effects of a force production task and a working memory task on pain perception

10

Click here to load reader

Upload: stephen-a

Post on 25-Dec-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Effects of a Force Production Task and a Working Memory Task on Pain Perception

R

The Journal of Pain, Vol 14, No 11 (November), 2013: pp 1492-1501Available online at www.jpain.org and www.sciencedirect.com

Effects of a Force Production Task and a Working Memory Task

on Pain Perception

Tiffany A. Paris, Gaurav Misra, Derek B. Archer, and Stephen A. CoombesLaboratory for Rehabilitation Neuroscience, Department of Applied Physiology and Kinesiology, University of Florida,Gainesville, Florida.

ReceivedThere areregard toNo fundiAddressRehabilitKinesioloE-mail: sc

1526-590

ª 2013 P

http://dx

1492

Abstract: The goal in the current study was to examine the analgesic effects of a pinch grip-force

production task and a working memory task when pain-eliciting thermal stimulation was delivered

simultaneously to the left or right hand during task performance. Control conditions for visual

distraction and thermal stimulation were included, and force performance measures and working

memory performance measures were collected and analyzed. Our experiments revealed 3 novel

findings. First, we showed that accurate isometric force contractions elicit an analgesic effect

when pain-eliciting thermal stimulation was delivered during task performance. Second, the

magnitude of the analgesic effect was not different when the pain-eliciting stimulus was delivered

to the left or right hand during the force task or the working memory task. Third, we found no

correlation between analgesia scores during the force task and the working memory task. Our

findings have clinical implications for rehabilitation settings because they suggest that acute force

production by one limb influences pain perception that is simultaneously experienced in another

limb. From a theoretical perspective, we interpret our findings on force and memory driven analgesia

in the context of a centralized pain inhibitory response.

Perspective: This article shows that force production and working memory have analgesic effects

irrespective of which side of the body pain is experienced on. Analgesia scores were not correlated,

however, suggesting that some individuals experience more pain relief from a force task as compared

to a working memory task and vice versa.

ª 2013 Published by Elsevier Inc. on behalf of the American Pain Society

Key words: Force, muscle contraction, working memory, analgesia, pain, cognition.

ehabilitation often involves the execution ofpain-eliciting movements, with patients requiredto execute many movements that are 10 to 20

seconds long during a 30- to 60-minute session. Forexample, one treatment approach for rotator cuffinjuries is exercise therapy and this often requirespatients to execute shoulder movements that canelicit acute musculoskeletal pain.59 One possibility foraltering the experience of pain is to simultaneouslycontract muscles in a noninjured limb.27,29 Althoughendogenous pain responses associated with movementscan help a clinician to gauge the acuteness of the

April 5, 2013; Revised July 8, 2013; Accepted July 10, 2013.no conflicts of interest, or any financial interests, to report withthis work for any of the authors.

ng sources were provided for this study.reprint requests to Stephen A. Coombes, PhD, Laboratory foration Neuroscience, Department of Applied Physiology andgy, University of Florida, PO Box 118206, Gainesville, FL [email protected]

0/$36.00

ublished by Elsevier Inc. on behalf of the American Pain Society

.doi.org/10.1016/j.jpain.2013.07.012

injury or to refine the treatment approach, developingnonpharmacologic treatment approaches that can alterpain perception is also important. Indeed, althoughpharmacologic treatments to manage pain in thesecontexts are often effective, their use has also beenassociated with inadequate pain relief and issuesrelated to dependence and addiction.21,34,37 The goalin the current study is to use an experimental model toexamine the effects of acute nonpharmacologicapproaches that alter the perception of exogenous pain.Previous studies have demonstrated that muscle con-

tractions are associated with a reduction in painperception, and this effect has been termed exercise-induced hypoalgesia.15,27,42 Analgesia has beenassociated with relatively long muscle contractionsthat are performed between 2 repetitions of a pain-pressure test on the same19,22,25,31,57 or differentlimb,27,29 and with very brief pain-eliciting stimulithat are delivered during longer force contractions ofthe same or different limb.29,50,54 Although acutepain typically coincides with movement of the injuredlimb in the clinical setting, in previous experimental

Page 2: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Paris et al The Journal of Pain 1493

paradigms muscle contractions and pain elicitationhave not occurred at the same time and for a similarduration. This has prevented firm conclusions frombeing drawn on the existence of a centralized paininhibitory response when force production and painoccur simultaneously.Previous studies have also shown analgesic effects

when a working memory task is coupled with pain-eliciting stimulation to the left hand3 and leftforearm.4,5,53 Other studies show similar effects duringa Stroop task58 and a verbal attention task16 withstimulation delivered to the right forearm and rightfoot, respectively. Together, these studies suggestthat cognition also influences pain perception via acentralized pain inhibitory response that is independentof the spatial location of the pain, but this suggestion hasnot been directly tested.We have developed a novel experimental paradigm

that allows us to separately measure pinch grip-forceproduction and working memory performance duringthe delivery of a constant pain-eliciting thermal stim-ulus to the left or right hand. This paradigm allows usto control for visual distraction during both tasks42

and to examine whether a centralized pain inhibitoryresponse during each task is related. We tested 3hypotheses: 1) pain perception will be reduced whenforce is produced by the right hand and pain is deliv-ered to the right or left hand, 2) pain perception willbe reduced when a working memory task is performedwhile pain is delivered to the left or right hand, and3) analgesic scores associated with force productionand working memory will be positively related, suggest-ing that a common centralized pain inhibitory responsemay be associated with both tasks.

