The effect of paired associative stimulation on fatigue resistance

The effect of paired associative stimulation on fatigue resistance

Susanne Kumpulainen

, Jussi Peltonen

, Markus Gruber

, Andrew Cresswell

, Sinikka Peurala

, Vesa Linnamo

, Janne Avela

aNeuromuscularResearchCenter,DepartmentofBiologyofPhysicalActivity,UniversityofJyväskylä,Finland

bSensorimotorPerformanceLab,DepartmentofSportScience,UniversityofKonstanz,Germany

cSchoolofHumanMovementStudies,UniversityofQueensland,Australia

dValidiaRehabilitation,Lahti,Finland

Keywords:

Fatigue Centralfatigue Motorcortex

Transcranialmagneticstimulation Pairedassociativestimulation

a b s t r a c t

Pairedassociativestimulation(PAS)isanon-invasivestimulationmethoddevelopedtoinducebidirec- tionalchangesintheexcitabilityofthecorticalprojectionstothetargetmuscles.However,veryfew studieshaveshownanassociationbetweenchangesinmotorevokedpotentials(MEP)afterPASand behavioralchangesinhealthysubjects.Inthepresentstudywehypothesizedthatthefunctionalrele- vanceofPAScanbeseenduringfatiguingexercise,sincethereisalwaysacentralcontributiontothe developmentoffatigue.Transcranialmagneticstimulationwasappliedoverthemotorcortextomea- surechangesintheMEPsofthesoleusmusclebeforeandafterPAS.Furthermore,fatigueresistancewas testedduring15ssustainedmaximalisometriccontractionsbeforeandafterPAS.Onaverage,fatigue resistancedidnotchangeafterPAS,howeverthechangeinexcitabilitycorrelatedsignificantlywiththe changeinfatigueresistance.Discussion:FunctionalityofPASinterventionwasnotdemonstratedinthis study.However,theobservedrelationshipbetweenexcitabilityandfatigueresistancesuggeststhatPAS mighthaveaffectedcentralfatigueduringshortmaximalcontractions.

1. Introduction

Pairedassociativestimulation(PAS)isanon-invasivemethod developedtoinducebidirectionalchangesintheexcitabilityofthe corticalprojectionstothetargetmuscles.PAScombineselectrical stimulationof aperipheralsomatosensory nervewithtranscra- nial magnetic stimulation (TMS) over the contralateral motor cortex. Depending on the interstimulus interval (ISI), PAS can produceeitherlong-termpotentiation(LTP)–orlong-termdepres- sion(LTD) – likeplasticityin thetarget synapse(Kumpulainen et al.,2012; Stefan et al., 2000;Wolters et al.,2003), showing properties such as rapid onset, associativity, duration, speci- ficity, and NMDA-receptor dependence (Ziemann et al., 2008).

Thus, spike-timing dependentplasticityis considered themost likelymechanismbehindPAS(Stefanetal.,2000;Woltersetal., 2003).PASand motortraining havebeenshowntoshare com- monneural mechanisms,which suggeststhat PAScan beused as a test for functionally relevant neuronal circuits within the motor cortex (Jung and Ziemann, 2009; Stefan et al., 2006;

∗ Correspondingauthorat:DepartmentofBiologyofPhysicalActivity,Box35, 40014UniversityofJyväskylä,Finland.Tel.:+358503049059;fax:+358142602071.

E-mailaddress:susanne.kumpulainen@jyu.fi(S.Kumpulainen).

Ziemann et al., 2004). However, very few studies have shown functionalityofPAS-inducedexcitabilitychangesinhealthysub- jects(Frantsevaetal.,2008;JungandZiemann,2009;Rajjietal., 2011).

FunctionalrelevanceofPAScanprobablybeseenduringfatigu- ingexercisessinceitiswellknownthatbothcentralandperipheral factorscontributetothedevelopmentoffatigue(Gandevia,2001).

Fatiguecanbedefinedasanyexercise-inducedreductioninthe abilityofamuscletogeneratemaximalforceorpower(Gandevia, 2001).Central fatigue referstoprocesses proximal tothe neu- romuscular junction and peripheral fatigue to processes at or distal to it (Gandevia, 2001). The relative contribution of the central and peripheral components depends on the intensity andduration ofthefatiguingexercise.Shortmaximal sustained contractionshavebeenshowntohaveasubstantialcentralcon- tribution tothedevelopment of fatigue(Gandeviaet al.,1996;

Hunter et al., 2006, 2008; Lentz and Nielsen, 2002; Szubski etal., 2007;Tayloret al.,1996,1999).Central fatiguehasbeen definedasaprogressivereductioninthevoluntaryactivationof amuscleduringexerciseanditcanoriginateatbothspinaland supraspinallevels(Gandevia,2001).Previousstudiessuggestthat centralfatigueatleastpartiallyoriginatesfrominadequatecortical drivetothemotorneurons(Gandevia,2001;Hunteretal.,2006, 2008).

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-0-283897 Erschienen in: Neuroscience Research ; 95 (2015). - S. 59-65

Fig.1.Schematicofthemeasurementsetupintheankledynamometer.(a)Astim- ulatingcoilwasplacedandsecuredoverthelefthemisphereand(b)subject’sneck wascomfortablysupportedbyaheadrest.(c)Bodymovementwasrestrictedwith seatbeltsandakneestrap.(d)Rightfootwasinstalledtoaforcepedaland(e)the forcewasdisplayedonacomputerscreeninfrontofthesubject.

