w w w . e l s e v i e r . c o m / l o c a t e / p a i n

PAIN�

153 (2012) 636–643

Similar alteration of motor unit recruitment strategies during the anticipationand experience of pain

Kylie Tucker a,⇑, Anna-Karin Larsson b, Stina Oknelid b, Paul Hodges a

a School of Health and Rehabilitation Sciences, The University of Queensland, Brisbane, Australiab Department of Community Medicine and Rehabilitation, University of Umeå, Umeå, Sweden

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

a r t i c l e i n f o a b s t r a c t

Article history:Received 8 August 2011Received in revised form 20 November 2011Accepted 22 November 2011

Keywords:PainAnticipationMotor unitRecruitmentForceAdaptation

0304-3959/$36.00 � 2011 International Associationdoi:10.1016/j.pain.2011.11.024

⇑ Corresponding author. Address: NHMRC Centre oin Spinal Pain, Injury and Health, School of Health andUniversity of Queensland, Brisbane, Qld 4072, Austra

E-mail address: k.tucker1@uq.edu.au (K. Tucker).

A motor unit consists of a motoneurone and the multiple muscle fibres that it innervates, and forms thefinal neural pathway that influences movement. Discharge of motor units is altered (decreased dischargerate and/or cessation of firing; and increased discharge rate and/or recruitment of new units) duringmatched-force contractions with pain. This is thought to be mediated by nociceptive (pain) input onmotoneurones, as demonstrated in animal studies. It is also possible that motoneurone excitability isaltered by pain related descending inputs, that these changes persist after noxious stimuli cease, and thatdirect nociceptive input is not necessary to induce pain related changes in movement. We aimed to deter-mine whether anticipation of pain (descending pain related inputs without nociceptor discharge) altersmotor unit discharge, and to observe motor unit discharge recovery after pain has ceased. Motor unit dis-charge was recorded with fine-wire electrodes in the quadriceps of 9 volunteers. Subjects matched iso-metric knee-extension force during anticipation of pain (anticipation: electrical shocks randomly appliedover the infrapatellar fat-pad); pain (hypertonic saline injected into the fat-pad); and 3 intervening con-trol conditions. Discharge rate of motor units decreased during pain (P < .001) and anticipation (P < .01)compared with control contractions. De-recruitment of 1 population of units and new recruitment ofanother population were observed during both anticipation and pain; some changes in motor unitrecruitment persisted after pain ceased. This challenges the fundamental theory that pain-relatedchanges in muscle activity result from direct nociceptor discharge, and provides a mechanism thatmay underlie long-term changes in movement/chronicity in some musculoskeletal conditions.

� 2011 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

1. Introduction

It is generally accepted that changes in activity of muscle duringpainful noxious stimuli is mediated by inputs from nociceptorafferents onto motoneurones in the spinal cord. Consistent withthis view excitatory and inhibitory inputs on motoneurones havebeen shown in spinalised animals after noxious stimuli [20,21].The alternative view is that modulation of motoneurone dischargeproperties may be mediated by descending inputs. It is well recog-nised that noxious stimuli are interpreted by broad networks in thebrain, that have descending inputs that modulate spinal networksand can influence motor control [23,29,37]. This has been theorised[9] to contribute to changes in motoneurone output during pain,but has never been tested. As the descending inputs will be

for the Study of Pain. Published by

f Clinical Research ExcellenceRehabilitation Sciences, The

lia. Tel.: +61 7 3365 4589.

activated in any human exposed to pain, this cannot be excludedas a means of altered motor behaviour in human studies. The chal-lenge is to provide either descending input without nociceptivedischarge, or the converse.

Here we overcame this problem by using a pain condition thatonly has descending components. This involved anticipation of pain,known to activate brain areas [33,35] associated with the painexperience and to have descending input to spinal networks, butwithout nociceptor activation. Subjects performed contractiontasks at 2 force levels in multiple conditions: with nociceptor dis-charge and descending inputs (pain: �5 minutes of pain was in-duced in the patella fat pad by injection of hypertonic saline);with descending inputs alone (anticipation: short bursts of electri-cally induced pain were provided at random intervals on the skinabove the patella fat pad and were anticipated throughout thesecontractions but did not burst during periods used for analysis);and 3 intervening control trials that were performed before eachtest condition (C1, C2) and after completion of the trials (C3). Dis-charge properties of motoneurones were studied: the identity of

Elsevier B.V. All rights reserved.

Fig. 1. Experimental set up and pain location. Placement of fine-wire and surface(black circles) EMG electrodes (A); electrodes used for electrical stimulation (A:grey filled circles); location of painful saline injection (B: arrow); and area of pain(open grey circles) that was reported by subjects on completion of the anticipationof pain (A) and pain (B) trials are shown.

K. Tucker et al. / PAIN�

153 (2012) 636–643 637

the population of units that discharged in each condition weredetermined by evaluation of motor unit action potential morphol-ogy; discharge rate of units that were active in each condition wascompared between conditions. We provide the first published evi-dence that the same changes in motoneurone firing that are ob-served during pain (when nociceptor and pain related descendingdrive can influence motoneurone activity) also occur during antic-ipation of pain (when descending drive alone can influence moto-neurone activity). This refutes the long-accepted theory thatchanges in activity of muscle during painful stimuli is mediatedby inputs from the nociceptor afferents onto motoneurones inthe spinal cord. Changes in motor unit firing include a decreasein discharge rate including cessation of firing of 1 population andan increase in discharge rate and recruitment of a new populationof units during anticipation of pain and pain alike. This challengesthe fundamental theory of generalised inhibition of the motoneu-ron pool during pain [26]. This study also provides evidence ofincomplete recovery of motor unit discharge properties after antic-ipation/pain has ceased. The results have widespread implicationsfor movement rehabilitation during and after resolution of an epi-sode of musculoskeletal pain.