Methods

SubjectsThirty-eight healthy right-handed adults with

normal or corrected-to-normal vision and not takingany pain medications participated (19 female, 19male; mean = 20.78 y, range: 18–43 y). Each subjectprovided informed consent to all procedures, whichwere approved by the local institutional reviewboard and were in accord with the Declaration ofHelsinki. Exclusion criteria were any history of neuro-logic disease, including any history of chronic pain orany current episode of acute pain. Subjects whoverbally self-reported any history of these disorderswere excluded during prescreening. All subjectscompleted the state and trait segments of theState-Trait Anxiety Inventory (STAI-S, STAI-T),52 theBeck Depression Inventory,2 Pain Anxiety SymptomsScale,38 Pain Catastrophizing Scale,55 and the TampaScale of Kinesiophobia.28 All scores were within thenormal range of responses typically given by healthyadult subjects (STAI-S: mean = 28.5, SD = 9.20; STAI-T:mean = 33.30, SD = 8.87; Beck Depression Inventory:mean = 2.0, SD = 2.67; Pain Anxiety Symptoms Scale:mean = 54.83, SD = 17.4; Pain Catastrophizing Scale:

mean = 11.5, SD = 8.55, Tampa Scale of Kinesiopho-bia: mean = 29.78, SD = 4.25).2,44,52,56

Pinch Grip-Force TaskPinch grip-force was always produced by the index

finger and thumb of the right hand. During the practicesession, each subject’s maximum voluntary contraction(MVC) was estimated using a force transducer (JamarHydraulic Pinch Gauge; Lafayette Instrument Co,Lafayette, IN). Subjects were asked to sustain a contrac-tion of maximum force for three 5-second trials. Trialswere separated by 60-second periods of rest. The MVCwas calculated as the average of the 3 peak force levels.MeanMVC across subjects was 64.75 N (SD = 16.4 N). Theforce transducers used were ELFF-B4 model load cellsconstructed from piezoresistive strain gauges measuringforce up to 100 N (Measurement Specialties, Hampton,VA). Force datawere collected by Coulbourn InstrumentsType B V72-25 amplifiers (Coulbourn Instruments, Allen-town, PA) at an excitation voltage of 5 V. The force signalwas transmitted via a 16-bit analog-digital converter anddigitized at 125 Hz. The summed output from the 2 forcetransducers (index finger and thumb) was presented tothe subject using a visual display on the computer screen.The force outputwas displayed on a 40-inch liquid crystaldisplay screen at a resolution of 1,600 � 1,024 pixels anda refresh rate of 59 Hz. The subject sat 50 inches from thescreen. Force production was guided by real-time visualinformation consistent with the paradigm shown inFig 1A, and with previous force control studies.11,41

During the rest period and during the task, 2 bars werevisible to the subject. A white bar represented thetarget force level that was set at 25% of each subject’sMVC. A red/green force bar was controlled by thesubject and provided real-time visual information offorce production. When the force bar turned green,the subject’s goal was to match the green bar with thewhite bar. The subject produced force until the greenbar turned red. Each task was 15 seconds long andincluded 5 force pulses. Each pulse was 1.8 secondslong (bar turns green) and pulses were separated by a1.2-second rest period (bar turns red). Each trial wasfollowed by a 7.5-second rating period in which subjectsrated the level of pain they experienced during theprevious 15-second trial. Subjects were instructed tomake a rating after every trial. This was especiallyimportant for the trials where no stimulation wasdelivered because we wanted to determine whetherany pain was experienced as a function of thegrip-force task alone. Ratings were always made withthe left hand using a keyboard to control a cursor on avisual analog scale (VAS) presented on the screen. Asshown in Fig 1A, during experimental trials, 2 verbaldescriptors were visible to the subject: ‘‘no sensation’’on the left side of the scale and ‘‘intolerable pain’’ onthe right side of the scale. The range of the ratingbar was 0 to 10, with a resolution of .1. As shown inFig 1A, subjects did not see numbers on the scale.Subjects were familiarized with the rating scale duringa practice session. Fig 1B shows an example time series

Page 3: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Figure 1. Experimental paradigm for the force task and theworkingmemory task. (A) The red bar indicates rest. When the bar turnedgreen, subjects produced force during the active condition or watched the green bar move based on a prerecorded force trace from thesame subject during the passive condition. Subjectswere informedbefore each blockwhether the upcoming blockwas active or passive.During the active force task, 5 pulses of force (1.8 seconds each)were produced separated by 1.2-second periods of rest. Note that only 2pulses are shown in (A). After each15-secondtrial, subjects ratedthe levelof sensation theyfeltduring thepreceding trial. (B)Anexampleof the force pulses produced by 1 subject during 1 trial at each temperature condition (baseline, warm, hot). The black lines (left y-axis)represent the force produced by the subject and the red lines (right y-axis) reflect the temperature of the thermode during each repre-sentative active trial. (C)Working memory task. Rest was represented by a fixation cross. During each 15-second trial, a sequence of 10letterswas presentedon the screen, one at a time. Note that only 2 letters are shown in (C). During the active task, subjectswere requiredto count thenumberof timesa letterwas the sameas theonepresented2 letters previously. Theykept countof thenumberofhits duringeach trial and thenrespondedat theendofeach trial using their left handonakeyboard to control a slider onan integer scale. (D)Duringthe presentation of the letters, subjects also received baseline, warm, or hot stimulation (red lines, right y-axis). After the n-back rating,subjects reported the level of sensation they experiencedduring theprevious 15-second trial using the same scale as in the force task. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

1494 The Journal of Pain Effects of Force and Memory on Pain

of 1 trial of a subject’s force production (black lines: lefty-axis) overlapped with the temperature of the ther-mode (red lines: right y-axis) for the 3 differenttemperature conditions.

