Interaction of support surface stability and Achilles tendon vibration

14 Pages·2013·0.47 MB·English

by  Marius Dettmer

HumanMovementScience32(2013)214–227
ContentslistsavailableatSciVerseScienceDirect
Human Movement Science
journal homepage: www.elsevier.com/locate/humov
Interaction of support surface stability and
Achilles tendon vibration during a postural
adaptation task
⇑
Marius Dettmer , Amir Pourmoghaddam, Daniel P. O’Connor,
Charles S. Layne
DepartmentofHealthandHumanPerformance,UniversityofHouston,104CGAR,3855HolmanSt.,Houston,TX77204,USA
CenterforNeuromotorandBiomechanicsResearch,JohnP.McGovernCampus,2450HolcombeBoulevard,Houston,TX
77021-2040,USA
a r t i c l e i n f o a b s t r a c t
Articlehistory: Orchestration of sensory-motor information and adaptation to
Availableonline5March2013 internalorexternal,acuteorchronicchangesisoneofthefunda-
mental features of human postural control. The postural control
PsycINFOclassification: systemischallengedonadailybasis,anddisplays aremarkable
2221 abilitytoadapttobothlongandshorttermchallenges.Toexplore
theinteractionbetweensupportsurfacestabilityandAchillesten-
Keywords:
Learning donvibrationduringaperiodofadaptationweusedbothalinear
Sensoryadaptation measureandanon-linearmeasurederivedfromcenter-of-pressure
Proprioception (COP)data.Anequilibriumscore(ES),baseduponpeakamplitude
of anterior-posterior sway towards theoretical limits of stability
wasthelinearmeasureusedtoassessposturalperformance.We
observedearlyeffectsofvibrationonposturalstability,depending
onsupportcharacteristics.Participantswereabletodecreasesway
withextendedpracticeoverdays,independentofsupportsurface
stability.ApproximateentropyanalysisofCOPdataprovidedaddi-
tionalinformationaboutcontroladaptationprocesses.
(cid:2)2013ElsevierB.V.Allrightsreserved.
1.Introduction
Maintainingposturalstabilitywhetherforstanding,locomotionorothertasks,requirescomplex
interactionsbetweenneuromotorprocesses,biomechanicalconstraintsandthegoalsoftheindividual.
⇑
Correspondingauthor.Address:3855HolmanSt.,104CGAR,Houston,TX77204,USA.Tel.:[phone];fax:+1(713)
7439860.
E-mailaddress:[email protected](M.Dettmer).
0167-9457/$-seefrontmatter(cid:2)2013ElsevierB.V.Allrightsreserved.
http://dx.doi.org/10.1016/j.humov.2012.12.002
M.Dettmeretal./HumanMovementScience32(2013)214–227 215
The central nervous system (CNS) constantly processes information regarding position and torques
providedbythreeprimarysensorysystems,thevestibular,visualandsomatosensory(Horak,Nashner,
&Diener,1990;Lackner&DiZio,2005).Informationfromthesesystemsresultsinthedetectionofpo-
sitionandmovementandtheinitiationofadequatemuscularresponsestomaintainbipedalpostural
control.
Ithasbeendemonstratedthatwithpractice,theCNSexhibitstheabilitytoadapttoexternalper-
turbations and thereby to decrease postural sway (Fransson et al., 2003). This adaptive process is
potentiallyduetoseveral,distinctshort-andlongertermmechanismsofadaptation:oneproposed
mechanismisdynamic,sensoryreweightingofavailablesensoryinformationsuchthattheCNSuti-
lizes a weighted sum of available sensory information to maintain postural control. If one sensory
channelprovidesunreliableinformation,itisdownweightedinfavorofmorereliablechannels(Car-
ver,Kiemel,&Jeka,2006;Haran&Keshner,2008).Neurophysiologicalmechanismsunderlyingthis
theory are not yet fully understood, but several theories and models for sensory integration and
reweightinghavebeenpresented(Maurer,Mergner,&Peterka,2006,2002,2003;Peterka&Loughlin,
2004).Asecondelementofadaptationinvolvesthemodificationofposturalstrategies.Thiscanin-
cludeshiftstowardsahiporanklestrategy,‘‘leaning’’andassociatedchangesofbiomechanicalcon-
straints, or general stiffening of lower leg joints and increased ankle musculature co-contraction
(Almeida,Carvalho,&Talis,2006;Benjuya,Melzer,&Kaplanski,2004;Brumagne,Janssens,Knapen,
Claeys,&Suuden-Johanson,2008;Termozetal.,2008).Additionaladaptationhasbeenproposedto
beencompassedbyaprocesslabeledlong-termconsolidation,emergingfromnumerousrepetitions
of the task, representing an enduringmotor learning effect that integratesthe aforementioned fea-
tures of adaptation for more efficientperformance (Fransson, Johansson,Tjernström, & Magnusson,
2003;Tjernström,Bagher,Fransson,&Magnusson,2010;Tjernström,Fransson,&Magnusson,2005).