2. Materials and methods

2.1. Subjects

Fourteen volunteers (n = 5 females and 9 males) with no historyof neurological disorder or significant knee/leg pain in their test leg(n = 6 right leg, n = 8 left leg) participated in these studies. Ninesubjects participated in study 1 (age, mean [SD] 26.6 [5.1] years),and 5 participated in an additional control study (age, 27.7[9.4] years). All subjects gave written informed consent before par-ticipation. The Institutional Ethics Committee approved the study,and all procedures conformed to the principles of the Declarationof Helsinki.

2.2. Experimental design

Participants were seated on a plinth with their back and upperlegs supported and their lower legs relaxed over the end. A strapwas fixed firmly over the pelvis to limit changes in body postureduring a knee extension task [45]. Force transducers (Gedge Sys-tems, Melbourne, Australia) were attached via a �5-cm-wide ther-moplastic band (1.6-mm UltraPerf, Sammons Preston, Bolingbrook,Illinois, USA) that was moulded to the subject’s lower leg justabove the ankle, and to the plinth for support.

2.3. Electromyography

Gross electromyography (EMG) was recorded with bipolarsurface electrodes (Noraxon Dual Electrodes, Noraxon: Scottsdale,Arizona, USA, or Cleartrace Conmed, Utica, NY) placed 20 mm apartover the belly of the medial (vastus medialis: VM) and lateral (vas-tus lateralis: VL) heads of the quadriceps (Fig. 1A). The skin wasabraded (Nuprep, D.O. Weaver & Co: Aurora, Colorado, USA) andcleaned with water before attaching surface electrodes. Single mo-tor unit EMG was recorded with 4 pairs of fine-wire intramuscularelectrodes (2x Teflon-coated stainless steel wire [California FineWire Co., Grover Beach, California, USA], 100 lm diameter with0.5 mm Teflon removed and tips bent back) that were threadedinto a hypodermic needle (23–25G; Becton Dickinson, Singapore)and inserted into the VM and VL (Fig. 1A). Skin was cleanedwith antiseptic (Persist Plus; Becton Dickinson Infusion TherapySystems, Franklin Lakes, New Jersey, USA) before insertion.

EMG data were pre-amplified 1000 times (NL824, Digitimer,UK), amplified 2 more times, band-pass filtered (20–5000 Hz),

and notch filtered at 50 Hz (Neurolog, Digitimer, UK), then sampledat 10,000 Hz (Power1401 Data Acquisition System with Spike2software; CED, UK). A large ground electrode (Universal Electrosur-gical Pad: 3M Health Care NSW, Australia) was placed over the ti-bia �5 cm below the patella of the test leg.

Subjects generated an isometric knee extension torque by con-tracting their quadriceps. The experimental tasks consisted of3 � 15- to 30-second contractions at 2 force levels (see below).Subjects rested for 20 seconds between each contraction. In study1, the knee extension task was performed in 5 conditions: control 1(C1), anticipation of pain (anticipation), control 2 (C2), pain (pain),and control 3 (C3). Conditions were not randomised, as it wasimportant to complete the anticipation of pain trial before the paincondition for 2 reasons. First, the primary goal of this study was todetermine the effects of anticipation of pain on single motor unitdischarge parameters; and second, the time course of recovery ofmotor unit discharge properties after resolution of pain has notbeen determined. Randomisation of the order of conditions mayhave compromised the integrity of the results.

Before testing began, 2 contraction force levels were deter-mined. At the lower and higher force levels, the aim was to recruit2 to 4 and 4 to 8 motor units, respectively, from any of the 4 fine-wire electrodes. During the testing sessions extension efforts wereperformed with visual feedback of force on a computer screenpositioned �1 m in front of the subject. In each condition subjectsincreased their force level from rest to the target force level over 5to 8 s, and maintained the given force level until instructed to ramptheir force back down to rest over 5 to 8 sequences. The sequenceof testing for study 1 is shown in Fig. 2.

2.4. Anticipation of pain: electrical stimulation

During the anticipation of pain condition subjects performedthe task while expecting painful electrical stimulation of the skinover the infrapatella fat pad. The stimulating electrodes (NoraxonDual Electrodes, Noraxon: Scottosdale, Arizona, USA) were placeover the medial inferior edge of the patella just before the com-mencement of the anticipation trial (Fig. 1A). Electrical stimuli(200 Hz, 750 ms, 200-ls pulse duration) similar to that used

Fig. 2. Sequence of testing. (A) The lower and higher force levels required to recordmotor unit discharge were determined, and the feedback screen set to display theseforces as required. (B) Control 1: 3 times lower and then 3 times higher forcecontractions were performed. (C) Stimulating electrodes were applied, and thestimulus intensity required to elicit pain �5/10 was determined. (D) Anticipation: 3times lower and then 3 times higher force contractions were performed. Painfulelectrical stimuli (#) were provided at random intervals during the contractions.Pain level was reported at the end of each contraction and stimulus intensityincreased if required to maintain pain level �5/10. (E) Participants were informedthat no further electrical stimuli would be applied. (F) Control 2: 3 times lower andthen 3 times higher force contractions were performed. (G) Hypertonic saline wasinjected into the patella fat pad. (H) Pain: 3 times lower and then 3 times higherforce contractions were performed, pain was reported throughout the contractions.(I) Control 3: 3 times lower and then 3 times higher force contractions wereperformed after resolution of experimental pain.