Working Memory TaskThe visual display during the working memory task is

shown in Fig 1C. Subjects completed a standard 2-backworking memory task, which is also termed an n-backtask. During each trial, a series of 10 white letters werepresented one at a time on a black background on thevisual display. Each letter was visible for 1 second andletters were separated by .5 seconds. Subjects were

required to count the number of hits, which was thenumber of times the presented letter was the same asthe letter presented 2 letters previously. Following eachtrial, subjects reported the number of hits during the15-second trial using their left hand to control akeyboard to move a cursor on an integer scale shownon the screen.53 The scale range was 0 to 5. Subjectshad 6 seconds to register their response. The correctnumber of hits in each trial was pseudo-randomizedacross conditions, and a unique series of letters waspresented during each trial. Next, during a 7.5-secondrating period, subjects used their left hand tomake an in-tensity rating on the same VAS as in the force condition.Aside from the additional n-back response time

Page 4: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Paris et al The Journal of Pain 1495

(6 seconds), the trial timing (15 seconds), rating duration(7.5 seconds), and rest period (7.5 seconds) werethe same for the force trials and theworkingmemory tri-als. Fig 1D shows the time series of the 3 temperaturetraces (red lines: right y-axis) relative to the n-backtask. Note that only 4 of the 10 letters are shown in thefigure.

Thermal StimulationThermal stimuli were delivered to the thenar

eminence of the hand using a 573-mm2 CHEPS thermode(PATHWAY System; Medoc Advanced Medical Systems,Ramat Yishai, Israel). Consistent with previous work,4,5

the pain threshold of each subject was determinedthrough a calibration protocol. Subjects first wentthrough a practice session where they were madefamiliar with the rating scale. The scale ranged from0 = ‘‘no sensation’’ to 10 = ‘‘intolerable pain.’’ Thecalibration paradigm consisted of 16 trials; theduration of each trial was 15 seconds. Eachtemperature between 41 and 48�C was delivered twicein a random order. After each trial, subjects ratedperceived pain intensity during the trial by moving acursor on a digital VAS. For each subject, ratings wereplotted against their corresponding temperatures. Toqualify for the study, an individual’s ratings of theapplied temperatures had to be consistent betweenratings of 2 and 10 (R2 > .65).5 Data from 2 subjects (1male, 1 female) were excluded from the analysis becauseof inconsistent ratings during the calibration period(R2 < .65). Hence, 36 participants (18 female, 18 male)were included in the analyses. The mean R2 value was.85 (SD = .06). Full calibrations were completed on thefirst hand that received stimulation (which was counter-balanced across subjects). After calibration, the tempera-ture corresponding to a rating of 6 to 8 was selected asthe hot temperature and the temperature correspond-ing to a rating of 2 to 3 was selected as thewarm temper-ature. When switching stimulation to the other hand, ashort calibration procedure was used to determinewhether the hot temperature was rated similarly acrossboth hands. In all subjects, the ratings were similar andso the same temperature was used for both hands forall subjects. The baseline temperature was set at 32�C.The mean warm temperature was 42.51�C (SD = .91)and the mean hot temperature was 46.01�C (SD = 1.97).For all tasks, the left and right handwere in a pronated

position. Thermal stimuli were delivered to the thenareminence of the hand. The thermode was strapped tothe hand using hook-and-loop, and cushioning wasplaced under the forearm to prevent downward pressurefrom the hand to the thermode. The thermode remainedin place during the entire block of trials and was notremoved prior to each rating being made. The grip-force apparatus (right hand) and the keyboard (lefthand) were raised and positioned so that when the ther-mode was attached to the hand, finger and thumbmovements occurred with minimal change in positionor pressure of the thermode on the hand.The LabVIEW program (National Instruments, Austin,

TX) that collected and displayed the force task and the

working memory task also controlled the timing of thethermal stimulation via a transistor-transistor logic pulseto the PATHWAY system. During each 15-second trial, thetemperature of the thermode was set at the baselinetemperature or at each subject’s warm or hot tempera-ture. During warm and pain-eliciting stimulation trials,the thermode surface was heated to the target tempera-ture at a rate of 70�C/s, and this temperature wasmaintained for the duration of the 15-second trial.The temperature returned to the baseline at a rate of40�C/s at the end of the 15-second trial. The fast rate ofheating and cooling of the thermode ensured that thethermal stimulation was synchronized with the forcetask. During the rest period between trials, the thermodewas set at the baseline temperature. Note that althoughthe thermal stimulation was constant for 15 seconds, theforce task cycled between force pulses and brief restperiods, and the working memory task presented lettersthat were also separated by brief rest periods. Hence,thermal stimulation was matched in terms of total trialduration (ie, 15 seconds), but brief rest periods didoccur within the force production task and theworking memory task during which thermal stimulationremained constant.

Experimental ParadigmThe experimental paradigm for the force production

task and the working memory task were designed tobe as consistent as possible in terms of task timing,hand position, stimulation duration, and control condi-tions. In addition to allowing us to examine analgesiaduring each task separately (hypotheses 1 and 2), ourapproach also allowed us to examine the relation inanalgesia scores between tasks (hypothesis 3). For theforce task and the working memory task, there wereactive and passive conditions. During active conditions,subjects produced force with their right hand orperformed the working memory task. During passiveconditions, subjects did not produce force or completethe working memory task but viewed a prerecordedforce trace or a series of letters that did not have to beremembered. Visual information was thereforeconsistent across active and passive conditions for eachtask. For the force task, the prerecorded force tracewas recorded during each participant’s practice session.During the practice session subjects produced forceduring each temperature condition to familiarizethemselveswith the experimental paradigm. Force traceswere recorded during accurate practice trials and wereplayed back to the subject later during experimentaltrials for the corresponding passive conditions. Thismeant that the passive condition could occur first inthe experimental paradigm and order effects wereavoided. Because of the relative ease of the task, subjectswere able to perform the task accurately during thepractice session after 1 or 2 trials. Thermal stimulationwas delivered during both active and passive conditions,and ratings were always made using the left hand inboth conditions. This allowed us to calculate analgesiascores while controlling for visual stimuli betweenconditions. Subjects completed a total of 8 blocks of

Page 5: Effects of a Force Production Task and a Working Memory Task on Pain Perception