Given the important role of sensory contributions to action, researchers are often interested in
challengingtheposturalcontrolsystembymodifyingspecificinputofvarioussensorychannels.Mus-
cletendonvibration(mostcommonattheAchillestendon),isatechniquethatallowsresearchersto
effectively challenge the postural control system by modifying sensory inputs primarily emerging
from ankle muscle spindles in healthy or pathologic populations (Fransson, Johansson,Hafström, &
Magnusson,2000;Gurfinkel’,Kireeva,&Lebik,1996;Rolletal.,1993;Thompson,Bélanger,&Fung,
2007, 2011). During vibration, Ia-afferents of the muscle spindles are preferentially stimulated (Ri-
bot-Ciscar,Rossi-Durand,&Roll,1998).Thismusclespindlestimulationcreatesproprioceptivemisin-
formation about the current muscle length and stretch (Matthews, 1986; Polónyová & Hlavacka,
2001), thereby generating sensations of muscle lengthening (Goodwin, McCloskey, & Matthews,
1972).Acommonlydescribedeffectisabackwardsleaningforcompensationoftheperceived(false)
musclelengtheningandassociatedposturalsway.Thisimmediateeffectonbodyorientationincreases
bodyswayandCOPfluctuations(Barbierietal.,2008;Ceyteetal.,2007;Lackner,1988;Polónyová&
Hlavacka,2001;Roll,Vedel,&Roll,1989;Thompsonetal.,2007)especiallywhencombinedwithoc-
cludedvision.Achillestendonvibrationcreatesadiscrepancybetweenperceivedafferentinformation
fromstretchreceptorsofthelowerlimbmusculature,andothermorereliablesources(vestibularsys-
tem,jointorskinreceptors,etc.),whichultimatelyaffectsposturalstability.
Themagnitudeofinfluenceofvibrationonposturalstabilityishighlycontext-dependent.Forin-
stance,posturalswayinresponsetomuscletendonvibrationwhilestandinguprightonanunstable
supportsurface isreducedwhencomparedtothevibration-induced swaywhile standingona sta-
blesupportsurface(Ivanenko,Solopova,&Levik,2000;Ivanenko,Talis,&Kazennikov,1999).Ithas
been suggested that if vibration is introduced while standing on an unstable support surface, the
CNS is able to partially suppress or downweight proprioceptive sensory input that is of no or little
informational value or is detrimental to postural stability (Hatzitaki, Pavlou, & Bronstein, 2004).
However, the possible interaction between vibration and support stability could affect the time
course of postural adaptation. Numerous investigations have combined different support surface
characteristics, manipulatingsensory conditionsand investigatingrepetitionof tasks to assesspos-
tural adaptation.
Linearanalysishastraditionallydominatedresearchandclinicalassessments.However,analyses
basedonnon-lineartechniquesarenowmorefrequentlybeingusedtoinvestigateposturalcontrol
dynamics. Among the non-linear tools applied to postural control research is approximate entropy
216 M.Dettmeretal./HumanMovementScience32(2013)214–227
(ApEn), a regularity statistic determining the predictability/regularity of variation in a time-series
(Pincus,1991;Pincus&Goldberger,1994).ApEnhasbeenshowntodetectchangesinmotorcontrol
thatmightnotberevealedbylinearanalysis,andresultshavebeeninterpretedaspotentialmodifi-
cationsofmotorcontrolorbiomechanicalconstraints(Cavanaughetal.,2006).Recentdatahashigh-
lighted the role of ApEn as a potentially useful supplementary analysis technique to traditional
outcomemeasuresofposturalstability(Cavanaugh,Guskiewicz,&Stergiou,2005;Cavanaugh,Mercer,
& Stergiou, 2007; Dusing, Kyvelidou, Mercer, & Stergiou, 2009; Hong, Manor, & Li, 2007; Stergiou,
2004;Turnock&Layne,2010).