BecausePAScanbeusedtosystematicallyaltertheresponsive- nessofneuronsintheprimarymotorcortex,thecurrentstudywas designedtoinvestigatetheeffectoftwodifferentPASinterventions onfatigueresistanceduring15ssustainedmaximalisometriccon- tractions.ThePASinterventionstargetedthesoleusmuscle(SOL) asthisisanimportantantigravitymuscleduringstandinganda majorcontributortoforceproductionduringtheimpactphaseof walkingandrunning(Ishikawaetal.,2005).Itwashypothesized thatafterPAS_LTPfatigueresistancewouldincreasewhereasafter PASLTDfatigueresistancewoulddecrease.

2. Materialsandmethods 2.1. Ethicalapprovalandsubjects

Thirtyhealthy subjectsvolunteered for the study and were dividedinto two groups:PASinduced LTP-like plasticitygroup (PASLTP; 9 females and 6 males, 25±4 years, 62±10kg, 168±11cm)andLTD-likeplasticitygroup(PAS_LTD;9femalesand6 males,25±4years,63±7kg,168±7cm).Subjectswereblindedto thePASinterventiontheywereundergoingandthusthePASLTPand PAS_LTDwereconsideredaseachother’scontrol.Noneofthesub- jectshadanyhistoryofneuromuscularororthopedicdiseasesand allsubjectswerenaïvetotheexperiments.Beforetesting,subjects wereinformedabouttheproceduresand gavewrittenconsent.

ThestudywasapprovedbytheethicsboardfromtheUniversityof Jyväskyläandwasperformedinconformitywiththelatestrevision ofthedeclarationofHelsinki.

2.2. Experimentaldesign

Participants were positioned on a custom built ankle dynamometer (University of Jyväskylä, Finland) with the hip at110^◦,thekneeinanextendedpositionat180^◦,theankleat90^◦ andtherightfootrestingonapedal(Fig.1).Aseatbeltrestricted movement of theupper body and strapssecured theright leg andfoot.Handswererestinginthelapduringallmeasurements.

Prior to the measurements, the participants performed three

procedureswereconductedsothatfatiguewouldnotaffectMEPs orPASintervention.Toavoidpossiblefatigueeffectstherewasat least40minbetweenthelastfatiguingcontractionandthePAS interventioninthepremeasurements.Inthepostmeasurements, fatigueresistancewasmeasuredafteralltheMEPswererecorded but within25minof thePASprotocol, becauseLTP/LTDeffects havebeenshowntolastforaminimumof30min(Kumpulainen etal., 2012;Mrachacz-Kerstinget al.,2007;Stefanet al.,2000;

Woltersetal.,2003).

2.3. Recordings

For electromyographic (EMG) measurements, a pseu- domonopolar electrode placement protocol was used where onesurfaceelectrodeofapair(Unilect,Ag/AgCl,UnomedicalLtd., Redditch,UK)wasplacedontherightSOLandtheotherovera bonysurfaceofthetibia.Agroundelectrodewasplacedoverthe lateralmalleolus (Hoffman et al., 2009). The pseudomonopolar setupallowedMEPs ofhigher amplitudetoberecorded,which in turn alsodecreased the intensity of thestimulus neededto evokeadetectableMEP.Priortoelectrodeplacement,theskinwas shaved, abradedand cleaned withalcohol to reduce resistance below5k.EMGsignalswereamplified(100×),band-passfiltered (10–1000Hz)andsampledat5kHz(NeuralSystemsNL900Dand NL844,DigitimerLtd.,Hertfordshire,UK).EMGdataandreaction forcesfromthepedalwerecollectedwithacomputervia16-bit AD converter (CED power 1401, Cambridge Electronics Design Limited,UK)andstoredforlateranalysis.

2.4. Procedures

A rectangular current pulse with a duration of 1ms was deliveredto thecommon tibial nerve using a constant-current stimulator(DS7AH,DigitimerLtd.,Hertfordshire,UK)forthePAS protocolinadditiontoevokingH-reflexesandmaximalM-waves (Mmax).Acircularcathodewithapickupareaof77mm²(Unilect short-term ECG Electrodes, Ag/AgCl, Unomedical Ltd., UK) was placed over thetibial nerve on thepoplitealfossa and anoval shaped(5.08cm/10.16cm)anode(V-trodesneurostimulationelec- trodes,MattlerElectronicsCorp.,USA)wasplacedabovethepatella.

Motorthreshold(MT)wasdefinedastheminimalintensitythat induced a visually identifiable muscletwitch in SOL.To quan- tify reliable Mmax, supramaximal stimulus intensity was used, being 150% of the current needed to elicit maximal stimulus response.

Forthefatiguetestsubjectswereinstructedtoproducetheir maximalisometricplantarflexionforceandmaintainitfor15s (Fig.3),duringwhichtheforcedeclinedtowardtheend.Verbal encouragementwasgiventhroughoutthetrial.Attheendofthe fatiguingcontraction,neuraldeficitwasestimatedusingtheinter- polationtwitchtechnique(ITT);asupramaximal(M_maxintensity) doublepulsewith10msintervalwasdeliveredtothetibialnerve toquantify possible increment in force (superimposed twitch).

Fig.2. Theexperimentalprotocol.Timeline’supperpanelrepresentsthemainprotocolandlowerpaneladditionalproceduresforsub-groupsonly.TMS,transcranialmagnetic stimulation;MVC,maximalvoluntarycontraction;ISI,inter-stimulusinterval;PAS,pairedassociatedstimulation.

ImmediatelyfollowingthesustainedMVC,thesamesupramaxi- maldoublepulsewasdeliveredtotherelaxedmuscletoquantify restingtwitch.Neuraldeficitindicatesinsufficientcentraldriveto themotorneuronsandisattributabletocentralfatigue.Test–retest repeatabilitywasdeterminedbyrepeatingthefatigueresistance testpriortointerventionintensubjects.Therewasamandatory breakof10minbetweenthetestandtheretesttoreducethepossi- bilityofanyfatigueeffect.Ithasbeenshowninpreviousstudiesthat 10ssustainedmaximalforceisrecoveredwithin5min(Benwell etal.,2006).

