to the enhancedforce production of mammalian muscle when one ortwo short interdischarge intervals occur at the onsetof a discharge or during a steady background dis-charge.6,8,9 In human thenar motor units, “doublets”at 5–15 ms can produce a force that is 48 � 13% ofthe maximal tetanic force. The enhancement isgreatest for units with low twitch/tetanus ratios30

and may be, in part, due to stretch of elastic ele-ments (see earlier). However, the enhancement in adoublet is diminished by the administration of dan-trolene (which reduces Ca2� availability) and thepotentiation presumably also involves increased re-lease of Ca2� from the sarcoplasmic reticulum22 orincreased sensitivity to Ca2�.

Accordingly, the phenomena studied here usingbrief trains produce changes similar to those de-scribed for the catchlike property of muscle. It islikely that: (1) the catchlike property operates whentrains are longer than the 2 or 3 discharges that havenormally been studied; (2) it is not confined to thefirst 2 or 3 discharges in a contraction; and (3) itdoes not require the interpolation of a brief intervalin a steady discharge. The contraction force of mus-cle may be modified within a contraction by mech-anisms over and above those that outlast the contrac-tion and produce postactivation potentiation.

Novel Findings of the Present Study and Their Implica-

tions. Four novel findings have been presented.First, slowing of the contraction at short ISIs canresult in contraction force peaking at much the samelatency after the first stimulus of the train, regardlessof the ISI. Although the rate of force increase isgreater the shorter the interval, the force maximumproduced by a doublet or a brief train of stimuli lagsbehind and may be less than the (theoretical) con-traction produced by summing single twitch contrac-tions in which there are no history-dependent effectson contraction force produced by preceding stimuli.Clearly, a preceding contraction does not always in-crease the force of subsequent contractions, and thechange involves both the peak force and the time-course of force generation. It is notable that thesechanges are more prominent at intervals tradition-ally associated with the catchlike property—that is,5–15 ms.

Second, the augmentation of the force producedby the second stimulus of a doublet occurred acrossa range of ISIs (Figs. 1D and 4A). There was no clear

peak at 5–15 ms in either the augmentation of thecontraction force produced by the second stimulusor the total force produced by the doublet (Fig. 4B),although this might have been expected from theliterature.5,9,10,30 A factor that may be relevant is thedegree of stretch of the thenar muscles, which mayhave been greater in these studies.30 The quotedstudies correctly emphasize that an additional stim-ulus at a short interval will accelerate the overall rateof force production and achieve a higher force levelmore quickly. Nevertheless, the present studies indi-cate that this occurs despite the fact that the time topeak force produced by a doublet is actually slowerthan would be the case if the contractile mechanismhad not been altered by the first contraction of thedoublet.

Third, maximal augmentation of the force pro-duced by the last stimulus in brief trains occurred atintervals of 56–75 ms and was greater the longer thestimulus train. The force–time integral (i.e., thework done by the muscle) was greatest at these in-tervals, presumably because of the greater contrac-tion force produced by the individual stimuli. Thisimplies that there may be a preferred rate for opti-mal force generation, a rate of some 15–20 Hz,which corresponds reasonably well to the adaptedfiring rate of thenar motor units in maximal volun-tary contractions.4 Fourth, at these intervals, thepeak of the force produced by the last discharge dueto trains of 2–5 stimuli occurred after the peak of theforce produced by the train, that is, it prolonged theactive state and thereby contributed more to sustain-ing force than to achieving the force maximum. Thiscontrasts with the force–time integral, but peak forceis affected more by the synchrony of individual con-tractions, whereas the integral measures the totalwork done by the muscle even when contractions arenot fused into a smooth profile. It should be remem-bered, however, that in a steady voluntary contrac-tion, motor units discharge asynchronously and, un-der these circumstances, the force–time integral willreflect the overall contraction force better than thesynchronized discharge of motor units.

We conclude that the mechanical properties ofmuscle undergo profound changes when there arepreceding contractions, and that these changes in-volve both the force and time course of the evokedcontraction, to varying extents, depending on inter-stimulus (interdischarge) interval. It is likely thatrates of 15–20 Hz are optimal for force generationby the thenar muscles during steady contractions.

This study was presented in part at Australian Association ofNeurologists Tenth Biennial Clinical Neurophysiology Workshop,Southport, Queensland, Australia, October 2003, and subse-


quently published in the conference proceedings (Howells J,Jankelowitz SK, Trevillion L, Burke D. Clini Neurophysiol 2004;115:992). This study was supported by the National Health &Medical Research Council of Australia and the University of Syd-ney Sesquicentenary R&D Scheme.

REFERENCES

1. Ali A, Hamman J, Mass DP. The biomechanical effects ofangulated boxer’s fractures. J Hand Surg 1999;24A:835–844.

2. Basmajian JV, De Luca CJ. Muscles alive, their functions re-vealed by electromyography, 5th ed. Baltimore: Williams &Wilkins; 1985. p 301–309.

3. Bawa P, Calancie B. Repetitive doublets in human flexor carpiradialis muscle. J Physiol (Lond) 1983;339:123–132.

4. Bigland-Ritchie B, Johansson R, Lippold OCJ, Smith S, WoodsJJ. Changes in motoneurone firing rates during sustainedmaximal voluntary contractions. J Physiol (Lond) 1983;340:335–346.