Thecurrentstudywasdesignedtoevaluateshort-orlongertermadaptationinresponsetodiffer-
ent support surface characteristics and Achilles tendon vibration using both linear and non-linear
measurements.ItwashypothesizedthatinitialapplicationofvibrationwouldincreaseCOPmotion,
withmoresevereeffectsinthefixed-supportcondition;repeatedpracticewithvibrationwouldde-
creaseCOPmotioninbothsupportsurfaceconditions,bothintheshort-termandinthelongerterm.
InitialvibrationwouldaffectApEn,butthiseffectwouldbeinfluencedbysupportsurfacecondition.
Repeatedpracticewithvibrationwould affectApEnin bothsupportsurfaceconditions,both inthe
short-termandinthelongerterm.
2.Methodsandmaterials
2.1.Subjects
TheprotocolwasapprovedbytheCommitteefortheProtectionofHumanSubjects(CPHS)atthe
UniversityofHouston.ParticipantsforthestudywererecruitedfromtheUniversityofHoustoncam-
pus.NeuromuscularhealthwasexaminedbyhavingprospectiveparticipantsfilloutaPhysicalActiv-
ityReadinessQuestionnaireandperformapartialSensoryOrganizationTest(SOT),intherespective
conditiontheyhadbeenrandomlyassignedto(either2or5).Exclusioncriteriawerereportedpoten-
tialneuromotordeficitsorwhentheSOTshowedanabnormalscore.Thecutoffscoreforexclusion
was determined by normative data provided with the NeuroCom system (NeuroCom International,
Clackamas,OR).Foranagegroupof20–59years,thecutoffscoreswere85(SOT2)and52(SOT5).
Allrecruitedparticipantswereincludedinthestudy.Eighteen(n=18)young,healthyindividualsbe-
tween ages 18 to 35 (10 females; 8 males, mean age 24.1±4.2yrs; mean height 168.3±10.8cm;
M±SD)participatedinthestudy.
2.2.Design
AllexperimentalsessionswereconductedontheNeuroComBalanceManager(NeuroComInterna-
tional,Clackamas,OR)intheCenterforNeuromotorandBiomechanicsResearch(CNBR).Allpartici-
pants were randomly assigned to one of two experimental conditions (condition 2 and 5) of the
Sensory-Organization-Test(SOT).SOT2requiresofparticipantstoclosetheireyesandmaintainup-
right stance on a fixed-support surface (FS). SOT 5 requires of participants to close their eyes and
maintainuprightstancewhilethesupportsurfaceisunstable,andsway-referenced(SRS).InSOT5,
thesurfacerotatesinanteriorandposteriordirectionaccordingtoforcesappliedtotheplatformby
thesubjectwhiletryingtomaintainquietstance.Thisconditionisdesignedtodampenproprioceptive
signals,whereasmusclespindleinformationnolongerprovidesreliableinformationaboutbodyver-
ticality. This is combined with removal of visual reference,creating a challenging postural task. All
testingwascompletedinstockingfeetwitharmscrossedinfrontofthechest.Duringalltestsessions,
COPdatawerecollectedat100Hz.
VibrationwasappliedbilaterallyontheAchillestendons(seeFig.1)atafrequencyof70Hz(VB
115,Techno-Concept,Cereste,France).70Hziswithinarangeoffrequenciesshowntobeeffective
instimulatingmusclespindlesaroundtheankles(Polónyová&Hlavacka,2001)andevokepostural
responses.
Priortothefirstsession,informedconsentandanthropometricdata(age,genderandheight)were
obtainedfromeachparticipant.Participantswerefamiliarizedwiththesensationofmuscletendon
M.Dettmeretal./HumanMovementScience32(2013)214–227 217
Fig.1. Experimentalsetupforthecurrentstudy.Forceplatform(NeuroComBalanceManager)andIa-afferentstimulation
appliedvialocalmuscletendonvibration(TechnoConceptVB115).
vibration(whilesittinginachair)andperformedthreeacclimatizationtrialsoftheirassignedsupport
condition.Duringtheseinitialtests,themusclevibratorswereattachedattheparticipants’anklesbut
werenotactivated.Inthenextsession(followingday),threetrialsofbaselinedatawerecollected(no
vibration).Immediatelyafterbaselinetesting,thefirstvibrationtestsessionwasconducted.Eachtest
sessionconsistedofninetrialsof20secondseach,withfive-secondbreaksinbetweeneachsingletrial
andaoneminuterestperiodaftereverythreetrials.Participantsweretestedintheirrespectivegroup
supportsurfacecondition.Anauditorystimulussignalledthestartofthetest.Vibrationwasalways
activatedsimultaneouslywiththestartofatrialandwasturnedoffbetweentrials.Testsessionsof
ninetrialsperdaywereconductedforthreeconsecutivedaysforallparticipants.