638 K. Tucker et al. / PAIN�

153 (2012) 636–643

previously [11,14,30] were delivered using a constant currentstimulator (DS7A, Digitimer, UK). Stimulus intensity was graduallyincreased from 0 mA in �4 mA increments until the subject re-ported verbally the experience of pain of �5/10 on an 11-point nu-meric rating scale (NRS) during the stimuli (where 0 = ‘no pain’ and10 = ‘worst pain imaginable’). The stimulus intensities required forthis pain level ranged between 7 and 21 mA. To provide an antici-pated pain, subjects were told that they would receive the painfulelectrical stimulation at random intervals between 0.5 and 8 sec-onds during each contraction in the anticipation of pain condition.The electrical stimulation then appeared as proposed during thesecontractions without further warning. During this condition, sub-jects maintained the required level of contraction until 2 � 5-sec-ond periods of ‘anticipation of pain’ (periods with no painfulelectrical stimulus) were observed by the investigator. These2 � 5-second periods were later used for single motor unit analy-sis. Subjects were then instructed to ramp their force back downto rest over 5 to 8 seconds. The stimulator was turned off duringthe 20-second rest period between the contractions. After eachcontraction, subjects were asked to rate the pain that they experi-enced during the electrical stimuli, using the 11-point NRS. If theirpain level dropped below 5/10, the stimulus intensity was in-creased by�1 to 3 mA. This was required in 5 of the 9 subjects dur-ing the anticipation trials. The area of pain experienced during theelectrical stimuli was recorded by the subjects on a schematic dia-gram after the completion of the anticipation of pain trial (Fig. 1A).

2.5. Pain: hypertonic saline injection

For the pain condition, pain was induced by a single bolus injec-tion of hypertonic saline (0.25 mL, 5% NaCl: 25-G 16-mm syringe)

into the medial side of the infrapatellar fat pad [1,45] (Fig. 1B). Thesaline was injected immediately (�30 s) after the second controlcondition ceased. Experimental pain was reported throughout thepain trial using a custom-made electronic 10-cm visual analoguescale (VAS) where 0 = ‘no pain’ and 10 = ‘worst pain imaginable’.Recordings began as soon as the pain reached 3/10 on the VAS.The area of pain was recorded by the subjects on a schematic dia-gram after the pain trial (Fig. 1B). Subjects were completely painfree before the post-pain control (C3) condition commenced(�5 min after completion of pain trials).

2.6. Study 2: ‘no-anticipation’ control trial

As the anticipation of pain trials involved intermittent nocicep-tive stimulation, although not at the time of evaluating motor unitdischarge, it was necessary to determine whether changes in mo-tor unit discharge were due to residual effects of the electricalnociceptor activation (wind up is a phenomenon by which spinalcircuits are argued to maintain heightened excitability or sustainedactivity after a nociceptive volley and may be argued to occur aslonger-term sequelae from the short bursts of painful electricalstimulation in the anticipation of pain trials). We therefore under-took a second study, on a separate day that involved comparison ofa control condition to a trial with no-anticipation of pain.

The no-anticipation trial involved the same painful electricalstimulation as described above, during the ramp up phase of thecontraction, but subjects were instructed that no further painwould be induced during the sustained contraction and thus, par-ticipants did not anticipate any further pain. In this study subjectsramped up their knee extension force (to a predetermined force le-vel that recruited �4–5 motor unit) from rest over �5 s, held thisforce for �15 s and ramped back down to rest over �5 s. Thiswas repeated 3 times with >20 s rest between contractions.

2.7. Data analysis

Single motor unit discharge properties were analysed as de-scribed by Tucker et al. [44]. In brief, 10 seconds of data from eachof the 3 repeated contractions, during the C1-3 and pain conditionswas chosen for analysis based on the force amplitude (equivalentbetween conditions). To avoid contamination of anticipation ofpain data with motor unit behaviour that results from direct noci-ceptive electrical stimulation, data were analysed from 6 � 5-sec-ond periods during the anticipation of pain trials that began>0.5 seconds after cessation of the electrical stimuli.

Single motor unit action potentials were identified based onmorphology using Spike2 software (CED, UK) from the 4 fine-wireelectrodes pairs. Averages of single motor unit fine-wire and sur-face recordings were then triggered from the discharge of each dis-criminated unit over the 3 � 10-second analysis periods (or 6 � 5-second analysis periods in the anticipation of pain trial) to generatea template of the motor unit morphology. Using the surface repre-sentation of the motor unit morphology is advantageous, as thereis less potential for change in shape over the period of the experi-ment than with recordings made with intramuscular electrodesthat may move slightly during contractions. Both fine-wire andsurface recorded motor unit profiles were compared visually with-in subjects to determine whether the same unit was present ineach contraction. Motor units that fired consistently in 2 of 3 con-tractions (for both force levels separately) and for more than 50% ofthe time in those contractions were used for further analysis (forfurther discussion of this motor unit analysis, see Tucker et al.[44]). The discharge time of motor units that were recorded fromwithin the same muscle (but in different recording zones) werecompared to ensure that the same motor unit was not recordedfrom the 2 separate channels. In the case that the same unit was

K. Tucker et al. / PAIN�

153 (2012) 636–643 639

identified in both channels, this unit was included in the analysisonce.

Changes in motoneuron discharge within and between record-ing sites have been observed previous during pain [44–46], buthave not been described during anticipation of pain, or during con-tractions after pain has ceased. Altered motor unit discharge with-in a single recording site provides the best evidence available usingsingle motor unit analysis, for altered recruitment within a singlemotoneuron pool. Changes in the discharge characteristics of mo-tor units between recording sites and when data from all recordingsites from the synergist muscles are pooled provide evidence of amore gross change in motor strategy between conditions. As VLand VM are synergists during production of isometric knee exten-sion force, data are pooled from both muscles for the motor unitanalysis.