5. Bigland-Ritchie B, Zijdewind I, Thomas CK. Muscle fatigueinduced by stimulation with and without doublets. MuscleNerve 2000;23:1348–1355.

6. Binder-Macleod S, Kesar T. Catchlike property of skeletalmuscle: recent findings and clinical implications. MuscleNerve 2005;31:681–693.

7. Bostock H, Cikurel K, Burke D. Threshold tracking tech-niques in the study of human peripheral nerve. Muscle Nerve1998;21:137–158.

8. Burke RE, Rudomin P, Zajac FE. Catch property in singlemammalian motor units. Science 1970;168:122–124.

9. Burke RE, Rudomin P, Zajac FE. The effect of activationhistory on tension production by individual muscle units.Brain Res 1976;109:515–529.

10. Ding J, Storaska JA, Binder-Macleod SA. Effect of potentiationon the catchlike property of human skeletal muscles. MuscleNerve 2003;27:312–319.

11. Fuglevand AJ, Keen DA. Re-evaluation of muscle wisdom inthe human adductor pollicis using physiological rates of stim-ulation. J Physiol (Lond) 2003;549:865–875.

12. Garland SJ, Gossen ER. The muscular wisdom hypothesis inhuman muscle fatigue. Exerc Sport Sci Rev 2002;30:45–49.

13. Garland SJ, Griffin L, Ivanova T. Motor unit discharge rate isnot associated with muscle relaxation time in sustained sub-maximal contractions in humans. Neurosci Lett 1997;239:25–28.

14. Griffin L, Godfrey S, Thomas CK. Stimulation pattern thatmaximizes force in paralyzed and control whole thenar mus-cles. J Neurophysiol 2002;87:2271–2278.

15. Halonen J-P, Lang AH, Partanen VSJ. Change in motor unitfiring rate after double discharge: an electromyogram studyin man. Exp Neurol 1977;55:538–545.

16. Johanson ME, Valero-Cuevas FJ, Hentz VR. Activation pat-terns of the thumb muscles during stable and unstable pinchtasks. J Hand Surg 2001;26A:698–705.

17. Johnson MA, Polgar J, Weightman D, Appleton D. Data onthe distribution of fibre types in thirty-six human muscles. Anautopsy study. J Neurol Sci 1973;18:111–129.

18. Kaufman KR, An K-N, Litchy WJ, Cooney WP, Chao EYS.In-vivo function of the thumb muscles. Clin Biomech 1999;14:141–150.

19. Kiernan MC, Mogyoros I, Burke D. Differences in the recov-ery of excitability in sensory and motor axons of humanmedian nerve. Brain 1996;119:1099–1105.

20. Kudina LP, Alexeeva NL. Repetitive doublets of human mo-toneurones: analysis of interspike intervals and recruitmentpattern. Electroencephalogr Clin Neurophysiol 1992;85:243–247.

21. Marsden CD, Meadows JC, Merton PA. “Muscular wisdom”that minimizes fatigue during prolonged effort in man: peakrates of motoneuron discharge and slowing of discharge dur-ing fatigue. In: Desmedt JE, editor. Advances in neurology.Motor control mechanisms in health and disease. New York:Raven Press; 1983. p 169–211.

22. Parmiggiani F, Stein RB. Nonlinear summation of contrac-tions in cat muscles. II. Later facilitation and stiffnesschanges. J Gen Physiol 1981;78:295–311.

23. Partanen VSJ. Double discharges in neuromuscular diseases.J Neurol Sci 1978;36:377–382.

24. Pearlman JL, Roach SS, Valero-Cuevas FJ. The fundamentalthumb-tip force vectors produced by the muscles of thethumb. J Orthop Res 2004;22:306–312.

25. Rack PMH, Westbury DR. The short range stiffness of activemammalian muscle and its effect on mechanical properties.J Physiol (Lond) 1974;240:331–350.

26. Rack PMH, Westbury DR. Elastic properties of the cat soleustendon and their functional importance. J Physiol (Lond)1984;347:479–495.

27. Smutz WP, Kongsayreepong A, Hughes RE, Niebur G, CooneyWP, An K-N. Mechanical advantage of the thumb muscles.J Biomech 1998;31:565–570.

28. Søgaard K, Sjøgaard G, Finsen L, Olsen HB, Christensen H.Motor unit activity during stereotyped finger tasks and com-puter mouse work. J Electromyogr Kinesiol 2001;11:197–206.

29. Thomas CK, Johansson RS, Bigland-Ritchie B. Attempts tophysiologically classify human thenar motor units. J Neuro-physiol 1991;65:1501–1516.

30. Thomas CK, Johansson RS, Bigland-Ritchie B. Pattern ofpulses that maximize force output from single human thenarmotor units. J Neurophysiol 1999;82:3188–3195.

31. Trevillion L, Howells J, Jankelowitz S, Burke D. Axonal excit-ability measured by tracking twitch contraction force. MuscleNerve 2004;30:437–443.

32. Whitelaw WA, Watson TW. Spike trains from single motorunits in human parasternal intercostal muscles. Respir Physiol1992;88:289–298.


augmentation of the contraction force of human thenar muscles by and during brief discharge trains

Documents