2.3.Datareduction
Anequilibriumscore(ES)foreverysingletrialwasobtainedusingEquiTest8.0software(Neuro-
ComInternational),whichisbasedonestimatedcenter-of-gravityvariationobtainedfromanterior-
posteriorCOPvalues.Thecenter-of-gravityisestimatedbyusingasecond-orderButterworthdigital
filter (cut-off frequency of 0.85Hz) on COP data in each time-series (Leitner et al., 2009). ES is ex-
pressedbytheratioofthepeak-to-peakcenter-of-gravity(anterior-posteriordirection,COGy)angular
displacement to the theoretical maximum angular displacement limits of stability (approximately
12.5(cid:3)inanterior-posteriordirection).ESrangesbetween0–100,whereasascoreof0denotesafall
and 100 denotes perfect stability (NeuroCom, 1991). This method is the basis of computerized dy-
namicposturography(CDP)andhasbeenusedinavarietyofclinicalandscientificenvironmentsin
ordertomakeconclusionsaboutposturalstabilityofhealthysubjectsandpatients(Roberts-Warrior
218 M.Dettmeretal./HumanMovementScience32(2013)214–227
etal.,2000;Yardley,Burgneay,Nazareth,&Luxon,1998).Sincethisinvestigationaimedatassessing
earlyeffects,short-orlongertermadaptation,weuseddifferentapproachesfordatareduction.
Toanswerthequestionofwhethersupportsurfaceconditionaffectedresponseduringtheinitial
exposuretovibration,whichwaslabelledasearlyeffects,wecomparedtheESandApEnbaselinedata
withESandApEndatafromtherespectiveinitialvibrationtrial.Themethodologyemployed(i.e.,data
collectionbeginningsimultaneouslywithvibrationinitiation)resultedinthewelldocumentedimme-
diatebackwardsswayinresponsetovibrationandsubsequentdriftbacktowardsbaselinebeingin-
cludedinthecalculationofboththeESand ApEnvalues.Wewerealsointerested inexploringthe
effectsofsupportsurfaceonthevibrationresponseoverasomewhatlongertimeperiod.Totestfor
short and long term adaptation, the ES and ApEn data for each trial were centered relative to the
respective baseline value for each subject, creating deviation scores to normalize for differences in
thebaselinemeans and variabilitybetweenthe fixed-support and sway-references support groups.
Fordeterminationofshort-termadaptation,weanalyzedthecentereddatafortrials1–9withineach
day.Toassesslongertermadaptation,weaveragedthecentereddatafromallninetrialsineachday
(day1throughday3)tocomparebetweentestdays.Averagingoftrialswithindayswasjustifiedby
thefactthattrial-to-trialtrendsinbothconditionsdidnotdifferbetweentestingdays(i.e.,nosignif-
icantday-by-trialinteraction.Sinceoneofthegoalsoftheexperimentwastodeterminewhetherpar-
ticipantsintheirrespectivetestconditionwouldbeabletofullyadapttothevibrationperturbation,
wedeterminedthetaskas‘‘learned’’(vibrationeffectsfullycompensated)whenparticipantshadES
withinonestandarddeviationoftheirinitialbaselinetestingmean.
ApEnmeasureswerecalculatedusingacustomizedMatLabR2008a(Mathworks,Natick,MA)code
andanterior-posteriorCOP-displacementdataforeach20-strialoftheexperiment(2000datapoints
pertrial).Accordingtodataprocessingprotocolsestablishedinformerstudies,thefollowingsettings
wereappliedforMatLabanalysis:Aserieslengthof2(m=2datapoints),anerrortolerancewindow
of0.2timesthestandarddeviationoftherespectivetimeseries(r=0.2);andalagvalueof10,reduc-
ingtheeffectivesamplingrateto10Hz(Cavanaughetal.,2006).UsingtheMatLabcode,weobtaineda
singleApEnvalueforeachtrial,wetheneitherusedsinglevaluesoraveragedthedataaccordingto
themethodpresentedforESabove.
Asavalidationfortheapplicationofapproximateentropyasanon-linearanalysistool,itneedsto
beidentifiediftheCOPtimeseriesdataaredeterministic.Thisisdonebyshufflingthedatapoints
using a transformation algorithm (Theiler, Eubank, Longtin, Galdrikian, & Farmer, 1992) in MatLab
2008r(Mathworks,Natick,MA).Usingthistransformation,asurrogatesetofdatawascreated.Surro-
gateandoriginaldata(ApEn)werethencomparedviat-tests(a=.05).Thereweresignificantdiffer-
ences betweenthe setsof data pointsfor each pair ofCOP time-series,demonstratingnon-random
characteristicsoftheoriginaltime-seriesdata.