1496 The Journal of Pain Effects of Force and Memory on Pain

experimental trials: 4 blocks for the force task and 4blocks for the working memory task. For each task, ther-mal stimulation was delivered to the left hand (2 blocks:1 active, 1 passive) and to the right hand (2 blocks: 1active, 1 passive). Within each block, the task type, thesite of stimulation, and the visual information wasconstant. The only independent variable that changedduring a block was the temperature of the thermalstimulation, which was at a baseline temperature (2 tri-als), a warm temperature (3 trials), or a hot temperature(3 trials). Temperatures within each block were pseudo-randomized such that the same temperature was neverpresented twice in a row. A 90-second rest interval sepa-rated trials 4 and 5 to provide a standardized rest periodwithin each block and to help avoid habituation, sensiti-zation, and fatigue affects. We also controlled for sensi-tization and habituation effects by pseudo-randomizingblock order with the following constraints: half the sub-jects completed the first 4 blocks with stimulation ontheir right hand and half completed the first 4 blockswith stimulation on their left hand. Half the subjectscompleted 2 blocks (active/passive) of the force task firstand half completed 2 blocks (active/passive) of the work-ingmemory task first. Finally, half the subjects completedthe passive block first, and half the subjects completedthe active block first. Nonsignificant paired t-tests be-tween the first and last rating, irrespective of task, forthe passive hot condition on each hand suggested thathabituation across trials was not an issue (left hand: t[35] = 1.95, P > .05; right hand: t[35] = �.954, P > .05).

Data Analysis

Force Task

Force data was analyzed using custom algorithms inLabVIEW. The force–time series data were digitallyfiltered using a fourth-order Butterworth filter with a20-Hz low-pass cut-off. Force production was character-ized bymeasures ofmean force amplitude and force vari-ability (SD). Force amplitude and force variability wereextracted from the middle 1 second of each pulse. Themiddle 1 second of the pulse was calculated followingidentification of the onset and offset of each pulse.The onset of each contraction was identified as thetime point where the force rose above 2 times the base-line value. The offset of each contraction was identifiedas the next time point where force fell below 2 times thebaseline value. Baseline values were calculated as themean force amplitude during the 300 ms prior toeach 15-second block. The summary statistic foreach force-dependent variable was the mean scorecalculated from all the pulses in each condition.Separate 2-way (temperature [baseline, warm,hot] � site [left stimulation, right stimulation]) repeatedmeasures analyses of variance (ANOVAs) were run foreach dependent variable.

Working Memory Task

Consistent with prior work, task performance wascalculated as the ratio between blocks with a correctly

indicated number of n-back targets and the totalnumber of blocks per condition.53 The summary statisticwas themean ratio score for each condition.Mean scoresfor each condition were compared with a 2-wayrepeated measures ANOVA (temperature [baseline,warm, hot] � site [left stimulation, right stimulation]).

Rating Task

Intensity ratings were made after every trial. A meanrating score was calculated from all trials in eachcondition. Our first 2 hypotheses predicted that forceproduction and working memory would influence theperception of a pain-eliciting stimulus delivered to theleft and right hand. We therefore ran a 3-way ANOVA(task: [force production, working memory] � site[left, right] � engagement [active, passive]) to examinedifferences in absolute ratings across all conditionswhen hot thermal stimulation was delivered. Controlanalyses used the same ANOVA model to examineratings between conditions during baseline and warmtemperature trials. Our third hypothesis predicted thatanalgesic scores during the force production task andthe working memory task would be related. To test thishypothesis, we first subtracted ratings during each activecondition from ratings during its corresponding passivecondition at the hot temperature. We then performeda correlation analysis between scores for the forceproduction task and the working memory task usingPearson correlation coefficient.For all analyses, significant effects were followed up

with paired t-tests corrected for multiple comparisonsusing the Bonferroni correction. For analyses in whichthe sphericity assumption was violated, Greenhouse-Geisser degrees of freedom corrections were appliedand are reported. Significance level was set at P < .05.

Results

Pinch Grip-Force TaskMean force amplitude and mean variability for all

conditions is shown in Table 1. The amplitude of forceproduction did not differ as a function of temperature(F[1.4, 49.01] = .39, P = .608), site (F[1, 35] = 2.79,P = .104), or an interaction between temperature andsite (F[1.45, 50.79] = 1.95, P = .163). Mean force acrossall conditions was 24.54% of MVC. Mean variability(.91% of MVC) also did not vary as a function oftemperature (F[1, 35] = 1.99, P = .143), site (F[1,35] = 1.99, P = .167) or an interaction between them(F[2, 70] = 1.40, P = .253).

Working Memory TaskOne subjectwas not included in the analysis because of

incomplete scoring data. Ratio scores from each site andtemperature condition were compared in the 2-wayrepeated measures ANOVA. There was no main effectof temperature (F[2, 68] = 1.91, P = .156) or site (F[1,34] = .77, P = .387), and the interaction of temperatureand site was not significant (F[2, 68] = .16, P = .853).

Page 6: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Table1. Performance andRatingData forActive andPassiveConditions for the Force Production Taskand theWorkingMemory Task for Each Site (Left, Right) and Each Temperature (Baseline,Warm,Hot)

SITE

ACTIVE PASSIVE

BASELINE WARM HOT BASELINE WARM HOT

M SD M SD M SD M SD M SD M SD

Force task

Amplitude (%MVC) Left 24.56 .65 24.57 .58 24.71 .56 - - - - - -

Right 24.49 .61 24.46 .69 24.43 .71 - - - - - -

Variability (%MVC) Left .87 .05 .89 .05 .88 .05 - - - - - -

Right .86 .05 .97 .05 .96 .06 - - - - - -

Rating Left .44 .82 2.16 1.51 6.41 1.81 .38 .73 2.31 1.35 7.02 1.48

Right .21 .40 2.21 1.37 5.90 1.89 .27 .57 2.63 1.54 6.73 1.62

Memory task

Performance (ratio) Left .59 .35 .67 .34 .58 .30 - - - - - -

Right .56 .40 .60 .31 .54 .30 - - - - - -

Rating Left .44 .58 2.18 1.57 5.71 1.86 .35 .81 2.15 1.50 6.73 1.56

Right .46 .56 2.43 1.54 5.60 1.73 .50 .88 2.30 1.30 6.86 1.69

Abbreviations: M, mean; SD, standard deviation.