2.8. Statistical analysis

Analysis was completed in 2 stages for study 1. First, data wasanalysed from the first 3 conditions (C1, anticipation and C2) todetermine whether (1) force was matched adequately betweenconditions, (2) anticipation of pain affected single motor unit dis-charge properties, and (3) the changes observed during anticipa-tion of pain resolved following this condition. Second, data fromall 5 conditions (C1, anticipation, C2, pain, and C3) were comparedin separate analysis to determine if the changes identified duringanticipation of pain were also observed during and after the exper-imental pain condition. This analysis was completed separately, as2 subjects did not participate in the pain and C3 trials, and another

Fig. 3. Altered motor unit recruitment during anticipation of pain. Recordings of knee extand second control and anticipation of pain trials performed at the lower force level.unchanged between conditions (top trace). Changes in motor unit discharge are visible in2 s represented in the figure are shown overlaid, with their mean discharge rates. In this eduring anticipation of pain. In fine-wire electrode 1, the discharge rate of unit B decreasesneurone, at the same time a new unit (A) is recruited which demonstrates a simultaneourecruited during anticipation of pain, however unit E, which is recorded from the sameprofile of the motor units present during each condition (B, E), and a unit that is only recindicates the electrode placement did not change between conditions.

subject recruited too many units to be accurately discriminatedduring the higher-force pain trial. Data from the second analysisis therefore based on 6 and 7 subjects in the higher and lower forceconditions, respectively.

Repeated-measures analyses of variance (ANOVAs) were usedto compare motor unit discharge rate and force. If there was a maineffect of condition, a Scheffé post-hoc test was used.

The level of pain reported during the bursts of pain in the antic-ipation of pain trial, and the average pain reported throughout thepain trial was compared using a t test. For the additional controlstudy (study 2), a paired t test was used to compare motor unit dis-charge rate between the control and ‘no-anticipation’ trials.

All data are original for this manuscript. Data are reported asmean (SD), and significance was set at P < .05.

3. Results

Short bursts of electrical stimulation of the skin over the infra-patella fat pad used during the anticipation of pain trials (anticipa-tion) produced a reported pain intensity [mean (SD)] of 5.2 (0.7)/10. After the bolus injection of hypertonic saline into the infra-pa-tella fat pad during the pain trials, participants reported a slightlylower average pain intensity of 4.3 (0.7)/10 (P = .02). Participantsperceived the pain along the medial side of the patella during boththe pain and the anticipation of pain trials (while being electricallystimulated) (Fig. 1A and B).

The isometric knee extension force (N) used to discharge therequired number of motor units varied between subjects, but wasmatched between conditions (when comparing control 1, anticipation,

ension force and muscle activity are shown for a representative subject from the firstFive motor units were discriminated from 2 fine-wire EMG channels. Force wasthe 2 raw EMG traces. The voltage profile of the discriminated motor units over thexample, units B and E discharge in all conditions, and their discharge rate is reducedduring anticipation of pain demonstrating a reduction in the excitability of its motor

s increase in the excitability of its motoneuron. In fine-wire electrode 2, unit C is notrecording zone continues to discharge throughout the testing sessions. The voltageruited during the control conditions (C) remain the same throughout the trials. This

640 K. Tucker et al. / PAIN�

153 (2012) 636–643

and control 2: the average force produced during all conditions atlower [14.6 (8.5)], and higher [20.7 (11.3)] force levels, was not dif-ferent between conditions (main effect for condition: higher[P = .73] and lower [P = .18] force).

The population of motor units that were recruited to maintainforce was altered during the contractions with anticipation of paincompared with the control contractions (Figs. 3 and 4C and D). Ofthe 63 motor units that were discriminated during the lower forcecontractions, �37% (n = 23) discharged in all 3 conditions (C1,anticipation, C2). A second population of units (n = 8) were re-cruited only during the control conditions, ie, they were recruitedduring C1, then did not fire during anticipation, and recommencedfiring during C3 when no further anticipation of pain was present.These units were replaced with another population of units (n = 8)that only discharged during the anticipation of pain contractions,and other units (n = 7) that fired first during the anticipation ofpain, then continued to fire during the C2 contractions (Fig 3, rep-resentative data; Fig. 4C, group data). This pattern of recruitmentwas also observed during the higher force contractions (Fig. 4D).

The discharge rate of single motor units that were identified inall 3 conditions (ie, excluding any units that were newly recruitedor only recruited in 1 or 2 of the 3 conditions) was significantlylower during anticipation of pain compared with the control con-tractions at both force levels (see raw data in Fig 3; Main effectfor Condition P < .01; post hoc statistics in Fig. 4A and B). This sig-nificant reduction in discharge rate exists even though 7 of the 23motor units observed in all 3 conditions increased their dischargerate during the anticipation of pain contractions compared withthe first control (from 8.1 [1.1] to 8.6 [1.0] Hz). These 7 units con-tinued to discharge at a higher rate during the second control con-tractions [8.7 [0.8] Hz].