StatisticalanalysiswasperformedusingIBMSPSSStatistics20.0.Mixedeffectslinearmodelanal-
yseswithpolynomialcontrastsoftherepeatedmeasureswereusedtotestbothdependentvariables
(ES, ApEn) for analysis of between-groups, within-groups, and interaction effects when comparing
eitherbaselinevs.firstvibrationtrial(earlyeffects),theninevibrationtrialswithindays(short-term
adaptation),orthreetestingdays(long-termadaptation).Randomeffectsforsubjectswereincluded
inthemodelstoaccountfordependencyoftherepeatedmeasures.Themodelfortestingearlyeffects
(firsttrialofvibration)onESincludedheterogenousvariancesbecausethevariancesestimatedfrom
theobserveddataweresubstantiallydifferent;allotheranalyseswereunadjustedsincetheobserved
datawereconsistentwithhomogeneousvariance.TheHuynh–Feldtepsilonwasusedasappropriate
toadjustthedegreesoffreedomfordeviationfromsphericityofrepeatedmeasures.
3.Results
Eighteenparticipantscompletedthefamiliarizationtrials,baselinetestingandthreeconsecutive
daysofvibrationtests.Allparticipantsassignedtothesway-referencedsupportconditionwereable
toreturntobaseline(novibration)performancewithinthevibrationtestingperiod,whereasnopar-
ticipantinfixedsupportwasabletoachieveascorewithinonestandarddeviationfrombaselineper-
formancewithinthecourseofthetestingperiod.
M.Dettmeretal./HumanMovementScience32(2013)214–227 219
3.1.Earlyeffects–equilibriumscore
Data from the first trial of vibration exposure revealed an early effect on postural performance
regardless of surface support condition (see Fig. 2) as measured by equilibrium score,
F(1,19.98)=30.479,p<.001.Theequilibriumscoreinthefixed-supportgroupdecreasedfromabase-
line value of 93.2±2.3 to 67.4±14.5, F(1,8)=26.044, p=.001, and decreased from 74.0±7.0 to
61.6±12.9inthesway-referencedsupportgroup,F(1,8)=10.010,p=.013.Therewasnosignificant
group-by-trials interaction effect, F(1,19.98)=3.742, p=.067. There was a significant group effect,
F(1,19.98)=13.221,p=.002;scoresinthefixed-supportgroupweregenerallyhigherduringbaseline
andinitialvibration.
3.2.Short-termadaptation–equilibriumscore
Analysisofshort-termadaptationfortheninetrialswithineachdayundervibrationshowednosta-
tisticallysignificantchangesofcenteredequilibriumscoreforDay1,F(6.85,109.64)=1.076,p=.383,
Day 2 F(7.02,112.31)=1.327, p=.224, and Day 3, F(6.01,96.19)=2.146, p=.055. There was no
group-by-trials interaction for Day 1, F(6.85,109.64)=1.503, p=.175, Day 2, F(7.02,112.31)=0.286,
p=.959,andDay3,F(6.01,96.19)=0.485,p=.818.Fig.3illustratesthatposturalstabilityshowsnoevi-
denceofadaptationwithinatestingday.Thefixed-supportgroupalsodisplayedhigheroverallpostural
performance(i.e.,higherequilibriumscores)comparedtothesway-referencedsupportgroup.
3.3.Longer-termadaptation–equilibriumscore
Longer-term adaptation over the three days increased the centered equilibrium score in both
groups, F(2,32)=14.674, p<.001 (see Fig. 3). There was a significant difference in centered ES be-
tween Days 1 and 2, F(1,32)=14.155, p=.001 and between Days 2 and 3, F(1,32)=16.005,
p<.0001.Nosignificantgroup-by-daysinteractioneffectwasfound,F(2,32)=.749,p=.481;repeated
administration of the test under vibration lead to similar improvements of balance performance in
bothsurfaceconditiongroups.Overthecourseofthethreetestingdays,bothgroupsimprovedtheir
posturalcontrolduringvibrationatthesamerate.Centeredequilibriumscoredifferedsignificantly
betweengroups,F(1,16)=1.149,p=.006.Posturalperformance,unadjustedequilibriumscores,over
thethreetestingdayswasoverallhigherinthefixed-supportgroup(Fig.3).However,Fig.4illustrates
thatthefixed-supportgroupdisplayedsignificantlygreaterdecrements(centeredequilibriumscores)
inresponsetovibrationrelativetotheirbaselinemeasureswhencomparedtothesway-referenced
group.