NOTE. Performance data were not collected during passive conditions. Variability during the force task represents the standard deviation of the force output.

Paris et al The Journal of Pain 1497

The overall mean ratio across all conditions was .59(SE = .039).

Figure 2. Rating data for the force production task (left) andthe working memory task (right) for passive (P) and active (A)trials at each temperature. Data are collapsed across the leftand right sites. Different temperature conditions arerepresented by circles (hot), triangles (warm), and squares(baseline). As expected, ratings increased as a function of tem-perature, and analgesic effects were not found at warm orbaseline temperatures for either task. At the hottemperature, ratings during both tasks were significantlyhigher for the passive trials as compared to active trials. Errorbars represent 1 SD. The asterisk represents a significantt-test finding (*P < .05).

Rating DataThe primary aim was to examine the effect of task,

site, and engagement on stimulation intensity ratingswhen a hot pain-eliciting stimulus was delivered tothe subject during task performance. The correspond-ing 3-way repeated measures ANOVA revealed a signif-icant main effect of engagement (F[1, 35] = 48.97,P < .001) but not task (F[1, 35] = 3.72, P > .05) or site(F[1, 35] = .63, P > .05). The main effect of engagementwas superseded by a significant task � engagementinteraction (F[1, 35] = 7.93, P < .01). Data werecollapsed across site, and paired t-tests comparingactive and passive conditions revealed significanteffects of engagement for the force task (t[35] = 5.43,P < .001) and the working memory task (t[35] = 5.67,P < .001). Fig 2 shows rating data for the forceproduction task (left) and the working memory task(right) for passive (P) and active (A) trials at eachtemperature. Data are collapsed across left and rightsite. The temperature conditions are represented bycircles (hot), triangles (warm), and squares (baseline).For the force task, 31/36 (86%) subjects had a positiveanalgesic score (passive rating > active rating), with rat-ings decreasing by 11% from the passive to the activetask. For the working memory task, 32/36 subjects(89%) had positive analgesic scores and ratingsdecreased by 17% from the passive to the active task.A control analysis was used to examine ratings at thewarm temperature and a significant task � engage-ment interaction was found (F[1, 35] = 4.826, P < .05).Follow-up t-tests comparing active and passive condi-tions for each task were not significant. Main effectsof site, task, and engagement were not significant at

the warm temperature. At the baseline temperatureno significant main effects or interactions were found.Our third hypothesis was designed to test the relation

between analgesic scores for the force production taskand the working memory task. For each task, wesubtracted the active score (hot condition) from the

Page 7: Effects of a Force Production Task and a Working Memory Task on Pain Perception

1498 The Journal of Pain Effects of Force and Memory on Pain

passive score (hot condition) and then tested for acorrelation between scores. The correlation was notsignificant (r = .092, n = 36, P > .05).

DiscussionThe goal in the current study was to examine the

analgesic effects of force production and a workingmemory task when pain-eliciting thermal stimulationwas delivered to the left or right hand duringtask performance. Our experiments revealed 3 novelfindings. First, we showed that acute isometric forcecontractions elicit an analgesic effect when pain-eliciting thermal stimulation was delivered during taskperformance. Second, the magnitude of the analgesiceffect was not different when the pain-eliciting stim-ulus was delivered to the left or right hand during theforce task and the working memory task. Third, wefound no correlation between analgesia scores duringthe force production task and the working memorytask.During rehabilitation, patients are often required to

make movements that are short in duration, have to berepeated frequently, and elicit pain for the durationof the movement. The current study was partiallymotivated by these observations because althoughprevious studies have demonstrated analgesic effects offorce contractions, they have done so using serialparadigms in which the contraction and the pain donot occur simultaneously.19,25,27 We extend thesefindings by demonstrating an analgesic effect whenacute force production and pain-eliciting stimuli aredelivered simultaneously, for similar amounts of time,and involve the same or different hand. Our findingscontrast a previous study that did not find analgesiceffects when a 5-second pain-eliciting stimulus wasdelivered 30 or 60 seconds after the onset of acontraction but did find effects when the same stimulusoccurred 90 seconds after onset of the contraction.54 Onepossibility is that analgesia can occur during the first15 seconds of a contraction and then dissipate beforereemerging 90 seconds after onset. An alternativeexplanation is that a pulsing pinch grip-force taskincreases motor system activity to a greater extent thana sustained contraction,43 which led to an analgesiceffect emerging earlier in the current study. Othermethodological differences may also account for thecontrasting findings. In the current study, the durationof the thermal stimulus was 15 seconds as compared to5 seconds. Subjects in the current study also reportedstimulus intensity after the trial using a digital VAS ascompared to a verbal report after each stimulus butnot during the contraction. We also used a passive con-trol condition to account for visual distraction asopposed to a baseline condition without a visual control.Finally, our findings extend previous work because theyare the first to collect and analyze force production dur-ing the delivery of a pain-eliciting stimulus. Subjectsproduced accurate force during all trials, allowing us toconclude that our analgesic effects were not related towithin-trial changes in force production.