The average force produced during all 5 test conditions at lower(14.2 [6.3]), and higher (19.3 [7.0]) force level, was not different be-tween conditions (main effect for condition: higher [P = .20] andlower force [P = .44]). The changes in motor unit discharge patterns

Fig. 4. Group motor unit data during anticipation of pain. The discharge rate (A, B)and number of units (C, D) recruited during the different conditions for both forcelevels are shown from 9 subjects. The units reported as ‘other’ (C, D) are those unitsrecruited in only 1 of the control conditions (low force: n = 13; high force: n = 16) orcontrol 1 (C1) and anticipation of pain (Ant), but not the second control trial (C2).The units shown in black were recruited for the first time (or only) duringanticipation of pain. The change in recruitment of motor units during anticipationof pain includes the de-recruitment of a population of units, recruitment of newunits (C, D) and the reduction in discharge rate of units during anticipation of pain,that are recruited in all conditions (A, B: ⁄P < .05, ⁄⁄P < .001. Discharge rate shown asmean [SD]).

during contractions with anticipation of pain were similar to thoseobserved during contractions with pain. Approximately 29% and50% of units were recruited in all 5 conditions (C1–C3, anticipationand pain) during the lower and higher force contractions, respec-tively. Discharge rate of these units decreased during anticipation,and during pain, compared with the 3 control trials (pain: bothforce levels; anticipation of pain: higher force only: all P < .001;Fig. 5A and B). Although a significant decrease in discharge rateis observed during both anticipation and pain compared with thecontrol contractions in the grouped results, not all of the motorunits have a lower discharge rate during anticipation/pain (exam-ples of the complex changes in motor unit discharge characteristicsare shown in Fig. 5C and D and Fig. 6). For descriptive purposes, thedata from both force levels are combined. To determine whethermotor unit recruitment strategies are altered during anticipationof pain and pain alike, the changes in motor unit discharge behav-iour during control 1, anticipation, and control 2 (n = 104 motorunits) are compared with the changes observed during control 2,pain, and control 3 (n = 105 motor units). Approximately half (53and 55) of these motor units discharged consistently in all 3 condi-tions, respectively. Of these, 8 and 12 increased their dischargerates during the test condition. The remaining 45 and 43 decreasedtheir discharge rate during the test condition (Fig. 5C and D andFig. 6). In addition, 20 and 24 units ceased firing, and 14 and 13units fired for the first time, during the anticipation and pain con-ditions, respectively. This provides further evidence of nonuniformchanges in motor unit discharge within a muscle during bothanticipation of pain and during pain. The units that increased/de-creased firing during the test conditions (pain and anticipation ofpain) were not necessarily the same between conditions.

Motor unit discharge rate recovered to initial control levels afterthe anticipation of pain and pain trials at both force levels (Fig. 5A).In contrast to the discharge rate, not all of the units that either didnot discharge or were newly recruited during the test conditionsreturned to their pre test-condition state following completion ofthe test condition. For example, of the 20 and 24 motor units thatceased firing during anticipation and pain (discussed above),respectively, 13 and 10 of these units returned to discharged againonce anticipation/pain had ceased. In addition, of the 14 and 13units that fired for the first time during anticipation of pain, 7and 6 units continued to discharge after anticipation/pain hadceased. This provides evidence of a mechanism that maintainslonger-term adaptations to acute musculoskeletal pain.

In study 2, force (N) was well maintained between conditions[control: 7.5 (7.1); no-anticipation: 7.5 (7.0); P = .33]. A total of25 motor units were discriminated from 5 subjects. Of these, 80%(n = 20) motor units were present in both the control and no-antic-ipation trials. The motor unit discharge rate (Hz) did not decreaseduring these trials compared with control trials when pain was notanticipated [control: 8.6 (1.8); no anticipation: 8.9 (1.9); P = .19].

4. Discussion

There are 2 novel findings from this study. First, motor unitrecruitment is altered during anticipation of pain, when there isno direct input from nociceptive discharge onto the motoneuronepool. The changes in motor unit discharge properties are similarto those observed during periods of direct nociceptor stimulation(pain), and include reduced discharge (including de-recruitment)in 1 population of motor units, and increased discharge (includingnew recruitment) in another population of motor units. The alteredrecruitment is observed both within the selective fine-wire record-ing zones, and throughout large synergistic muscles [17,44–46]. Asthere was no ongoing nociceptor discharge during the anticipationof pain contractions, these data do not support the theory thatchanges in motor unit discharge during pain depend on direct

Fig. 5. Altered discharge properties during anticipation of pain and pain alike. The mean (SD) discharge rate (A) and the percentage change in discharge rate from the firstcontrol (C1) contraction (B) of the units that discharged in all 5 conditions are shown. During the higher force contractions (dark grey), when a greater number of units werepresent in all conditions, the reduction in discharge rate was similar in both the anticipation of pain (Ant) and pain conditions compared with the 3 control conditions (C1–C3)(⁄⁄P < .001; ⁄P < .05; symbols adjacent to asterisk [⁄] refer to data compared for that analysis). When considering the whole population of motor units identified in anycondition, a complex change in discharge behaviour is observed. Data from lower and higher forces are combined for graphical purposes, from 7 participants who completedthe 5 conditions (C, D). Each circle represents a unit that fired during both control 1 and anticipation (white) or control 2 and pain (black). Mean discharge rate (C) and percent change in discharge rate from the preceding control contraction (D) are shown. Units that are below the line of equality (dotted line) decreased in discharge rate duringthe test (anticipation or pain) condition compared with their respective control condition. Units below this line and encased in the grey box did not fire during the testcondition (ie, 0 Hz [left panel]; 100% decrease [right panel]). Units above the line of equality increased in discharge rate during anticipation/pain. Units above this line andencased in the grey box only fired during the test condition (and not during their respective control trial).

K. Tucker et al. / PAIN�

153 (2012) 636–643 641

nociceptor input to the motoneurone pool, although nociceptorinput may contribute when it is present. Second, we have shownthat some changes in motor unit discharge during pain/anticipa-tion of pain do not resolve once the threat of pain is ceased. Thismay be because the goal of reducing pain (for example) drives achange in the motor strategy during pain or when pain is antici-pated, but there is no specific drive to return the motor strategyback to its original pattern once the painful stimuli have ceased.These results have widespread implications in understanding themechanisms underlying movement changes and musculoskeletalrehabilitation, during and after resolution of acute pain.

