1 Biomechanics and energetics of running on uneven terrain 2 Alexandra S. Voloshina and Daniel ...

37 Pages·2014·2.74 MB·English

©2015.PublishedbyTheCompanyofBiologistsLtd|TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
RESEARCHARTICLE
Biomechanics and energetics of running on uneven terrain
AlexandraS.Voloshina1,2,*andDanielP.Ferris1
ABSTRACT McAndrew et al., 2010) and could utilize the same strategy to
improve running stability as well. A recent study showed that
In the natural world, legged animals regularly run across uneven
assistingwithlateralbalanceduringrunningresultedinreducedstep
terrain with remarkable ease. To gain understanding of how
widthvariabilityandenergyexpenditureinhumans(Arellanoand
runningonuneventerrainaffectsthebiomechanicsandenergetics
Kram, 2012). In contrast, if uneven terrain leadsto increased step
of locomotion, we studied human subjects (N=12) running at
2.3ms−1onanuneventerraintreadmill,withuptoa2.5cmheight width variability, it may contribute to increased energetic coststo
maintainbalance.
variation. We hypothesized that running on uneven terrain would
Changes in surface height variability are likely to alter muscle
show increased energyexpenditure, step parameter variabilityand
activation patternsand mechanical work duringrunning.Running
leg stiffness compared with running on smooth terrain. Subject
energy expenditure increased by 5% (0.68W kg−1; P<0.05) when onsand,forexample,resultsingreatermuscleactivity,greaterhip
andkneemotion,andgreaterpositivemechanicalworkcompared
runningonuneventerraincomparedwithsmoothterrain.Stepwidth
with running on a smooth flat surface (Lejeune et al., 1998;
andlengthvariabilityalsoincreasedby27%and26%,respectively
Pinningtonetal.,2005).Thesechangesarelikelycontributorstothe
(P<0.05). Positive and negative ankle work decreased on uneven
terrainby22%(0.413Jkg−1)and18%(0.147Jkg−1),respectively increasedenergyexpenditureforrunningonsandcomparedwitha
smooth,hardsurface(Pinningtonetal.,2005).Inaddition,wehave
(P=0.0001 and P=0.0008). Mean muscle activity increased on
previouslyfoundthathumansubjectsshowedgreaterhipandknee
uneventerrainforthreemusclesinthethigh(P<0.05).Legstiffness
jointflexionmotionsduringswingwhenwalkingonuneventerrain
also increased by 20% (P<0.05) during running on uneven terrain
comparedwitheventerrain(Voloshinaetal.,2013).Itisreasonable
comparedwithsmoothterrain.Calculationsofgravitationalpotential
toexpectsimilarmodificationsinswinglegdynamicsforrunning
energyfluctuationssuggestthatabouthalfoftheenergeticincreases
onasimilarunevensurface.Inadditiontoincreasedmuscleactivity,
canbeexplainedbyadditionalpositiveandnegativemechanicalwork
runningontheunevensurfacemayalsodisruptpatternsofmuscle
for up and down steps on the uneven surface. This is consistent
recruitment.Effectiveenergystorageandreturnduringlocomotion
betweenwalkingandrunning,astheabsoluteincreasesinenergetic
requiresmuscleactivationtoproduceaconcertedcontraction(Hof
costforwalkingandrunningonuneventerrainweresimilar:0.68and
0.48W kg−1, respectively. These results provide insight into how etal.,1983),orcontractionsthatminimizethechangeinlengthof
themusclefiberandmaximizetendonandaponeurosisstretchand
surface smoothness can affect locomotion biomechanics and
recoil. Because work produced from elastic energy storage and
energeticsintherealworld.
return contributes to about half of the total mechanical work
KEYWORDS:Energyexpenditure,Jointwork,Kinematics,Leg performed during running (Cavagna et al., 1964), a reduction in
stiffness elastic work would require increased muscle work and would be
more energetically costly. As a result, these factors have the
INTRODUCTION potentialtocontributetorunningenergeticcostsrelatedtosurface
Empiricalmeasurementshavedocumentedthatrunningonnatural smoothness.