Our findings have theoretical relevance because theysuggest that analgesia related to force production isassociated with activity in the central nervous system.Three explanations may account for our findings:1) diffuse noxious inhibitory control, 2) distraction, and3) a centralized pain inhibitory response. First, a diffusenoxious inhibitory control interpretation would requirea second painful stimulus to occur simultaneously withthe pain-eliciting stimulus. We used a submaximal targetforce level in the current study to avoid the possibilitythat the force production task itself could becomepainful. Our approach was confirmed by near-zeroratings during the force-baseline condition, whichshowed that the pinch grip-force task alone did not elicitpain. A second explanation for our finding is distraction,either due to the visual stimulus or due to the parallelexperimental paradigm employed. One advantage ofthe parallel experimental design is that passiveconditions can be matched to active conditions suchthat thermal stimulation is paired with visual informa-tion but force is not produced. This allowed us to ruleout visual distraction as an explanation for differencesbetween the active and passive conditions. Given thatforce production is one task and evaluating thestimulation could be considered a second task, wecontrolled for a dual-task interpretation by including aforce-warm condition. If lower ratings in the activeforce–hot condition were due to dual-task effects, onewould expect the active force–warm condition to alsobe rated lower than its corresponding passive controlcondition. This was not the case. Accounting for visualdistraction and dual-task conditions is consistent with arecent review article that supported the use ofappropriate control conditions when examiningexercise-induced analgesia.42

A third explanation relates to a centralized pain inhibi-tory response that suggests that engaging motor circuitsinfluences activity in pain circuits at the level of the spinalcord and/or brain. This suggestion is based on ourfinding that analgesia was not different when force pro-duction and pain-eliciting stimuli were delivered to thesame as compared to different hands, which is consistentwith previous studies that used serial27,47 and parallelexperimental paradigms.29,54 The gate control theorysuggests that the sensory input generated by forceproduction may have activated inhibitory interneuronsthat inhibited pain signals prior to them reaching thespinal cord. However, this theory cannot explain theanalgesic effect that emerged when force productionand pain-eliciting stimuli occurred simultaneously ondifferent hands. Direct inhibition at the spinal level viathe crossing of sensory and motor neurons is also a pos-sibility, but the available evidence suggests that acentralized inhibitory circuit is the more likely explana-tion. This interpretation is consistent with repetitivetranscranial magnetic stimulation studies (rTMS) thathave shown that rTMS over one side of primary motorcortex (M1) leads to an analgesic effect on both sidesof the body even though no force isproduced.18,32,33,35,46 This suggests that M1 may be anentry port for the modulation of pain irrespective of its

Page 8: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Paris et al The Journal of Pain 1499

spatial location. Previous neuroimaging studies havedemonstrated M1 activity during a pinch grip-forcetask similar to the one used in the currentstudy.8-10,14,23,43,60 Hence, our analgesic effects mayhave been driven by increased activity in motor circuitsthat influenced activity in downstream pain processingnetworks that include midcingulate cortex, insula,and periaqueductal gray.1,6,12,13,30,36,39,48,51 Althoughstimulation of M1 has been shown to attenuate chronicpain in conditions such as fibromyalgia, analgesiceffects associated with voluntary isometric contractionsin patient populations have been mixed.20,22,31,54

These observations, coupled with differences betweenexogenous and endogenous pain, voluntary andinvoluntary engagement of the motor system, andpotential age- and sex-related differences, suggest thattranslating the current findings to clinical populationsmay not be straightforward. Future studies thatreplicate our protocol but also assess other potentialmechanisms such as skin temperature, increasedsecretion of b-endorphins, and blood pressure22,24,26,42

may help translate the current findings to clinicalpopulations.Although previous studies have examined the effect of

working memory on pain perception,3-5,53 they have notmanipulated the side of stimulation in the sameindividual. Here, we extend these findings bydemonstrating that a working memory task has ananalgesic effect irrespective of which side of the bodythe pain is experienced on. Our findings haveimplications for advancing the shared resources modelof pain and cognitive performance because theydemonstrate that cognitive performance and/or painprocessing share resources,4 consistent with neuroimag-ing evidence that working memory and pain processesare characterized by overlapping bilateral brainactivity.17,45 However, we found no evidence of apain-related decrement in working memory perfor-mance as subjects were able to perform the task withsimilar accuracy across all stimulation intensities.7,49

Performance ratios were consistent with a recentstudy that used a similar paradigm and scoringmethod.53 However, we are cognizant of work that hasdemonstrated pain-related decrements in cognitiveperformance.3-5,40,53 Task difficulty may account forwhy decrements in performance were not found in thecurrent study. Pairing pain-eliciting stimulation with

low and high working memory load tasks has shownimproved performance on low- as compared to high-load tasks and attenuated pain ratings for high- ascompared to low-load tasks.3,53 Effects of pain on taskperformance are mixed, however, and may be drivenby task difficulty such that harder tasks are moresusceptible to pain-related decrements as compared toeasier tasks.4,5 Together these findings suggest that thetask used in the current study was challenging enoughto drive an analgesic effect but not so challengingthat pain-related decrements in working memoryperformance emerged.The current study has a number of limitations. First, it

was designed to assess the effects of acute motor andcognitive tasks on exogenous pain perception. Hence,we did not examine the duration of the analgesic effect,which is an important question that needs to be ad-dressed. Second, our findings were limited to self-reported pain ratings, and physiological indices such asblood pressure and secretion of b-endorphins were notcollected. Third, our experiment did not assess other fac-tors that may alter pain ratings such as skin temperatureor factors that indicate whether the analgesic effectswere associated with changes in the spinal cord andbrain. Fourth, the generalizability of our findings toendogenous pain conditions is limited because we as-sessed changes in pain perception to an acute exogenousthermal stimulus in healthy adults. Future studies arenecessary to address these issues.Our findings suggest that both the force production

task and the working memory task engaged a central-ized inhibitory response, but no relation was foundbetween the analgesia scores for each task. This findingmay be explained by differences in trying to manipulatea visual target during the force task, which engages thevisuomotor system,8-10 and trying to remember andcount visually displayed letters, which engages theworking memory system.3,58 Hence, although eachsystem influences exogenous pain perception at thegroup level, at the individual level engaging onesystem instead of the other may have a greaterinfluence on pain perception. Our findings thereforehave implications for pain management approachesbecause they demonstrate that certain individuals mayengage a centralized pain inhibitory response more soduring a force production task as compared to aworking memory task and vice versa.