Lund et al. (1991) proposed theoretically that reduced agonistmuscle activity during pain results from generalised inhibitory in-put to the motoneurone pool from nociceptive afferents [26].Although not previously tested with respect to changes in moto-neurone discharge, other authors have proposed, but not shown,changes in drive to the motoneurone pool during pain by descend-ing inputs from supraspinal areas including the motor cortex

[9,18,33,35,40]. We have shown both increased and decreased mo-tor unit discharge activity in agonist muscles during pain [44–46]and anticipation of pain (current data). As motor unit dischargefrom the same muscle was simultaneously reduced[9,17,39,45,46], even to the point of de-recruitment [44–46] andincreased (increased firing of some already active units andrecruitment of new previously silent units) [44–46], our findingsreject the claim that the pool is uniformly inhibited during pain.This is consistent with findings from animal studies that demon-strate both inhibitory and excitatory post-synaptic potentials in al-pha motoneurones in response to stimulation of nociceptorafferents [21,22], and data from human studies that show variabil-ity (both increased and decreased agonist muscle activity) in re-sponse to pain [8,12,19,25,36,41].

The simultaneous increase and decrease of motor unit dischargeoccur within the same recording zone of a fine-wire electrodewithin a muscle, and between synergist muscles during pain oranticipation of pain compared with controlled trials [44–46]

Fig. 6. Changes in motor unit recruitment during pain and anticipation of pain aresimilar. Data from lower and higher forces are combined for graphical purposes,from n = 7 participants who completed the 5 conditions. Data are divided into unitsthat discharged during the C1, anticipation, C2 conditions (left) and C2, pain, C3conditions (right). The 3 test conditions are represented by the 3 symbols withineach division of the graph by a + (if the unit discharged) or a � (if the unit did notdischarge). Approximately 50% of the units that were identified in any of the 3conditions were present in all 3 conditions (white box: +++). Of these, �18%increased in discharge rate, and �72% decreased in discharge rate during antici-pation/pain (see arrows). Approximately 20% of the units active before anticipation/pain were not recruited during anticipation/pain (black box). Half of these unitsfired again after anticipation/pain (+�+), and half did not (+��). Approximately14% of the units appeared for the first time during anticipation/pain (dark grey).Half of the units continued to fire once anticipation/pain had ceased (�++), and halfdid not (�+�). Approximately 7% of the units appeared for the first time (��+) anddid not discharge again (++�) after anticipation/pain had ceased (light grey). Thisfigure demonstrates the similarity in the altered motor unit discharge behaviorsduring anticipation of pain and pain.

642 K. Tucker et al. / PAIN�

153 (2012) 636–643

(Fig. 3). We have previously argued [44] that changes in recruit-ment within a fine-wire recording zone may reflect changes in or-derly recruitment within a motoneurone pool [2,15,16,28] torecruit units with larger force production capabilities earlier[10,13,31,38]. Alternatively, changes within and between record-ing zones may reflect a change in the distribution of drive to motorunits that have a different direction of force production within themuscle [3,7,34,42,43,46] during pain.

This study is the first to investigate motor unit discharge whenthere is potential for descending inputs related to cognitions of painanticipation, but in the absence of concurrent nociceptive discharge.As no direct nociceptor stimulation occurred in the anticipation ofpain trials (during the period in which motor unit activity wasinvestigated), we argue that any changes in motoneurone dischargeduring these periods cannot be due to ongoing direct nociceptor in-put, and therefore must involve ‘‘top–down’’ mechanisms. It is likelythat these same mechanisms are involved with altered motoneu-rone discharge during pain, when both nociceptor afferents anddescending drives will be active. It is possible that these top–downmechanisms include changes in drive from the motor cortex, whichmay be either voluntary or involuntary. Motor drive may also beinfluenced by other brain centres that are affected by anticipationof pain, and during pain, such as the periaqueductal gray area ofthe midbrain [40], anterior cingulated cortex and parietal opercu-lum/posterior insula [35], and primary somatosensory cortex andanteroventral cingulated cortex [33]. This work also provides amechanism for previously reported changes in motor outcomes re-ported during both pain and anticipation of pain (including alteredgait [23,24], posture [6,30], and reflex activity [4,37].

Our data support previous work that shows recovery of de-creased discharge rate of motor units after resolution of pain

[17,39]. However, this study is the first to consider the recoveryof units that ceased discharge or were newly recruited during pain(or anticipation of pain). Our data show that �35% and 58% of theunits recruited during the preceding control, but then not recruitedduring anticipation of pain and pain trials, respectively, remainedinactive during the subsequent control contractions. Similarly,�50% and 53% of the units that were newly recruited during theanticipation of pain and pain trials, respectively, persisted to dis-charge during subsequent control contractions. This suggests thatalthough discharge rate resolves after cessation of pain, theadaptation in spatial distribution of activity does not return tothe initial state, that is, the nervous system maintains some com-ponents of the new strategy adopted during pain. This may be be-cause there is a clear goal for the nervous system to find a newstrategy during the painful contractions but not after pain hasceased. For example, a change in motor strategy during pain (oranticipation of pain) may assist in performing the task in a lesspainful manner. In contrast, once pain has ceased, there may beno specific motivation to return the altered motor strategy backto its original pattern. This may be 1 of the mechanisms thatunderlies long-term changes in motor control strategies after acutepain episodes, and will be important to consider in relation to reoc-currence and chronicity of symptoms.