TMSwasdeliveredusingamono-pulseMagstim2002stimula- torwitha9cmdoublebatwingcoil(Magstim,Whitland,UK).The optimalstimulussiteforSOLwasusuallylocated1cmlateraland 1cmposteriortothevertex.Acustom-madecoilholderandrubber strapswereusedtofixthecoilfirmlytothehead.Thepositionof thecoilwasmarkedonacloselyfittedcapwornbythesubjects.

Thecapalsoprotectedthehairandheadfromcoldspray(PRF101, Taerosol,Kangasala,Finland),whichwasusedtopreventthecoil fromoverheating.Restingmotorthreshold(RMT)wasdefinedas theloweststimulusintensityrequiredtoelicitaMEPwithapeak- to-peakamplitudeof50␮Vinthreeoutoffiveconsecutivetrials.

Stimulusintensitywasthensetto120%ofRMTandthisintensity wasusedthroughouttheexperiment.TMSwasdeliveredtentimes topassivemuscleandfivetimestoactivemuscleat20%ofMVC and50%ofMVC.Ithasbeenpreviouslydemonstrated thatMEP responsesincreaseprogressivelyfromresttoacontractionstrength of80%ofMVCintheSOL(Oyaetal.,2008).Astheforce–timecurve wasdisplayedonthescreeninfrontofthesubjects,theywereable toreachtheirtargetlevelwithanaccuracyof±2%duringtheactive conditions.

Fig.3. Examplefigureofthefatigueresistancetest.Forcetracesarefromonerep- resentativesubject.Twounitedthinlinesdemonstratethetest-retestrepeatability priortointervention(fatiguetest1=blackline,fatigueretest=grayline).Thethicker blacklinerepresentsthepost-interventiontest.Force(F)isnormalizedtomaximal voluntaryforce.

H-reflexesweremeasuredinsub-groupsofsixsubjectsfrom both PASLTP and PASLTD groups during rest. Mmax was elicited beforeallH-reflex measurements.Then10 submaximal stimuli wereappliedwithacurrentintensitythatevokedpeak-to-peak M-waveresponsesof20±5%ofMmaxtoquantifytheH-reflex.

PASconsistedofasingleelectricalstimulationdeliveredtothe tibialisnerveat150%ofMTandasingleTMSpulseat120%ofRMT.

ISIstoinduceLTP-andLTD-likeplasticitywereselectedbasedon previousexperiments.Aconstant ISIof 50mswasusedfor the PAS_LTPgroup(Kumpulainenetal.,2012;Mrachacz-Kerstingetal., 2007;Poonetal.,2008)andconstantISIof20mswasusedforthe PASLTDgroup(JayaramandStinear,2008;Poonetal.,2008;Stinear andHornby,2005).Atotalof200pairsofstimuliwereappliedat arateof0.2Hz.TooptimizethePASeffect,subjectswereaskedto produce5%ofMVCplantarflexionwiththeirrightlegduringthe PASprotocol(Mrachacz-Kerstingetal.,2007).

2.5. Dataanalysisandstatistics

Commerciallyavailablesoftware(Spike2,CED,Cambridge,Eng- land) was used for all offline analyses.Fatigue resistance was calculatedbydividingtheaverageforceduringthe15sisometric contractionbytheindividualMVC.Thusfatigueresistanceof100%

correspondstothetheoreticalsituationwheretheinitialMVCforce ismaintainedfortheentire15s.Amplitudesofthesuperimposed twitchandrestingtwitchweremeasuredandtheratiobetween themwascalculatedtoestimateneuraldeficit.Todeterminethe efficacyofPAS,peak-to-peakMEPamplitudesweremeasuredfrom SOLandaveraged.Spinal efficacywasdeterminedbytakingH- reflexandM_maxpeak-to-peakamplitudesandcalculatingtheratio ofthesetwomeasures(H/Mmax).Corticalsilentperiod(SP)was analyzedwhenMEPsweredeliveredduringcontractionsat50%of MVCasrecommendedbySäisänenetal.(2008)andalsoduring contractionsat20%ofMVC.SPreferstoasilencingofanyongo- ingEMGactivityaftertheMEPasaresultofTMSbeingdelivered duringavoluntarycontraction.ThedurationoftheSPwasdeter- minedby visualinspectionas thetime fromMEPoffset tothe time ofreoccurrenceofvoluntaryEMG activity.Eachindividual trialwasanalyzedseparatelyandthenaveragedacrosstrialsand subjects.

Normality of variables was tested with Shapiro–Wilk’s W- tests. Accordingly, SOL MEPs, neural deficit and H/Mmax were comparedusingWilcoxon’ssigned-ranktestseparatelyforboth groups.DifferencesbetweengroupsforPASeffect(postMEPsas percentageofbaselinevalues)weretestedwithMann–Whitney U-tests.MVC and fatigueforcewerecompared withtwo-tailed pairedt test.Fatigue resistancewascomparedwitha two-way repeatedmeasuresANOVAwithwithinfactortimeoftwolevels

LTD

inter-individualvariabilityofthePASeffects,whichhavebeenpre- viouslyreported(Ridding and Ziemann,2010).Thesignificance levelwassetatP<0.05.Alldataaregivenasmean±standarddevi- ation(SD).