surfaces such as sand, grass or irregular trails requires greater Legstiffnessisalsolikelytochangewithincreasedsurfaceheight
metabolicenergyexpenditurethanrunningonsmoothhardsurfaces variability. During human reaching in the presence of expected
(Jensenetal.,1999;Lejeuneetal.,1998;PinningtonandDawson, mechanicalperturbations(Burdetetal.,2001;Franklinetal.,2007), y
2001; Zamparo et al., 1992). Such natural terrain has many arm stiffness tends to increase compared with reaching without g
o
mechanicalpropertiesthatcaninfluencerunningbiomechanics.For perturbations.Thenervoussystemmayrespondtoexpectedlower l
o
example,humansandanimalsneedtoconstantlyadjustforchanges limbperturbationssimilarly.Inaddition,Grimmeretal.haveshown i
B
insurfacedamping,complianceandsmoothnessduringlocomotion thathumanrunnersincreaselegstiffnessinanticipationofasingle
l
in the real world. To identify the energetic and biomechanical step-up when running on an uneven track (Grimmeret al., 2008). ta
changes during running caused by increased surface height However, runners then decrease their leg stiffness for the actual n
e
variability, we studied human running on an uneven surface step-up, possibly to smooth out the perturbation to the center of m
designedtomimicnaturalterrain. mass trajectory. It is likely that runners would also adjust leg ri
e
Step parameter adjustments are one potential factor that could stiffnessinresponsetorunningonuneventerrain. p
contribute to increased energy expenditure during running on In this study, we examined the energetic and biomechanical x
E
uneven terrain. In particular, humans adjust step width for changes during running on uneven terrain when compared with
f
maintaining lateral balance during walking (Hak et al., 2012; runningonsmoothterrain.Weusedanuneventerrainsurfacethat o
l
was attached to a standard exercise treadmill (Fig. 1) to collect a
n
1SchoolofKinesiology,UniversityofMichigan,AnnArbor,MI,USA.2Departmentof continuous energetic and kinematic data of human runners. We r
MechanicalEngineering,UniversityofMichigan,AnnArbor,MI,USA. hypothesizedthat,onuneventerrain,subjectswouldshowgreater u
o
energy expenditure and step parameter variability. Based on J
*Authorforcorrespondence([email protected])
e
previous research indicating increased limb stiffness under
h
Received1July2014;Accepted18December2014 conditions of anticipated mechanical perturbations, we also T
711
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
Fig.1.Studyapparatus.(A)Uneventerraintreadmillusedfor
therunningstudies.(B)Schematicrepresentationoftheuneven
surface,withsteppingareasofthreedifferentheights(arrows
indicatethetreadmill’slongaxis).(C)Close-upoftheblocks
comprisingthesteppingareas.Dimensions:H,1.27cm;L,
15.2cm;W,2.54cm.
expected runners to increase leg stiffness on uneven terrain the percentage increase found during walking on uneven terrain.
compared with smooth terrain. Our overall objective was to Incontrasttorunning,metabolicenergyexpenditureduringwalking
provide insight into how the biomechanical adjustments lead to onuneventerrainincreasedfrom2.51±0.24to3.19±0.14Wkg−1
greaterenergyexpenditureduringrunningonunevensurfaces. (P=0.0004), or by ∼27%. This increase in metabolic cost was
consistentwiththe28%energyincreasefoundinourpreviousstudy
RESULTS onwalkingonuneventerrain(Voloshinaetal.,2013).Thissuggests
Runningonuneventerrainresultedinincreasedenergyexpenditure thatthebiomechanicaladaptationsduringwalkinginthisstudyare
compared with running on smooth terrain. Several biomechanical likely the same as we have previously described, even though
adjustmentscontributedtothisincreaseinenergeticcost.Subjects walking surfaces and subject footwear differed between the two
did not exhibit changes in mean step parameters between the two studies. Although the percentage increases in energy expenditure
conditions,butthereweredifferencesinstepparametervariability. weredifferentbetweenwalkingandrunning,theabsoluteincreases
Jointangle,torqueandpowerweremostlyunaffectedbytheterrain in energetic cost were similar: 0.68 and 0.48W kg−1 for walking
and only the ankle joint showed a significant decrease in joint and running, respectively. The mean standing metabolic rate was
power.Inaddition,weobservedincreasedmuscleactivityinthree 1.46±0.17Wkg−1.