100
90
e
or 80
c
S
m
u 70
bri
quili 60
E
50
40
no vibration vibration
Fig.2. EquilibriumScore(asameasureofposturalstability,rangingfrom0–100)meansandstandarddeviationoffixed
support condition (dashed line) and sway-referenced support condition (solid line) during both baseline testing and first
vibrationtrial.
220 M.Dettmeretal./HumanMovementScience32(2013)214–227
85
75
e
or
c
S
m
briu 65
uili
q FS
E
SRS
55
Day 1 Day 2 Day 3
45
Fig.3. ComparisonofEquilibriumScoresinfixed-supportcondition(dashedline)andsway-referencedsupportcondition(solid
line)duringninevibrationtrialsperday(27trialstotal).Openstarsrepresentthemeanvalue(plusorminusonestandard
deviation)forthetrialsofaparticulartestingday.
3.4.Earlyeffects–approximateentropy
Therewasasignificantmaineffectofvibration,F(1,16)=5.386,p=.034.Changesofapproximate
entropywerefoundonlyinthesway-referencedgroup(seeFig.5)overthecourseofthefirstvibration
trial. Approximate entropy values in the fixed-support group were 0.736±0.149 without vibration
and0.794±0.161duringinitialvibrationtrial,F(1,8)=1.002,p=.346.Inthesway-referencedgroup,
approximate entropy increased significantly from 0.740±0.134 to 0.853±0.206, F(1,8)=6.167,
p=.038. Therewas nogroup effect,F(1,16)=0.223,p=.643 or time-by-group effect,F(1,8)=0.556,
p=.467.
10
0
es
cor -10
S
e
c
n
e
er
Diff -20
S
E FS
SRS
-30
Day 1 Day 2 Day 3
-40
Fig.4. ComparisonofthedifferencescoresbetweenthebaselineESandtheESinfixed-supportcondition(dashedline)and
sway-referencedsupportcondition(solidline)duringninevibrationtrialsperday(27trialstotal).Openstarsrepresentthe
meanvalue(plusorminusonestandarddeviation)forthetrialsofaparticulartestingday.
M.Dettmeretal./HumanMovementScience32(2013)214–227 221
3.5.Short-termadaptation–approximateentropy
Analysisofshort-termadaptationforninetrialswithinadayundervibrationshowedstatistically
significant changes of approximate entropy values for Day 1, F(7.08,128)=6.953, p<.001, Day 2
F(8,128)=5.744,p<.001,andDay3,F(8,128)=2.434,p=.018(seeFig.6).Therewasnogroup-by-tri-
als interaction for Day 1, F(6.85,128)=1.612, p=.138, Day 2, F(8,128)=0.865, p=.548, and Day 3,
F(8,128)=0.655,p=.730;approximateentropyaveragesbetweengroupswerenotdevelopingdiffer-
ently between groups over the trials. Fig. 6 also illustrates that there was a general trend in both
groups toward increasing regularity within a testing day and this trend was more pronounced in
thefixed-supportsurfaceconditiongroup.Thefixed-supportgroupalsodisplayedmoreoverallregu-
larity(i.e.,lowerApEnscores)comparedtothesway-referencedsupportgroup.
3.6.Longer-termadaptation–approximateentropy
Repeatedtestingoverdayshadnostatisticallysignificanteffectoncenteredapproximateentropy,
F(1.65,32)=1.365,p=.269(seeFig.6).Therewasnostatisticallysignificantgroup-by-daysinteraction
effect,F(2,32)=2.915,p=.081,indicatingthatneithergroupdisplayedlong-termadaptationinthis
measure. Approximate entropy values were overall higher in the sway-referenced group,
F(1,16)=9.306,p=.008.
4.Discussion
Inthepresentstudy,theinfluenceofpracticeduringabipedalposturalcontroltaskwasinvesti-
gatedwithdifferentcombinationsofsupportsurfacestabilityandproprioceptiveperturbation(Achil-
lestendonvibration).Wesoughttocharacterizebothshortandlonger-termadaptationoutcomesthat
would suggest compensation to the proprioceptive perturbation. Further, we used both linear and
non-linearmeasuresofforceplate-derivedCOPdatatoexplorefeaturesofposturalcontrolandadap-
tationthatarenotcapturedbytraditionallinearmeasuresofposturalperformance.