3. Results

3.1. Changeinexcitability

Fig.4showsoriginalrestingMEPrecordingsforonerepresen- tativesubjectinthePASLTPandPASLTDgroupsbefore,immediately

trials.

afterand15minafterPAS.ThePASeffect(postMEPs/preMEPs) wassignificantlydifferentbetweengroupsatpost0andpost15at rest(P<0.05)butnotinactiveconditions(P>0.05).Thenormalized post-interventionpeak-to-peakMEPamplitudesarepresentedin Fig.5AandB.InthePAS_LTPgroup,MEPincreasedby51±108%at post0(P>0.05)andsignificantlyby73±123%(P<0.05)atpost15 atrest.Therewerenosignificantchangesinthe20%ofMVCcon- ditionbutMEPsdecreasedsignificantlyatbothpost0by9±12%

(P<0.05) and post15 by 8±14% (P<0.05) in the 50% of MVC

Fig.5. EffectofPASonMEPresponses.Meanpost-interventionMEPamplitudes(normalizedtobaseline)(A)inthePASLTPgroup(15subjects)and(B)inthePASLTDgroup (15subjects).Thesameresultsarepresentedalsoforrespondersonly(C)inthePASLTPgroup(11subjects)and(D)inthePASLTDgroup(12subjects),*P<0.05.

condition.In the PASLTD group, passive MEPdecreased signifi- cantly atpost0 by27±32% (P<0.05) but notat post15, where MEPdecreasedby20±38%(P>0.05).MEPdecreasedsignificantly by 15±25% (P<0.05) at post0 and by 9±18% (P<0.05) in the 20%ofMVCcondition.Therewerenosignificantchangesinthe 50% condition. The difference between the post0 and post15 wasneverstatisticallysignificant.SPsatcontractionlevelof50%

of MVCwere notaffected bytime (F_(2,56)=1.54, P>0.05)group (F_(1,28)=0.87,P>0.05)northeirinteraction(F_(2,56)=0.06,P>0.05).

SPvaluesforthePAS_LTPgroupwere75±27msatpre,78±23ms at post0 and 79±25ms at post15, and for the PASLTD group 68±19ms,72±15msand73±14ms,respectively.Correspond- ingly, there were no changes in SP duration at 20% of MVC;

time(F_(2,56)=1.83,P>0.05)group(F_(1,28)=1,66,P>0.05)northeir interaction(F_(2,56)=0.19,P>0.05).Thecoefficientofvariationwas significantly higher, 0.25 at 20% of MVC compared to 0.16 at 50%ofMVC(P<0.05).NeitherMmaxamplitudenorH/Mmaxratio changedsignificantlythroughouttheprotocol(P>0.05).Mmaxval- uesforthePAS_LTPgroupwere19±3mVatpre,18±3mVatpost0 and 18±3mV at post15, and for the PAS_LTD group 19±3mV, 19±3mVand19±3mV,respectively.H/MmaxratiosforthePASLTP

groupwere0.60±0.07atpre,0.58±0.12atpost0and0.60±0.09 atpost15,and forthePAS_LTDgroup0.49±0.24,0.47±0.22 and 0.51±0.24,respectively.Thesecondaryanalysiswithonlyrespon- dersisshowninFig.5CandB;MEPsincreasedsignificantlyatboth post0andpost15(P<0.05)inthePAS_LTPgroup(11subjects)and decreasedsignificantlyatbothpost0andpost15(P<0.05)inthe PASLTDgroup(12subjects).OtherwisethesignificancesinMEPand SPresultsdidnotdifferfromthewholegroupanalysis.

3.2. Fatigueresistance

Inboth groupsthe15sfatiguetest inducedsignificantforce reductionpriortoPAS;inthePAS_LTPgroupMVCwas1380±420N and average fatigue force was significantly less, 1230±340N (P<0.001). In the PASLTD group, MVC force was 1190±320N and average fatigue force was significantly less, 1110±280N (P<0.01). Correspondingfatigue resistancevalues were 90±8%

in thePASLTP group and 93±7% in thePASLTD group.Contrary to ourhypothesis, fatigue resistancewas not affected by time (F_(1,28)=0.11, P>0.05) group (F_(1,28)=0.59, P>0.05) nor their interaction (F_(1,28)=0.86, P>0.05). After the PAS interventions, fatigueresistancewas92±1%inthePASLTPgroupand92±7%in thePAS_LTDgroup.However,whenthegroupswerecombined,the changeinexcitability(post15MEPsasapercentageofbaselineval- ues)correlatedsignificantlywiththechangeinfatigueresistance (N=30,R=0.40,P<0.05),whichcanbeseeninFig.6.InthePAS_LTP group,neural deficitwas 1.6±2.4% prior toPAS and 0.9±4.4%

followingPAS.InthePASLTDgroup,neuraldeficitwas6.5±13.3%

priortoPASand5.9±11.2%followingPAS.Neuraldeficitdecreased non-significantly(P>0.05)by44±79%inPAS_LTPandby10±30%

in PASLTD. Test–retest comparison (N=10) revealed excellent repeatabilityofthefatigueresistancetest(P<0.001).Whenonly responders(PAS_LTPgroup=11subjectsandPAS_LTDgroup=12sub- jects)wereincludedinthesecondaryanalysis,two-wayrepeated measuresofANOVArevealedasignificantinteractionoftimeand group (F_(1,21)=4.8, P<0.05). Post hoc analysis using two-tailed paired t-test showed a non-significant improvement in fatigue resistanceby3.1±8.4%in thePASLTP groupand a reduction of 3.0±7.5%inthePAS_LTDgroup.Otherwisethesignificancesinforce resultsdidnotdifferfromthewholegroupanalysis.