proximal leg muscles (vastus medialis, rectus femoris and medial
hamstring), and increased muscle mutual contraction between the Kineticsandkinematics
vastus medialis and medial hamstring muscles. Subjects also Wesawnochangesinmeanstepparameters(width,length,height
demonstrated higher leg stiffness when running on uneven terrain and period), although step variability increased for all parameters
comparedwithsmoothterrain. during running on uneven terrain compared with even terrain
(Table1).Inparticular,stepwidth,lengthandheightvariabilityall
Metabolicenergyexpenditure increased significantly by approximately 27%, 26% and 125%,
Running andwalking ontheuneven terrainresulted insignificant respectively (P<0.05) on uneven terrain. In addition, step period
increases in energyexpenditurewhen compared with runningand variability on uneven terrain increased significantly by 30%
walking on the even surface (Fig. 2). During running, energetic (P<0.05).
cost increased from 9.72±0.65 to 10.2±0.94W kg−1 (mean±s.d., Subjects showed few changes in joint kinematics and kinetics
P=0.008),orabout5%fromeventouneventerrain.Thispercentage duringrunningonuneventerraincomparedwitheventerrain,with
increase in running energy expenditure was much lower than mostnotablechangesoccurringattheanklejoint.Duringrunning
onuneventerrain,jointanglesinthesagittalplaneshowedslightly
higherpeakflexionanglesinthekneeandhipduringmid-stance,
possibly to allow for greater leg clearance (Fig. 3). Qualitative y
examinationoftheankleangleshowedaslightlydecreasedrangeof g
o
motion on uneven terrain compared with even terrain, although l
o
subjectsappearedtomaintainasimilarheel-strikefootfallpatternon i
B
thetwosurfaces.Thisreducedrangeofanklemotionsuggeststhat
l
a
subjectsranwithslightlyflatterfeetwhenonunevenground.The t
n
ankle joint also showed an approximately 14% decrease in peak e
joint moment, around mid-stance. In contrast, the knee and hip m
jointsshowedlittlechange.Changesinjointpowerwereonlyseen ri
e
aroundtheankleandknee,withthetwojointsshowingdecreasesin p
power around mid-stance (by 29% and 23%, respectively), when x
E
running on uneven terrain. Ankle power also decreased by 25% f
Fig.2.Netmetabolicrateforwalkingandrunningontheevenanduneven priortopush-offandat∼40%ofstridetime.Thetimingoftoe-off o
l
surfaces.Allnetmetabolicratesarenormalizedtosubjectmassandshowthe withrespecttostridetimingdidnotdifferbetweenthetworunning a
n
absolutechangesinenergeticswhenwalkingandrunningonuneventerrain conditions. r
comparedwithsmoothterrain.Percentagesindicatetheincreasesinenergetic u
Joint motion variability was also greater on uneven terrain
o
costcausedbyuneventerrainwhencomparedwithevenwalkingorrunning.
comparedwitheventerrain(Fig.3).Surfaceunevennessincreased J
Asteriskssignifyastatisticallysignificantdifferencebetweentheevenand e
unevenwalkingandrunningconditions(posthocpair-wisecomparisons, ankle and knee angle variability by approximately 25%, and hip h
α=0.05). angle variability by 35% when compared with even terrain (all T
712
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
Table1.Stepparametersforrunningontheevenandunevensurfaces
Even Uneven P-value
Mean±s.d. Variability(mean±s.d.) Mean±s.d. Variability(mean±s.d.) Mean Stepvariability
Stepwidth 0.055±0.029 0.022±0.004 0.059±0.033 0.028±0.006* 0.353 <0.0001
Steplength 0.881±0.051 0.035±0.009 0.884±0.044 0.044±0.011* 0.385 0.0002
Stepheight – 0.004±0.001 – 0.009±0.002* – <0.0001
Stepperiod(s) 0.729±0.041 0.010±0.003 0.731±0.033 0.013±0.003* 0.427 0.0014
Parametersincludestepperiodandthemeanstepwidth,lengthandheightandtheirrespectivevariabilities(allnormalizedtosubjectleglength,mean0.944m).