An additional controlled trial with no anticipation of pain wasconducted to ensure that altered recruitment during the anticipa-tion of pain trials (when short bursts of pain were administered atrandom intervals) was not due to ‘‘wind up’’ in the spinal cord ordue to incomplete resolution of changes in motor unit activity in-duced by the short bursts of painful stimulation used during theanticipation of pain trials. Wind up refers to a change in respon-siveness of spinal cord circuitry in response to intermittent painfulstimulation [5,27,32,47,48]. When the painful stimuli were pro-vided at the commencement of a sustained contraction but theparticipants were instructed that there would be no further pain(ie, no anticipation of pain), there was no change in discharge rateand relatively few units altered their discharge characteristicscompared with that observed during the anticipation of pain trials.Thus the anticipation of future pain was sufficient to induce achange in motor strategy, but a single predictable short burst ofpain did not induce these changes.

4.1. Conclusion

This study supports and extends our previous work which sug-gest that the nervous system uses a different motor unit recruit-ment strategy to achieve the same force output during pain andanticipation of pain, and that the change in recruitment of unitsare likely due to uneven distribution of synaptic input acrossthe motoneurone pool. We argue that changes in motor unitrecruitment are likely to involve top–down mechanisms and donot simply involve direct nociceptive input to the motoneuronepool, although the effects of nociceptive inputs may summatewith the effects due to top–down mechanisms during pain. Thechanges in motor unit discharge properties may contribute tothe altered motor output previously observed during pain andanticipation of pain experiments. The changes in motor unit firingdid not completely resolve after resolution of pain, and wereobserved after short exposure to pain when further pain wasanticipated. These findings are important to consider in relationto rehabilitation and chronicity of musculoskeletal painconditions.

Conflict of interest statement

There were no conflicts of interest.

K. Tucker et al. / PAIN�

153 (2012) 636–643 643

Acknowledgments

Financial support was provided by the National Health andMedical Research Council (NHMRC) of Australia (Research Fellow-ships [PH] ID401599 [KT] ID1009410; Project grant ID569744).

References

[1] Bennell K, Hodges P, Mellor R, Bexander C, Souvlis T. The nature of anteriorknee pain following injection of hypertonic saline into the infrapatellar fat pad.J Orthop Res 2004;22:116–21.

[2] Burke R. Selective recruitment of motor units. In: Humphrey D, Freund H,editors. Motor control: concepts and issues. Hoboken, NJ: Wiley; 1991. p. 5–21.

[3] Butler JE, McKenzie DK, Gandevia SC. Discharge properties and recruitment ofhuman diaphragmatic motor units during voluntary inspiratory tasks. J Physiol1999;518:907–20.

[4] Cadden SW. Modulation of human jaw reflexes: heterotopic stimuli and stress.Arch Oral Biol 2007;52:370–3.

[5] Cook AJ, Woolf CJ, Wall PD. Prolonged C-fibre mediated facilitation of theflexion reflex in the rat is not due to changes in afferent terminal ormotoneurone excitability. Neurosci Lett 1986;70:91–6.

[6] della Volpe R, Popa T, Ginanneschi F, Spidalieri R, Mazzocchio R, Rossi A.Changes in coordination of postural control during dynamic stance in chroniclow back pain patients. Gait Posture 2006;24:349–55.

[7] Desnedt HE, Gidaux E. Spinal motoneuron recruitment in man: rankdeordering with direction but not with speed of voluntary movement.Science 1981;214:933–6.

[8] Farina D, Arendt-Nielsen L, Graven-Nielsen T. Experimental muscle paindecreases voluntary EMG activity but does not affect the muscle potentialevoked by transcutaneous electrical stimulation. Clin Neurophysiol2005;116:1558–65.

[9] Farina D, Arendt-Nielsen L, Merletti R, Graven-Nielsen T. Effect of experimentalmuscle pain on motor unit firing rate and conduction velocity. J Neurophysiol2004;91:1250–9.

[10] Garnett R, Stephens JA. Changes in the recruitment threshold of motor unitsproduced by cutaneous stimulation in man. J Physiol (Lond) 1981;311:463–73.

[11] Gasser HS, Erlanger J. The role of fiber size in the establishment of a nerveblock by pressure or cocaine. Am J Physiol 1929;88:581–91.

[12] Graven-Nielsen T, Svensson P, Arendt-Nielsen L. Effects of experimentalmuscle pain on muscle activity and co-ordination during static and dynamicmotor function. Electroencephalogr Clin Neurophysiol 1997;105:156–64.

[13] Grimby L, Hannerz J. Firing rate and recruitment order of toe extensor motorunits in different modes of voluntary contraction. J Physiol (Lond)1977;264:865–79.

[14] Handwerker HO, Kobal G. Psychophysiology of experimentally induced pain.Physiol Rev 1993;73:639–71.

[15] Heckman CJ, Binder MD. Neural mechanisms underlying the orderlyrecruitment of motoneurons. In: Binder MD, Mendell LM, editors. Thesegmental motor system. New York: Oxford University Press; 1990.

[16] Henneman E. Relation between size of neurons and their susceptibility todischarge. Science 1957;126:1345–7.

[17] Hodges PW, Ervilha UF, Graven-Nielsen T. Changes in motor unit firing rate insynergist muscles cannot explain the maintenance of force during constantforce painful contractions. J Pain 2008;9:1169–74.

[18] Hodges PW, Richardson CA. Inefficient muscular stabilization of the lumbarspine associated with low back pain. A motor control evaluation of transversusabdominis. Spine 1996;21:2640–50.

[19] Hodges PW, Tucker K. Moving differently in pain: a new theory to explain theadaptation to pain. Pain 2011;152:S90–8.

[20] Kniffki KD, Schomburg ED, Steffens H. Synaptic effects in alpha-motoneuronesof spinal cats induced by chemical algesic stimulation of skeletal muscle[proceedings]. J Physiol 1978;284:174P–5P.

[21] Kniffki KD, Schomburg ED, Steffens H. Synaptic responses of lumbar alpha-motoneurones to chemical algesic stimulation of skeletal muscle in spinal cats.Brain Res 1979;160:549–52.