4. Discussion

Theaimofthisstudywastoinvestigatethefunctionalrelevance ofPASbydeterminingtheeffectoftwodifferentPASintervention

Fig.6. Relationshipbetweenexcitabilityandfatigueresistance.Spearman’scoeffi- cientwasusedtocorrelatePAS-inducedchangesinexcitability,presentedonthe horizontalaxis,withchangesinfatigueresistance,presentedontheverticalaxis (N=30,R=0.40,P<0.05).

onfatigueresistanceduring15ssustainedmaximalisometriccon- traction.OnaveragefatigueresistancedidnotchangeafterPAS_LTP orPASLTD,howeverthechangesinMEPpeak-to-peakamplitudes afterPAScorrelatedwiththechangesinfatigueresistance.Subjects whose MEPsize increasedalso demonstrated improvedfatigue resistanceandviceversa.

4.1. Neuralcorrelates–changesincortico-spinalandspinal excitabilitiesafterPAS

Theresultsofthepresentstudyindicatethatassociativemod- ulation of excitability to the cortical projections of SOL was achievedafterPASLTPandPASLTDinterventionswithconstantISIs.

Accordingly,PAS_LTP inducedenhancementofsynaptictransmis- sionwhereasPASLTDinducedweakeningofsynaptictransmission.

TherewasnosignificantincreaseinMEPsizeatpost0afterPASLTP, butasignificantincreaseatpost15.Asimilartrendhasbeenseen inpreviousstudiesthathaveshownanincreasingPAS_LTPeffect over time (Kumpulainen et al., 2012; Prior and Stinear, 2006;

Stefanetal.,2000).On thecontrary,MEPsizedecreased signif- icantlyatpost0butnot atpost15afterPAS_LTD,whichis inline withapreviousstudybyDiLazzaroetal.(2011)whereasignif- icanteffectwasachievedimmediatelyafterbutnot30minafter PAS_LTD.MEPswerenotincreasedintheactiveconditionsfollow- ingPASLTP;onthecontrary,MEPsweresignificantlydecreasedin the50%ofMVCconditionafterPASLTP.Thisisaninterestingresult since,toourknowledge,PAS-inducedeffectshavenotbeenmea- suredatsuchhighcontractionlevels,and asignificantdecrease inMEPsizeafterPASLTPhasnotbeenreportedbefore.However, therearestudiesshowingnochangeinMEPsduringslightcon- tractions(Kumpulainenetal.,2012;Stefanetal.,2000,2004).The interpretationoftheeffectofPASonactivemusclesismorecom- plicated ascan beseen in the studyof Lu et al.(2009) where a decrease of the movement-relatedcorticalpotentials (MRCP) inelectroencephalographyrecordingswasreportedafteranLTP- likePAS-protocol.Accordingly, performingthesamemovement pattern, simple thumb abduction, generated a decreased MRCP negativity after the PAS intervention. MRCP reflects executive aspectsoftheforthcomingmotoractionanddecreasedMRCPneg- ativityindicatesweakervolitionalmotoroutput.Thismayindicate that the LTP-like effect decreasesthe level of effort needed to

(Kumpulainenetal.,2012;Stefanetal.,2000;Woltersetal.,2003).

4.2. Functionalimplications–changesinfatigueresistanceafter PAS

The15ssustainedmaximalisometriccontractioninducedsig- nificantforcelossonalloccasionsbutfatigueresistancedidnot changeafterinterventions.Thisisinlinewitha previousstudy byMilanovi ´cetal.(2011)whichshowednochangesinafatigue testinvolvingsustainedisometriccontractionat50%ofMVCafter PASLTP.However,asignificantrelationshipwasfoundbetweenthe changein excitability of thecorticalprojections and change in fatigueresistanceinthepresentstudy,whichwasnotreported inthestudybyMilanovi ´cetal. (2011).Onepossiblereason for thisdifferencecouldbethattheyusedthedurationofsubmaxi- malcontractionasanindicationoffatigue,whichmayinvolvea greaterperipheralcomponentoffatiguecomparedtothepresent protocol(LentzandNielsen,2002).Ithasbeendemonstratedthat endurance time and changes in maximal capacity to generate forceprovideinformationaboutdifferentprocessesinducedbythe exercise(Vøllestad,1997).Thus,endurancetimemaynotdirectly correlatewithmotorcortexexcitability.Benwelletal.(2006)found asignificantreductionintherateofforcelossduringa10-sMVC ofhandmusclesafterincreasingcorticomotorexcitabilityusinga spiketiming-dependentrepetitiveTMSintervention.Inthecurrent studythelackofchangesinfatigueresistanceatthegrouplevel maybeduetoconsiderableinter-individualvariabilityinthePAS- inducedplasticitychangeswithinthegroups.Thechosenconstant ISIsmightbeonereasonforrelativelyhighinter-individualvari- ability(SD=±108%and123%)observedinthisstudy.Nevertheless, thepresentvaluesarecomparabletothevariability(SD=±105%

and52%)andnumberofresponders(12respondersfrom16sub- jects)obtained in a previousstudy(Kumpulainen et al., 2012), where the ISI wasoptimized to the individualsomatosensory- evokedpotentialandwas50±2msonaverage.Sinceanearlier studyby Mrachacz-Kersting et al. (2007) foundsignificant LTP duringa10mstimewindowwithconstantISIsof45–55ms,the constant ISI of 50ms used here shouldbe sufficient toinduce LTPin most subjects.The optimalISI to induceLTD in theleg areahas not been studied but an ISI of 20ms is mostly used.