Stepvariabilityisdefinedasthes.d.ofstepdistancesoveratrial,reportedasthemean(±s.d.)acrosssubjects.
Asteriskssignifyastatisticallysignificantdifferencebetweentheevenandunevenrunningconditions(posthocpair-wisecomparisons,α=0.05).
P<0.05). Ankle moment variability also increased by about 60%, All but three muscles showed a significant increase in
while knee and hip moment variability more than doubled on electromyographic (EMG) variability when running on the
uneven terrain (all P<0.05). Joint power variability increased by uneven terrain (Fig. 5). Only two muscles in the lower leg and
50%fortheankleand∼70%forthekneeandhip(allP<0.05). three muscles in the thigh showed increases in variability (s.d. of
Running on the uneven surface also affected the amount of muscleactivity),withthemeanincreaseinvariabilitybeingslightly
positiveandnegativejointworkdoneattheankle(Fig.4).Positive higherinthethighmuscles(meanincreaseof14%and25%inthe
ankleworkdecreasedby0.413Jkg−1(22%)whilenegativeankle lowerlegandthighmuscles,respectively).Inthelowerleg,boththe
work decreased by 0.147J kg−1 (18%; P=0.0001 and P=0.0008, soleus(SO)andlateralgastrocnemius(LG)showed14%increases
respectively).Positiveandnegativejointworkforthekneeandhip in s.d. (P<0.05), while the tibialis anterior (TA) and medial
werenotstatisticallydifferentbetweenthetworunningconditions. gastrocnemius(MG)showednosignificantchanges.Forthethigh
muscles,VM,RFandMHshowed15%,35%and26%increasesin
Muscleactivation musclevariability,respectively(P<0.05).
Subjects showed increases in muscle activity variability, mean Out of the three pairs of antagonistic muscles, we observed
muscleactivityandmuscleco-activationinthethighmuscleswhen increasedmuscleco-activationovertheentirestrideonlyintheMH/
runningonuneventerraincomparedwithrunningonevenground VMmusclepair(Table2).However,whenwebrokedownthestride
(Fig. 5). Significant increases in mean muscle activity were only into 1%increments,we noticedsignificant increases in muscle co-
noted in three of the thigh muscles; vastus medialis (VM), rectus activationinthefirst5%ofthestrideintheMH/VLmusclepairas
femoris(RF)andmedialhamstring(MH)activityincreasedby7%, well. Similarly, the MH/VM pair showed increased muscle co-
20%and19%,respectively(P<0.05).However,thevastuslateralis activationduringthefirst5%ofthestridebutalsoslightlybeforeand
(VL) and all muscles in the lower leg showed no significant duringtoe-off.Themusclepairalsodemonstratedincreasedmuscle
differencesinmeanmuscleactivitybetweenconditions(Fig.6). co-activation during swing, although, because of minimal muscle
activityofthetwomusclesduringthistime,theseincreasesarelikely
Fig.3.Jointangle,torqueandpowerversusstride
timeforrunningonevenanduneventerrain.Plotted
insolidlinesarethemeantrajectoriesfortheankle,
kneeandhipagainstpercentagestridetimeforrunning
duringunevenandeventerrainconditions.Shaded
areasdenotethemeans.d.envelopesacrosssubjects
fortheunevencondition,dashedlinesfortheeven
condition.Stridesstartandendatsame-sideheel-
strike.Dashedverticalgraylineindicatestoe-off.Ext.,
extension;Flex.,flexion.
y
g
o
l
o
i
B
l
a
t
n
e
m
i
r
e
p
x
E
f
o
l
a
n
r
u
o
J
e
h
T
713
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
compared with even terrain (P=0.0002). In addition, vertical
ground reaction force variability more than tripled when running
onuneventerrain(P<0.05).