4.1.Balanceperformanceduringbaselineandfirstvibrationtrial
Theobservedbaselinevaluesexhibitedtheexpectedcharacteristicsofposturalstabilityundersta-
ble(FS)orunstablesupport(SRS)conditions:ESweresignificantlybetterinFSthaninSRS.Standing
1.2
1.1
1
0.9
n
pE 0.8
A
0.7
0.6
0.5
0.4
no vibration vibration
Fig.5. ApEn(asameasureofswayregularity)meansandstandarddeviationoffixedsupportcondition(dashedline)andsway-
referencedsupportcondition(solidline)duringbothbaselinetestingandfirstvibrationtrial.
222 M.Dettmeretal./HumanMovementScience32(2013)214–227
1
0.8
n
E 0.6
p
A
0.4
FS
SRS
Day 1 Day 2 Day3
0.2
Fig.6. ComparisonofApEnscoresinfixed-supportcondition(dashedline)andsway-referencedsupportcondition(solidline)
duringninevibrationtrialsperday(27trialstotal).Openstarsrepresentthemeanvalue(plusorminusonestandarddeviation)
forthetrialsofaparticulartestingday.
quietlyonastablesurfaceprovidesmoreopportunityforretrievalofsensoryinformationusedinthe
controlofbipedalposture.IntheFS-conditionwithoutvibration,musclespindlesandmechanorecep-
torsofthesolesofthefeetareabletoprovidetheCNSwithreliableinformationrelatedtopostural
swaythatisusedtoregulateCOPmotion(Maureretal.,2006).SRSreducesopportunitiesforobtaining
reliableinformationtomanageposturalcontrol,sincereliableproprioceptiveinformationisgreatly
minimizedduetoswayreferencing.Theearlyeffectsofanklevibrationonlinearmeasuresofpostural
controlwasconsiderablylessinSRS(ESdecreasedbyapproximately9%)comparedtoFS(ESdecreased
byapproximately28%).Althoughthegroup-by-trialsinteractioneffectapproachedbutfailedtoreach
statisticalsignificance(p=.067),themagnitudeofthedifferencesobservedinbothFig.2andthedif-
ferencebetweengroupsoftheday1trial1deviation-from-baselinevaluesplottedinFig.4suggest
thatthesupportsurfaceinfluencestheresponsetovibration(seefurtherdiscussionbelow).
4.2.Short-andlongertermdevelopmentofbalanceperformance
Franssonandcolleaguesinvestigatedadaptationeffectsinaseriesofposturaltasktrialsinvolving
sensory perturbation in the form of musculo-tendon vibration (Fransson, Johansson, et al., 2003;
Tjernströmetal.,2005).Theypresentedanexplanationforthedevelopmentofposturalperformance
curvesovertime,basedonseveralmajorfactors:1)apotentialsuppressionof,oradaptationtovibra-
tionstimulibasedonsensoryreweightingandassociatedshifttowardsunperturbedafferentssources,
2)posturalstrategychangesincludingthepotentialmodificationofkinematicdegrees-of-freedomby
alteringposturalorientation,associatedwithchangesinmuscleco-contractionandanklestiffnessand
3)consolidation,representinglonger-termchangesofposturalcontrolbasedonmotormemoryand
experience,thatfunctiontoimproveresponsestoappliedconstraintsandperturbations.Thisexpla-
nationprovidesaframeworkinwhichtointerpretthecurrentdata.
Theequilibriumscoredatarevealednosignificantshorttermadaptationacrosstheninetrialsona
givenday,regardlessofcondition.Thisfindingsuggeststhatduringthisshorttimeperiod,potential
postural control changes did not assist in reducing postural sway significantly. However, repeated
administration of vibration over consecutive days led to improvements in ES in both support
conditions and the rate of this improvement was similar. This finding suggests that with repeated
exposuretovibration,adaptationdidoccurwiththeassociatedassumptionthattheposturalcontrol
systemwassuppressingunreliableinputsandplacinggreaterweightonmorereliablesensoryinfor-
mation.Repeatedadministrationofposturaltaskshaveshowntoevokereductionofbodymovements
due to adaptation (Fransson, Johansson, et al., 2003; Nashner, Black, & Wall, 1982; Wrisley et al.,
2007).
M.Dettmeretal./HumanMovementScience32(2013)214–227 223
Thecurrentdataareconsistentwithearlierworkconcerningtheattenuationoftheeffectsofmus-
cletendonvibrationwhenadynamicproprioceptivestimulusintheformofanunstableplatformis
added(Ivanenkoetal.,1999).Ithasbeenproposedthatthisdampenedresponsetovibrationinthe
face of postural instability is the result of sensory down-weighting of ankle proprioceptive input.