StinearandHornby(2005)reportedthenumberofLTDrespon- ders and it was quite similar to ours; 12 responders from 14 subjects.Alsopreviousstudieshavereportedwidevariabilityof PAS-inducedeffectsbetweensubjects,whichhasbeenassociated withdifferencesinbrainanatomy,genes,andtrainingbackground (RiddingandZiemann,2010).Becauseofthedissimilaritiesinbrain anatomy,TMScanpreferentiallyactivatedifferentneuronalcircuits indifferentsubjects(Sakaietal.,1997)causingvariabilityinthe PAS-inducedeffects.Thegeneticpolymorphismsofneurotrophins caninfluencethe inductionof plasticity(Ridding and Ziemann, 2010).However,thereisonlyonebrain-derivedneurotrophicfac- tor–polymorphism,whichhasbeenshowntolimitPAS-induced motorcortexplasticity(Cheeranetal.,2008).Furthermore,ithas

haveoccurredwithinthevariationofITT.Lowsensitivityofthe ITTmethodhasbeenreportedatmaximalcontractionintensities (Taylor,2009).

4.3. CandidatemechanismsbehindPASandfatigue

IthasbeenproposedthatPAS-inducedLTPoccurswithinthe motorcortex(DiLazzaroetal.,2009;Stefanetal.,2000),wherethe somatosensorystimuluscanarriveviathedorsalcolumn-medial leminiscalrouteorviaalongerpathwayinvolvingthecerebellum (Strigaroetal.,2014).Intracorticalneuralcircuitsthroughwhich theeffectsofPASandfatigueemergehavebeeninvestigatedwith singleandpairedpulseTMSinseveralpreviousstudies(Carson andKennedy,2013;Gruetetal.,2013;TaylorandGandevia,2001).

PairedpulseTMSisatooltoexaminefacilitatoryandinhibitory circuitsinthecerebralcortex.However,asdifferentstudieshave useddifferent interventionsand targetmuscles, theresultsare difficulttointerpret.ProlongationofSPhasbeenobservedaftervar- iousfatiguingexercisesincludingsustainedMVCofsoleusmuscle (IguchiandShields,2012;McKayetal.,1996).Sinceprolongation ofSPislessaftercervicomedullarystimulation-inducedMEP,addi- tionalinhibitionatthecorticallevelhasbeensuggested(Levenez etal., 2008;Taylor etal., 1996).Inaddition, Hiltyet al.(2011) showedthatcentralprojectionsofgroupIII–IVmuscleafferents mayfacilitateafatigue-inducedincreaseinSP.Therefore,itseems thatGABA^b-mediatedintracorticalinhibitorycircuitshavearole inthedevelopmentofcentralfatigue.SinceSPdurationremained unchanged afterPAS interventionsin this study, it seemsthat, atleastpartly,fatigueandPASaffectdifferentcorticalinterneu- rons.Thiscorrespondstothecurrentresultsshowingnosignificant effectofPASonfatigueresistance.However,amongPASresponders thePAS-inducedLTPmighthavecompensatedforfatigue-induced inhibitioninthemotorcortexandthusaffectedfatigueresistance.

Inaddition,PASinducedLTP/LTD-likeplasticityhasbeenshownto beaccompaniedwithotherneuronalcircuitsinthecerebralcor- tex(CarsonandKennedy, 2013), whichmight havecontributed totheobservedsignificantcorrelation.Forexample,longafferent inhibition(LAI),whichisalsoGABA^b-mediated,isdecreasedfol- lowingPASLTP(Meunieretal.,2012;Russmannetal.,2009)and mightincreasemotordrivetotheexercisingmuscleduringfatigu- ingtasks.LAIreflectstheactivityofsomatosensoryinputs,andis obtainedwhentheintervalbetweenperipheralafferentstimula- tionandsubsequentTMSisintheregionof200ms.

5. Conclusions

Inthepresentstudy,PASinducedassociativeplasticitychanges inthecorticalprojectionstotherestingSOL.Onaverage,fatigue resistanceduringa15-ssustainedmaximalisometriccontraction didnotchangefollowingPASinterventions.Thus,functionalityof PASinterventionswasnotevidentwiththecurrentexperimen- taldesign.However,amongrespondersfatigueresistanceshowed

significant interaction of time and group and PAS-induced excitabilitychangescorrelatedsignificantlywithchangesinfatigue resistance. Thissuggests that PAS might have slightly affected centralfatigueduringshortmaximalcontractions.Therefore,PAS mighthaveimplicationsforimprovingperformanceinrehabilita- tionsettings.

Acknowledgments

Theauthorsareverygratefultothesubjectswhoparticipated inthestudyandtothelaboratorystafffromtheNeuromuscular ResearchCenter(DepartmentofBiologyofPhysicalActivity)ofthe UniversityofJyväskylä,Finland,fortheirvaluablecontributionsto thisproject.

References

Benwell,N.M.,Mastaglia,F.L.,Thickbroom,G.W.,2006.Paired-pulserTMSattrans- synapticintervalsincreasescorticomotorexcitabilityandreducestherateof forcelossduringafatiguingexerciseofthehand.Exp.BrainRes.175,626–632.

Carson,R.G.,Kennedy,N.C.,2013.Modulationofhumancorticospinalexcitabilityby pairedassociativestimulation.Front.Hum.Neurosci.7.

Cheeran,B.,Talelli,P.,Mori,F.,Koch,G.,Suppa,A.,Edwards,M.,Houlden,H.,Bhatia, K.,Greenwood,R.,Rothwell,J.C.,2008.Acommonpolymorphisminthebrain- derivedneurotrophicfactorgene(BDNF)modulateshumancorticalplasticity andtheresponsetorTMS.J.Physiol.586,5717–5725.

DiLazzaro,V.,Dileone,M.,Pilato,F.,Profice,P.,Oliviero,A.,Mazzone,P.,Insola,A., Capone,F.,Ranieri,F.,Tonali,P.A.,2009.Associativemotorcortexplasticity:

directevidenceinhumans.Cereb.Cortex19,2326–2330.