References

1. Augustine JR: Circuitry and functional aspects of theinsular lobe in primates including humans. Brain Res BrainRes Rev 22:229-244, 1996

2. BeckAT, Steer RA: BeckDepression InventoryManual. SanAntonio, TX, The Psychological Corp, 1987

3. Bingel U, Rose M, Glascher J, Buchel C: fMRI reveals howpain modulates visual object processing in the ventral visualstream. Neuron 55:157-167, 2007

4. Buhle J, Wager TD: Performance-dependent inhibition ofpain by an executive working memory task. Pain 149:19-26,2010

5. Buhle JT, Stevens BL, Friedman JJ, Wager TD: Distractionand placebo: Two separate routes to pain control. PsycholSci 23:246-253, 2012

6. Cauda F, D’Agata F, Sacco K, Duca S, Geminiani G,Vercelli A: Functional connectivity of the insula in the restingbrain. Neuroimage 55:8-23, 2011

7. Coen SJ, Aziz Q, Yaguez L, Brammer M, Williams SC,Gregory LJ: Effects of attention on visceral stimulus intensity

Page 9: Effects of a Force Production Task and a Working Memory Task on Pain Perception

1500 The Journal of Pain Effects of Force and Memory on Pain

encoding in the male human brain. Gastroenterology 135:2065-2074, 2008. 2074.e1

8. Coombes SA, Corcos DM, Pavuluri MN, Vaillancourt DE:Maintaining force control despite changes in emotionalcontext engages dorsomedial prefrontal and premotorcortex. Cereb Cortex 22:616-627, 2012

9. Coombes SA, Corcos DM, Sprute L, Vaillancourt DE:Selective regions of the visuomotor system are related togain-induced changes in force error. J Neurophysiol 103:2114-2123, 2010

10. Coombes SA, Corcos DM, Vaillancourt DE: Spatiotem-poral tuning of brain activity and force performance.Neuroimage 54:2226-2236, 2011

11. Coombes SA, Gamble KM, Cauraugh JH, Janelle CM:Emotional states alter force control during a feedbackoccluded motor task. Emotion 8:104-113, 2008

12. De Oca BM, DeCola JP, Maren S, Fanselow MS: Distinctregions of the periaqueductal gray are involved in theacquisition and expression of defensive responses. J Neuro-sci 18:3426-3432, 1998

13. Dum RP, Levinthal DJ, Strick PL: The spinothalamicsystem targets motor and sensory areas in the cerebralcortex of monkeys. J Neurosci 29:14223-14235, 2009

14. Ehrsson HH, Fagergren A, Jonsson T, Westling G,Johansson RS, Forssberg H: Cortical activity in precision-versus power-grip tasks: An fMRI study. J Neurophysiol 83:528-536, 2000

15. Ellingson LD, Colbert LH, Cook DB: Physical activity isrelated to pain sensitivity in healthy women. Med Sci SportsExerc 44:1401-1406, 2012

16. Frankenstein UN, Richter W, McIntyre MC, Remy F:Distraction modulates anterior cingulate gyrus activationsduring the cold pressor test. Neuroimage 14:827-836, 2001

17. Friebel U, Eickhoff SB, Lotze M: Coordinate-basedmeta-analysis of experimentally induced and chronic persis-tent neuropathic pain. Neuroimage 58:1070-1080, 2011

18. Hirayama A, Saitoh Y, Kishima H, Shimokawa T,Oshino S, Hirata M, Kato A, Yoshimine T: Reduction ofintractable deafferentation pain by navigation-guidedrepetitive transcranial magnetic stimulation of the primarymotor cortex. Pain 122:22-27, 2006

19. Hoeger Bement MK, Dicapo J, Rasiarmos R, Hunter SK:Dose response of isometric contractions on pain perceptionin healthy adults. Med Sci Sports Exerc 40:1880-1889, 2008

20. Hoeger Bement MK, Weyer A, Hartley S, Drewek B,Harkins AL, Hunter SK: Pain perception after isometricexercise in women with fibromyalgia. Arch Phys MedRehabil 92:89-95, 2011

21. Juurlink DN, Dhalla IA: Dependence and addictionduring chronic opioid therapy. JMed Toxicol 8:393-399, 2012

22. Kadetoff D, Kosek E: The effects of static muscularcontraction on blood pressure, heart rate, pain ratings andpressure pain thresholds in healthy individuals and patientswith fibromyalgia. Eur J Pain 11:39-47, 2007

23. Keisker B, Hepp-Reymond M-C, Blickenstorfer A,Kollias SS: Differential representation of dynamic andstatic power grip force in the sensorimotor network.Eur J Neurosci 31:1483-1491, 2010

24. Koltyn KF: Analgesia following exercise: A review.Sports Med 29:85-98, 2000

25. Koltyn KF, Trine MR, Stegner AJ, Tobar DA: Effect ofisometric exercise on pain perception and bloodpressure in men and women. Med Sci Sports Exerc 33:282-290, 2001

26. Koltyn KF, Umeda M: Exercise, hypoalgesia and bloodpressure. Sports Med 36:207-214, 2006

27. Koltyn KF, Umeda M: Contralateral attenuation ofpain after short-duration submaximal isometric exercise.J Pain 8:887-892, 2007

28. Kori SH, Miller RP, Todd DD: Kinesiophobia: A new viewof chronic pain behavior. Pain Manag 3:35-43, 1990

29. Kosek E, Lundberg L: Segmental and plurisegmentalmodulation of pressure pain thresholds during static musclecontractions in healthy individuals. Eur J Pain 7:251-258,2003

30. Kurth F, Zilles K, Fox PT, Laird AR, Eickhoff SB:A link between the systems: Functional differentiationand integration within the human insula revealed bymeta-analysis. Brain Struct Funct 214:519-534, 2010

31. Lannersten L, Kosek E: Dysfunction of endogenous paininhibition during exercise with painful muscles in patientswith shoulder myalgia and fibromyalgia. Pain 151:77-86,2010

32. Lefaucheur JP, Drouot X, Menard-Lefaucheur I,Keravel Y, Nguyen JP: Motor cortex rTMS restoresdefective intracortical inhibition in chronic neuropathicpain. Neurology 67:1568-1574, 2006