Subjectsraninaslightlymorecrouchedposturewhenonuneven
terrain compared with running on the even surface (Fig. 7A).
Subjects contacted the ground at heel-strikewith a more bent leg,
and hence a shorter leg length (0.992±0.022 and 0.984±0.022,
dimensionless, for even and uneven terrains, respectively;
P=0.0018), defined as the straight-line distance from the greater
trochanter marker to the fifth metatarsal markerof the stance foot
and normalized to mean subject leg length. Similarly, leg length
before toe-off decreased significantly from 1.02±0.011 to 1.01±
0.015 (dimensionless; P<0.0001). In addition, the minimum leg
length during mid-stance was longer on uneven terrain (0.907±
0.014 and 0.912±0.015, dimensionless, for even and uneven
terrains, respectively; P=0.0041). This resulted in a 15% decrease
in themaximum changein leg length, from 0.085±0.023on even
ground to 0.072±0.019 on uneven terrain (dimensionless;
P<0.0001).
Fig.4.Ankle,kneeandhipworkperstrideforthetworunningconditions.
Primarily due to different leg length dynamics, subjects ran on
Valuesareshownforpositiveandnegativeworkforthethreejoints,witherror
stifferlegswhenrunningonuneventerraincomparedwithrunning
barsdenotings.d.Dashedlinesindicatenetworkforthespecificjointand
condition,andasteriskssignifyastatisticallysignificantdifferencebetweenthe onevenground(Fig.7B).Usingthemoretraditional legstiffness
evenandunevenrunningconditions(posthocpair-wisecomparisons,α=0.05). calculation (defined as the ratio between the maximum vertical
groundreactionforceandthemaximumleglengthdisplacement),
inconsequential. The TA/SO muscle pair showed no significant wefounda20%differenceinlegstiffnessbetweenthetwosurfaces
increasesinmuscleco-activationatanypointinthestride(Fig.5). (from 27.9±6.40 on even terrain to 33.4±7.54 on uneven terrain,
dimensionless; P<0.0001). Similarly, the second leg stiffness
Verticalgroundreactionforcesandlegstiffness calculation(definedasthelinearfittotheverticalgroundreaction
Vertical ground reaction forces, normalized to subject weight, force versus leg displacement) showed a 10% increase in leg
remained largely unchanged for running on uneven terrain stiffness (from19.7±2.10oneventerrainto 21.8±2.74onuneven
compared with even terrain (Fig. 7A). The maximum force, fmax, terrain,dimensionless;P<0.0001).
occurred around 40% of stance and had a peak at 2.19±0.11
(dimensionless; P=0.753). The peak maximum force was not DISCUSSION
statisticallydifferentforthetworunningconditions.However,the In this study we quantified the changes in energetics and
impact peak increased byapproximately 17% (from 1.34±0.19 to biomechanics between running on uneven terrain and on flat,
1.57±0.25, dimensionless) when running on uneven terrain smooth terrain. Our findings supported our hypotheses, primarily
Fig.5.Averagedelectromyographic(EMG)
activityversusstridetimeforrunningoneven
anduneventerrain.AllEMGprofileswere
normalizedtothemaximummeanmuscleactivity
overthetworunningconditions,foreachmuscle
andsubject.Stridesstartandendatsame-side
heel-strikes.Dashedverticalgraylineindicates
toe-off.Shadedareasdenotethemeans.d. y
g
envelopesacrosssubjectsfortheuneven o
condition,dashedlinesfortheevencondition. l
o
Graybarsindicatestatisticallysignificant i
B
increasesinmutualmusclecontraction,with
l
darkercolorsindicatinglargerpercentage a
t
increasesfromtheeventotheunevenrunning n
condition.TA,tibialisanterior;SO,soleus;MG, e
m
medialgastrocnemius;LG,lateral
gastrocnemius;RF,rectusfemoris;MH,medial ri
e
hamstring;VL,vastuslateralis;VM,vastus p
medialis. x
E
f
o
l
a
n
r
u
o
J
e
h
T
714
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
Fig.6.MeanrectifiedEMGactivityvalues.MeansubjectEMGprofileswere
firstnormalizedtomaximummeanmuscleactivityoverthetworunning
conditions,foreachmuscleandsubject,andthenaveragedoverstridetimeto
producesubjectaverageEMGactivityvalues.SubjectaverageEMGactivity
wasthenaveragedoversubjectstoproducemeanEMGactivityvalues.Bars
indicates.d.acrosssubjects.Asteriskssignifyastatisticallysignificant
differencebetweentheevenandunevenrunningconditions(posthocpair-
wisecomparisons,α=0.05).