DiLazzaro,V.,Dileone,M.,Pilato,F.,Capone,F.,Musumeci,G.,Ranieri,F.,Ricci,V., Bria,P.,DiIorio,R.,deWaure,C.,Pasqualetti,P.,Profice,P.,2011.Modulation ofmotorcortexneuronalnetworksbyrTMS:comparisonoflocalandremote effectsofsixdifferentprotocolsofstimulation.J.Neurophysiol.105,2150–2156.

Frantseva,M.V.,Fitzgerald,P.B.,Chen,R.,Moller,B.,Daigle,M.,Daskalakis,Z.J.,2008.

Evidenceforimpairedlong-termpotentiationinschizophreniaanditsrelation- shiptomotorskilllearning.Cereb.Cortex18,990–996.

Gandevia,S.C.,2001.Spinalandsupraspinalfactorsinhumanmusclefatigue.Phys- iol.Rev.81,1725–1789.

Gandevia,S.C.,Allen,G.M.,Butler,J.E.,Taylor,J.L.,1996.Supraspinalfactorsinhuman musclefatigue:evidenceforsuboptimaloutputfromthemotorcortex.J.Physiol.

490(Pt2),529–536.

Gruet,M.,Temesi,J.,Rupp,T.,Levy,P.,Millet,G.,Verges,S.,2013.Stimulationofthe motorcortexandcorticospinaltracttoassesshumanmusclefatigue.Neuro- science231,384–399.

Hilty,L.,Lutz,K.,Maurer,K.,Rodenkirch,T.,Spengler,C.M.,Boutellier,U.,Jäncke,L., Amann,M.,2011.Spinalopioidreceptor-sensitivemuscleafferentscontribute tothefatigue-inducedincreaseinintracorticalinhibitioninhealthyhumans.

Exp.Physiol.96,505–517.

Hoffman,B.W.,Oya,T.,Carroll,T.J.,Cresswell,A.G.,2009.Increasesincorticospinal responsivenessduringasustainedsubmaximalplantarflexion.J.Appl.Physiol.

(1985)107,112–120.

Hunter,S.K.,Todd,G.,Butler,J.E.,Gandevia,S.C.,Taylor,J.L.,2008.Recoveryfrom supraspinalfatigueisslowedinoldadultsafterfatiguingmaximalisometric contractions.J.Appl.Physiol.(1985)105,1199–1209.

Hunter,S.K.,Butler,J.E.,Todd,G.,Gandevia,S.C.,Taylor,J.L.,2006.Supraspinalfatigue doesnotexplainthesexdifferenceinmusclefatigueofmaximalcontractions.

J.Appl.Physiol.(1985)101,1036–1044.

Iguchi,M.,Shields,R.K.,2012.Corticalandsegmentalexcitabilityduringfatiguing contractionsofthesoleusmuscleinhumans.Clin.Neurophysiol.123,335–343.

Ishikawa,M.,Komi,P.V.,Grey,M.J.,Lepola,V.,Bruggemann,G.P.,2005.Muscle- tendoninteractionandelasticenergyusageinhumanwalking.J.Appl.Physiol.

(1985)99,603–608.

Jayaram,G.,Stinear,J.,2008.Contralesionalpairedassociativestimulationincreases pareticlowerlimbmotorexcitabilitypost-stroke.Exp.BrainRes.185,563–570.

Jung,P.,Ziemann,U.,2009.Homeostaticandnonhomeostaticmodulationoflearning inhumanmotorcortex.J.Neurosci.29,5597–5604.

Kumpulainen,S.,Mrachacz-Kersting,N.,Peltonen,J.,Voigt,M.,Avela,J.,2012.The optimalinterstimulusintervalandrepeatabilityofpairedassociativestimula- tionwhenthesoleusmuscleistargeted.Exp.BrainRes.221,241–249.

Kumpulainen,S.,Avela,J.,Gruber,M.,Bergmann,J.,Voigt,M.,Linnamo,V.,Mrachacz- Kersting,N.,2014.Differentialmodulationofmotorcortexplasticityinskill-and endurance-trainedathletes.Eur.J.Appl.Physiol.,1–9.

Lentz,M.,Nielsen,J.F.,2002.Post-exercisefacilitationanddepressionofMwave andmotor evokedpotentialsin healthysubjects.Clin.Neurophysiol. 113, 1092–1098.

Levenez,M.,Garland,S.J.,Klass,M.,Duchateau,J.,2008.Corticalandspinalmodu- lationofantagonistcoactivationduringasubmaximalfatiguingcontractionin humans.J.Neurophysiol.99,554–563.

Lu,M.,Bliem,B.,Jung,P.,Arai,N.,Tsai,C.,Ziemann,U.,2009.Modulationofprepara- toryvolitionalmotorcorticalactivitybypairedassociativetranscranialmagnetic stimulation.Hum.BrainMapp.30,3645–3656.

McKay,W.B., Stokic,D.S.,Sherwood,A.M.,Vrbova,G.,Dimitrijevic,M.R.,1996.

Effectoffatiguingmaximalvoluntarycontractiononexcitatoryandinhibitory responseselicitedbytranscranialmagneticmotorcortexstimulation.Muscle Nerve19,1017–1024.

Meunier,S.,Russmann,H.,Shamim,E.,Lamy,J.,Hallett,M.,2012.Plasticityofcortical inhibitionindystoniaisimpairedaftermotorlearningandpaired-associative stimulation.Eur.J.Neurosci.35,975–986.

Milanovi ´c,S.,Filipovi ´c,S.,Blesi ´c,S.,Ili ´c,T.,Dhanasekaran,S.,Ljubisavljevi ´c,M.,2011.

Paired-associativestimulationcanmodulatemusclefatigueinducedmotorcor- texexcitabilitychanges.Behav.BrainRes.223,30–35.