Accordingtothishypothesis,unreliableinputsourcesarelessinfluentialwhileintegrationofreliable
sensoryinformationisweightedmoreheavilyforuseinthemaintenanceofposturalstability(Isableu
etal.,2010;Oie,Kiemel,&Jeka,2002;Vuillerme&Pinsault,2007).Fig.4illustratesthattheSRSgroup
waslessnegativelyimpactedbytheadditionofvibrationrelativetotheFSgroup,therebysuggesting
theywerepotentiallymoreeffectiveindown-weightingthedisruptedproprioceptionresultingfrom
the application of vibration to the ankle musculature. Conversely, the FS group continued to be
stronglyimpactedbyvibrationacrossallthreetestingdays.Thiscouldbetheresultofthesubjects
continuingtoattempttorelyonproprioceptiveinformationthatisnormallyreliableduringfixedsur-
facesupportconditions.Thedatadorevealthatdespiteexperiencinglargeperformancedecrements
withvibration,theFSgroupdidimproveoverthecourseofthethreetestingdays,suggestingthatcon-
solidationwasoccurring.
The potential underlying mechanisms associated with reweighing of sensory input are not fully
understood,butcouldbebasedoninhibitoryspinalpathwaysoraprocessthatsuppressestheeffect
of increased firing of proprioceptive afferents under vibration when the muscle is being stretched
(Bove, Nardone, & Schieppati, 2003; Radhakrishnan, Hatzitaki, Patikas, & Amiridis, 2011). This pro-
posedreweightingprocessreflectsanappropriateresponsetothedecreasedreliabilityofpropriocep-
tiveinputassociatedwithanunstablesupportsurfacethatpreventsonefromoptimallydetermining
posturalorientation.InthecaseofSRS,increasedswayresultingfromvibrationwouldbeattenuated
relativetoFS,duetodecreasedintegrationofunreliableanklemusclespindleinformationforpostural
control.Additionally,duetothenatureofthetaskinSRS,thereismoreposturalswayandcorrespond-
ingmultiplemuscleactivationinthelowerlimbs.Theincreasedextra-fusalactivationmighthavean
additionalimpactonattenuatingvibrationaleffects,byunloadingthespindlereceptorsandthereby
affectingtheirfiringrate(Radhakrishnanetal.,2011).
TheobservedeffectwouldrepresentanexampleoftheCNS’sabilitytodown-weightorsuppress
singlesensoryinformationchannelsdependingontheirpositiveornegativeinfluenceonposturalsta-
bilityunderspecificconditions(Hatzitakietal.,2004).Therehavebeenmultipleapproachestodeter-
mine the nature of sensory reweighting in postural control, but it remains unclear, to what extent
presynaptic mechanisms, spinocerebellar- or higher motor processing contribute to the changes in
weightingandintegrationofafferentinformation.Thedatapresentedinthisstudyareconsistentwith
thesensoryreweightinghypothesis.
OurfindingsareinaccordancewiththeoriesregardingthecapabilityoftheCNStoutilizedifferent
afferent information sources depending upon specific task demands and constraints. Over time, a
refiningofneuromotorprocessingimprovesfunctionalposturalcontrol.Inaddition,mechanismsof
consolidation (longer-term learning) are being utilized in order to improve postural balance
(Tjernström,Fransson,Hafström,&Magnusson,2002).Asourcriteriaforlearningthetaskwasare-
turnofESwithinonestandarddeviationofinitialbaselinetesting,onlygroupSRSparticipantswere
abletofullyadapttotheconditionsandreturntobaseline.Thissuggeststhattheprocessoflonger-
termadaptationwasmoreadvancedintheSRSconditionrelativetoFS.Althoughimprovementswere
observedinFSaswell,potentialmechanismsofsensoryreweighting,leadingtoposturalcontrolstrat-
egiesshiftsormotorlearningwerenotsufficienttofullyadapttotheperturbedsensoryinputemerg-
ingfromanklemusclespindles,atleastnotwithinthetrainingperiodofthreedays.SinceESinFSstill
showedanupwardtrendatthecompletionofthethree-daytestingperiod,itwouldbeinterestingto
investigatewhetherhealthyparticipantscouldcompletelyovercomemusclespindleperturbationon
thestablesupportwithadditionalpractice.
4.3.Reflectionofvibrationeffectsandadaptationinswayregularity
Thefoundationofapproximateentropyrepresentsthepredictabilityorrepeatabilityofdatapoints
inatimeseries.AnextensivesummaryofApEncomputationandpotentialapplicationsisprovided