that running is more energetically costly on uneven terrain
compared with even terrain. However, this increase was much
smaller than the increase caused by the same surface during
walking.Morespecifically,wefounda0.68Wkg−1(27%)increase
inmetaboliccostduringwalkinganda0.48Wkg−1(5%)increase Fig.7.Verticalgroundreactionforces,effectiveleglengthandleg
stiffnesscalculationsforthetworunningconditions.(A)Meanvertical
during running. Although the percentage changes were quite
groundreactionforcesnormalizedtosubjectweight(solidlines),andeffective
differentbetweenthetwolocomotiontypes,itisimportanttonote leglengthsnormalizedtomeansubjectleglength(thickdashedlines),versus
that the absolute increases were very similar. These absolute stancedurationforrunningonevenanduneventerrain.Shadedareadenotes
energeticincreasescouldberelatedtothetotalmechanicalenergy themeans.d.envelopeacrosssubjectsfortheverticalgroundreactionforce
fluctuations caused by the uneven surface. For example, running ontheunevencondition,thindashedlinesindicatetheenvelopefortheeven
condition.(B)Normalizedverticalgroundreactionforceplottedagainstthe
uphill and downhill for an equal distance would result in greater
normalizedeffectiveleglengthforthetworunningconditions.Meanleg
energyexpenditurethanrunningonlevelgroundforthesametotal
stiffnessvaluesarepresentedfortwolegstiffnesscalculationmethods:k
max
distance (Margaria, 1968). If we equate the uneven surface to a equalsthemaximumforcedividedbythemaximumleglengthdisplacement
series of steps up and down an incline, it would be reasonable to andk istheslopeofthelinearfittothelegstiffnesscurve.s.d.valuesare
fit
expectanincreaseinenergyexpenditureaswell.Ifweconsiderthe acrosssubjects.
meansteplengthofourrunners(0.884m)andthemaximumstep y
height change of the uneven terrain (0.025m), our uneven terrain smaller.Thissuggeststhatfactorsotherthanchangesinmechanical og
surface would be roughly equivalent to running up and down a work contribute to energy expenditure on uneven terrain during ol
1.6%incline.Inaddition,wecouldexpectthisinclinetoresultinan running. Bi
energyincreaseof∼0.35Wkg−1(Margariaetal.,1963).However, As expected, we saw changes in step length and width l
a
thetrueenergetic increaseduetoinclinevariationsislikelymuch variabilities across the two surfaces. On uneven terrain, runners t
n
showed33%and26%increasesinstepwidthandlengthvariability, e
respectively(P<0.05).Thisisconsistentwithpastresearch,which m
Table2.Musclemutualcontractionovertheentirestride
has shown that challenges to locomotor stability tend to produce ri
Even(mean±s.d.) Uneven(mean±s.d.) P-value morestepvariabilityduringwalking(Thiesetal.,2005;Voloshina pe
TA/SO 105.7±36.34 109.5±32.83 0.6527 et al., 2013). Greater step variability during walking also seems x
E
MH/VM 106.4±28.05 134.7±34.31* 0.0168 related to active stabilizing adjustments for maintaining lateral f
MH/VL 123.3±66.60 125.5±28.02 0.9244 balance(BaubyandKuo,2000;Donelanetal.,2001).Incontrast, o
l
Valuessignifythedimensionlessareaundertheminimumofthenormalized duringrunning,humanstendtoprefernarrowstepwidthscloseto a
n
EMGcurvesforthetwomusclesofinterest.Threemuscleantagonistpairs themidlineofthebody(Cavanagh,1987).Thisisbecausenarrow r
arecompared:tibialisanterior/soleus(TA/SO),medialhamstring/vastus u
step widths result in reduced lateral moments about the center of o
medialis(MH/VM)andmedialhamstring/vastuslateralis(MH/VL).s.d.are
mass and tend to reduce energetic cost compared with larger step J
calculatedacrosssubjects.Asteriskssignifyastatisticallysignificant e
differencebetweentheevenandunevenrunningconditions(posthocpair- widths. Based on previous research (Arellano and Kram, 2011), h