[22] Kniffki KD, Schomburg ED, Steffens H. Synaptic effects from chemicallyactivated fine muscle afferents upon alpha-motoneurones in decerebrate andspinal cats. Brain Res 1981;206:361–70.

[23] Lamoth CJ, Daffertshofer A, Meijer OG, Lorimer Moseley G, Wuisman PI, BeekPJ. Effects of experimentally induced pain and fear of pain on trunk

coordination and back muscle activity during walking. Clin Biomech2004;19:551–63.

[24] Lamoth CJ, Meijer OG, Daffertshofer A, Wuisman PI, Beek PJ. Effects of chroniclow back pain on trunk coordination and back muscle activity during walking:changes in motor control. Eur Spine J 2006;15:23–40.

[25] Le Pera D, Graven NT, Valeriani M, Oliviero A, Di Lazzaro V, Tonali P, Arendt-Nielsen L. Inhibition of motor system excitability at cortical and spinal level bytonic muscle pain. Clin Neurophysiol 2001;112:1633–41.

[26] Lund JP, Donga R, Widmer CG, Stohler CS. The pain-adaptation model: adiscussion of the relationship between chronic musculoskeletal pain andmotor activity. Can J Physiol Pharmacol 1991;69:683–94.

[27] Mendell LM. Physiological properties of unmyelinated fiber projection to thespinal cord. Exp Neurol 1966;16:316–32.

[28] Mendell LM. The size principle: a rule describing the recruitment ofmotoneurons. J Neurophysiol 2005;93:3024–6.

[29] Moseley GL, Brhyn L, Ilowiecki M, Solstad K, Hodges PW. The threat ofpredictable and unpredictable pain: differential effects on central nervoussystem processing? Aust J Physiother 2003;49:263–7.

[30] Moseley GL, Nicholas MK, Hodges PW. Does anticipation of back painpredispose to back trouble? Brain 2004;127:2339–47.

[31] Nielsen J, Kagamihara Y. Differential projection of the sural nerve to early andlate recruited human tibialis anterior motor units: change of recruitment gain.Acta Physiol Scand 1993;147:385–401.

[32] Peng HY, Chen GD, Lee SD, Lai CY, Chiu CH, Cheng CL, Chang YS, Hsieh MC,Tung KC, Lin TB. Neuroactive steroids inhibit spinal reflex potentiation byselectively enhancing specific spinal GABA(A) receptor subtypes. Pain2009;143:12–20.

[33] Porro CA, Baraldi P, Pagnoni G, Serafini M, Facchin P, Maieron M, Nichelli P.Does anticipation of pain affect cortical nociceptive systems? J Neurosci2002;22:3206–14.

[34] Riek S, Bawa P. Recruitment of motor units in human forearm extensors. JNeurophysiol 1992;68:100–8.

[35] Sawamoto N, Honda M, Okada T, Hanakawa T, Kanda M, Fukuyama H, KonishiJ, Shibasaki H. Expectation of pain enhances responses to nonpainfulsomatosensory stimulation in the anterior cingulate cortex and parietaloperculum/posterior insula: an event-related functional magnetic resonanceimaging study. J Neurosci 2000;20:7438–45.

[36] Schulte E, Ciubotariu A, Arendt-Nielsen L, Disselhorst-Klug C, Rau G, Graven-Nielsen T. Experimental muscle pain increases trapezius muscle activityduring sustained isometric contractions of arm muscles. Clin Neurophysiol2004;115:1767–78.

[37] Scott AJ, Cadden SW. Suppression of an inhibitory jaw reflex by theanticipation of pain in man. Pain 1996;66:125–31.

[38] Semmler JG, Turker KS. Compound group I excitatory input is differentiallydistributed to motoneurons of the human tibialis anterior. Neurosci Lett1994;178:206–10.

[39] Sohn MK, Graven-Nielsen T, Arendt-Nielsen L, Svensson P. Inhibition of motorunit firing during experimental muscle pain in humans. Muscle Nerve2000;23:1219–26.

[40] Sterling M, Jull G, Wright A. Cervical mobilisation: concurrent effects on pain,sympathetic nervous system activity and motor activity. Man Ther2001;6:72–81.

[41] Svensson P, Houe L, Arendt Nielsen L. Bilateral experimental muscle painchanges electromyographic activity of human jaw-closing muscles duringmastication. Exp Brain Res 1997;116:182–5.

[42] ter Haar Romeny BM, Denier van der Gon JJ, Gielen CC. Changes in recruitmentorder of motor units in the human biceps muscle. Exp Neurol 1982;78:360–8.

[43] Thomas JS, Schmidt EM, Hambrecht FT. Facility of motor unit control duringtasks defined directly in terms of unit behaviors. Exp Neurol 1978;59:384–97.

[44] Tucker K, Butler J, Graven-Nielsen T, Riek S, Hodges P. Motor unit recruitmentstrategies are altered during deep-tissue pain. J Neurosci 2009;29:10820–6.

[45] Tucker KJ, Hodges PW. Motoneurone recruitment is altered with pain inducedin non-muscular tissue. Pain 2009;141:151–5.

[46] Tucker KJ, Hodges PW. Changes in motor unit recruitment strategy during painalters force direction. Eur J Pain 2010;14:932–8.

[47] Wall PD, Woolf CJ. The brief and the prolonged facilitatory effects ofunmyelinated afferent input on the rat spinal cord are independentlyinfluenced by peripheral nerve section. Neuroscience 1986;17:1199–205.

[48] Woolf CJ, Wall PD. Relative effectiveness of C primary afferent fibers ofdifferent origins in evoking a prolonged facilitation of the flexor reflex in therat. J Neurosci 1986;6:1433–42.