33. Leo RJ, Latif T: Repetitive transcranial magneticstimulation (rTMS) in experimentally induced and chronicneuropathic pain: A review. J Pain 8:453-459, 2007

34. Li JX, Zhang Y: Emerging drug targets for paintreatment. Eur J Pharmacol 681:1-5, 2012

35. Lima MC, Fregni F: Motor cortex stimulation for chronicpain: Systematic review and meta-analysis of the literature.Neurology 70:2329-2337, 2008

36. Linnman C, Moulton EA, Barmettler G, Becerra L,Borsook D: Neuroimaging of the periaqueductal gray: Stateof the field. Neuroimage 60:505-522, 2012

37. Martell BA, O’Connor PG, Kerns RD, Becker WC,Morales KH, Kosten TR, Fiellin DA: Systematic review: Opioidtreatment for chronic back pain: Prevalence, efficacy, andassociation with addiction. Ann Intern Med 146:116-127,2007

38. McCracken LM, Zayfert C, Gross RT: The Pain AnxietySymptoms Scale: Development and validation of a scale tomeasure fear of pain. Pain 50:67-73, 1992

39. MesulamM-M, Mufson EJ: The insula of Reil in man andmonkey: Architectonics, connectivity, and function, inJones EG, Peters A (eds): Cerebral Cortex, Vol 4: Associationand Auditory Cortices. New York, Plenum Press, 1985, pp179-228

40. Moriarty O, McGuire BE, Finn DP: The effect of pain oncognitive function: A review of clinical and preclinicalresearch. Prog Neurobiol 93:385-404, 2011

41. Naugle KM, Coombes SA, Cauraugh JH, Janelle CM:Influence of emotion on the control of low-level forceproduction. Res Q Exerc Sport 83:353-358, 2012

42. Naugle KM, Fillingim RB, Riley JL 3rd: A meta-analyticreview of the hypoalgesic effects of exercise. J Pain 13:1139-1150, 2012

Page 10: Effects of a Force Production Task and a Working Memory Task on Pain Perception

Paris et al The Journal of Pain 1501

43. Neely KA, Coombes SA, Planetta PJ, Vaillancourt DE:Segregated and overlapping neural circuits exist for theproduction of static and dynamic precision grip force.Human Brain Mapp 34:698-712, 2013

44. OsmanA, Barrios FX, Kopper BA, HauptmannW, Jones J,O’Neill E: Factor structure, reliability, and validity of the PainCatastrophizing Scale. J Behav Med 20:589-605, 1997

45. Owen AM, McMillan KM, Laird AR, Bullmore E: N-backworking memory paradigm: A meta-analysis of normativefunctional neuroimaging studies. Human Brain Mapp 25:46-59, 2005

46. Passard A, Attal N, Benadhira R, Brasseur L, Saba G,Sichere P, Perrot S, Januel D, Bouhassira D: Effects ofunilateral repetitive transcranial magnetic stimulation ofthe motor cortex on chronic widespread pain in fibromyal-gia. Brain 130(Pt 10):2661-2670, 2007

47. Persson AL, Hansson GA, Kalliomaki A, Moritz U,Sjolund BH: Pressure pain thresholds and electromyograph-ically defined muscular fatigue induced by a muscularendurance test in normal women. Clin J Pain 16:155-163,2000

48. Picard N, Strick PL: Imaging the premotor areas. CurrOpin Neurobiol 11:663-672, 2001

49. Pud D, Sapir S: The effects of noxious heat, auditorystimulation, a cognitive task, and time on task on painperception and performance accuracy in healthy volunteers:A new experimental model. Pain 120:155-160, 2006

50. Ring C, Edwards L, Kavussanu M: Effects of isometricexercise on pain are mediated by blood pressure.Biol Psychol 78:123-128, 2008

51. Shackman AJ, Salomons TV, Slagter HA, Fox AS,Winter JJ, Davidson RJ: The integration of negative affect,pain and cognitive control in the cingulate cortex. Nat RevNeurosci 12:154-167, 2011

52. Spielberger CD: Manual for the State-Trait AnxietyInventory (STAI). Palo Alto, CA, Consulting PsychologistsPress, 1983

53. Sprenger C, Eippert F, Finsterbusch J, Bingel U, Rose M,Buchel C: Attention modulates spinal cord responses topain. Curr Biol 22:1019-1022, 2012

54. Staud R, Robinson ME, Price DD: Isometric exercise hasopposite effects on central painmechanisms in fibromyalgiapatients compared to normal controls. Pain 118:176-184,2005

55. SullivanMJL, Bishop SR, Pivik J: The Pain CatastrophizingScale: Development and validation. Psychol Assess 7:524-532, 1995

56. Thomas JS, France CR, Lavender SA, Johnson MR: Effectsof fear ofmovement on spine velocity and acceleration afterrecovery from low back pain. Spine (Phila Pa 1976) 33:564-570, 2008

57. Umeda M, Newcomb LW, Ellingson LD, Koltyn KF:Examination of the dose-response relationship betweenpain perception and blood pressure elevations induced byisometric exercise in men and women. Biol Psychol 85:90-96, 2010

58. Valet M, Sprenger T, Boecker H, Willoch F, Rummeny E,Conrad B, Erhard P, Tolle TR: Distraction modulatesconnectivity of the cingulo-frontal cortex and the midbrainduring pain—An fMRI analysis. Pain 109:399-408, 2004

59. van der Meijden OA, Westgard P, Chandler Z, Gaskill TR,Kokmeyer D, Millett PJ: Rehabilitation after arthroscopicrotator cuff repair: Current concepts review and evidence-based guidelines. Int J Sports Phys Ther 7:197-218, 2012

60. Wasson P, Prodoehl J, Coombes SA, Corcos DM,Vaillancourt DE: Predicting grip force amplitude involvescircuits in the anterior basal ganglia. Neuroimage 49:3230-3238, 2010