Mrachacz-Kersting, N.,Fong, M.,Murphy, B.A.,Sinkjaer, T.,2007. Changes in excitabilityofthecorticalprojectionstothehumantibialisanteriorafterpaired associativestimulation.J.Neurophysiol.97,1951–1958.

Oya,T.,Hoffman,B.W.,Cresswell,A.G.,2008.Corticospinal-evokedresponsesin lowerlimbmusclesduringvoluntarycontractionsatvaryingstrengths.J.Appl.

Physiol.(1985)105,1527–1532.

Poon,D.,Roy,F.,Gorassini,M.,Stein,R.,2008.Interactionofpairedcorticaland peripheralnervestimulationonhumanmotorneurons.Exp.BrainRes.188, 13–21.

Prior,M.M.,Stinear,J.W.,2006.Phasicspike-timing-dependentplasticityofhuman motorcortexduringwalking.BrainRes.1110,150–158.

Rajji,T.K.,Liu,S.,Frantseva,M.V.,Mulsant,B.H.,Thoma,J.,Chen,R.,Fitzgerald,P.B., Daskalakis,Z.J.,2011.Exploringtheeffectofinducinglong-termpotentiationin thehumanmotorcortexonmotorlearning.BrainStimul.4,137–144.

Ridding,M.C.,Ziemann,U.,2010.Determinantsoftheinductionofcorticalplas- ticitybynon-invasivebrainstimulationinhealthysubjects.J.Physiol.588, 2291–2304.

Russmann,H.,Lamy,J.,Shamim,E.A.,Meunier,S.,Hallett,M.,2009.Associative plasticityinintracorticalinhibitorycircuitsinhumanmotorcortex.Clin.Neu- rophysiol.120,1204–1212.

Säisänen,L.,Pirinen,E.,Teitti,S.,Könönen,M.,Julkunen,P.,Määttä,S.,Karhu,J.,2008.

Factorsinfluencingcorticalsilentperiod:optimizedstimuluslocation,intensity andmusclecontraction.J.Neurosci.Methods169,231–238.

Sakai,K.,Ugawa,Y.,Terao,Y.,Hanajima,R.,Furubayashi,T.,Kanazawa,I.,1997.Pref- erentialactivationofdifferentIwavesbytranscranialmagneticstimulationwith afigure-of-eight-shapedcoil.Exp.BrainRes.113,24–32.

Stefan,K.,Wycislo,M.,Classen,J.,2004.Modulationofassociativehumanmotor corticalplasticitybyattention.J.Neurophysiol.92,66–72.

Stefan,K.,Wycislo,M.,Gentner,R.,Schramm,A.,Naumann,M.,Reiners,K.,Classen, J.,2006.Temporaryocclusionofassociativemotorcorticalplasticitybyprior dynamicmotortraining.Cereb.Cortex16,376–385.

Stefan,K.,Kunesch,E.,Cohen,L.G.,Benecke,R.,Classen,J.,2000.Inductionofplas- ticityinthehumanmotorcortexbypairedassociativestimulation.Brain123, 572–584.

Stinear,J.W.,Hornby,T.G.,2005.Stimulation-inducedchangesinlowerlimbcor- ticomotorexcitability duringtreadmillwalkinginhumans.J.Physiol. 567, 701–711.

Strigaro,G.,Hamada,M.,Murase,N.,Cantello,R.,Rothwell,J.C.,2014.Interaction betweendifferentinterneuronnetworksinvolvedinhumanassociativeplastic- ity.BrainStimul.7,658–664.

Szubski,C.,Burtscher,M.,Loscher,W.N.,2007.Neuromuscularfatigueduringsus- tainedcontractionsperformedinshort-termhypoxia.Med.Sci.SportsExerc.39, 948–954.

Taylor,J.L.,Gandevia,S.C.,2001.Transcranialmagneticstimulationandhumanmus- clefatigue.MuscleNerve24,18–29.

Taylor,J.L.,Butler,J.E.,Gandevia,S.,1999.Alteredresponsesofhumanelbowflex- orstoperipheral-nerveandcorticalstimulationduringasustainedmaximal voluntarycontraction.Exp.BrainRes.127,108–115.

Taylor,J.L.,2009.Point:theinterpolatedtwitchdoes/doesnotprovideavalidmea- sureofthevoluntaryactivationofmuscle.J.Appl.Physiol.(1985)107,354–355.

Taylor,J.L.,Butler,J.E.,Allen,G.M.,Gandevia,S.C.,1996.Changesinmotorcortical excitabilityduringhumanmusclefatigue.J.Physiol.490(Pt2),519–528.

Vøllestad,N.K.,1997.Measurementofhumanmusclefatigue.J.Neurosci.Methods 74,219–227.

Wolters,A.,Sandbrink,F.,Schlottmann,A.,Kunesch,E.,Stefan,K.,Cohen,L.G., Benecke,R.,Classen,J.,2003.AtemporallyasymmetricHebbianrulegoverning plasticityinthehumanmotorcortex.J.Neurophysiol.89,2339–2345.

Ziemann,U.,Ilic,T.V.,Pauli,C.,Meintzschel,F.,Ruge,D.,2004.Learningmodifies subsequentinductionoflong-termpotentiation-likeandlong-termdepression- likeplasticityinhumanmotorcortex.J.Neurosci.24,1666–1672.

Ziemann,U.,2004.TMSanddrugs.Clin.Neurophysiol.115,1717–1729.

Ziemann,U.,Paulus,W.,Nitsche,M.A.,Pascual-Leone,A.,Byblow,W.D.,Berardelli, A.,Siebner,H.R.,Classen,J.,Cohen,L.G.,Rothwell,J.C.,2008.Consensus:motor cortexplasticityprotocols.BrainStimul.1,164–182.