wisecomparisons,α=0.05). reducingstepwidthandstepwidthvariabilityduringrunningleads T
715
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
to a reduction in energy expenditure. However, our subjects only stiffness adjustments across the surfaces. Vertical ground reaction
showeda27%increaseinstepwidthvariabilityandnochangein force profiles were also largely unchanged, with only the impact
the mean step width. These changes are relatively small and peak magnitude increasing by 17%. This increase suggests that
nowherenearthemagnitudenecessarytoproducea5%increasein subjectslandedwithahighercontactforce,likelyduetoflatterfeetat
energyexpenditure(ArellanoandKram,2011).Thissuggeststhat groundcontact.However,themaincauseofincreasedlegstiffnessis
the energetic increase caused by changes in step parameters was thechangeinlowerlimbposture.Whenrunningonuneventerrain,
negligible. subjectscontactedthegroundatheel-strikewithashorterleglength,
Asignificantfindingofourstudyisthattheabsolutechangesin and had a longer leg length during mid-stance and a shorter leg
energetic cost are independent of locomotor gait. Although lengthattoe-off.Thischangeinposture,andinturninlegstiffness,
percentage increases in energetic cost were significantly different canbetheresultofseveralfactors.Previousresearchonupperlimb
(27% and 5% for walking and running, respectively) the absolute movementshasshownthathumanstendtostiffentheirjointswhen
changeswererelativelyclose(0.68and0.48Wkg−1forwalkingand presented with unfamiliar tasks, likely in anticipation of potential
running,respectively).Thesimilarabsolutechangesforwalkingand perturbations. Similarly, Blum and colleagues (Blum et al., 2011)
runningsuggestthatthedominantfactorresponsibleforincreasesin haveshownthatamorecrouchedlegpostureduringavianrunning
metaboliccostduringlocomotiononunevensurfacesmayberelated maybeanadaptationmechanism,asitallowsforlengtheningand
tosurfaceheightvariabilityandthecorrespondingverticalmotionof shortening of the limb. As a result, the more crouched running
the center of mass. It would be interesting to examine a range of postureandoveralllargerlegstiffnessonuneventerrainislikelyan
surfaceheightvariabilitiesandtheireffectsonwalkingandrunning adaptationresponsetoanunfamiliarenvironment.Infuturestudies,
energetics. This could provide insight into whether walking and it may be interesting to look into the changes in leg stiffness
running have similar biomechanical mechanisms responsible for throughouttrainingperiods,wheresubjectsareallowedtobecome
energeticcostdifferences. familiarwiththesurfaceoverlongerperiodsoftime.
Another important finding of this study is that the lower limb This study had several limitations related to kinetic
joints that compensate for locomotion on uneven terrain are very measurements. The accuracy of the force measurements during
differentbetweenwalkingandrunning.Duringwalking,anklejoint uneven terrain running was one such limitation. As described
dynamics remain invariable while the knee and hip joints previously, the uneven terrain treadmill was placed atop two
compensate with greater positive work production (Voloshina supports,eachofwhichwasplacedontopofaforceplatform.The
et al., 2013). In contrast, running on uneven terrain only forces recorded from each platform were then added together to
significantlyaffects work done at the ankle joint. The most likely obtainthetotalgroundreactionforces.Thetreadmillwasnotrigidly
explanation for this contrast in joint kinetic adaptations is the attachedtothesupportsandtherewassomeslackinthebelttowhich
relianceondifferentbiomechanicalmechanismsforthetwogaits. the uneven surface was attached. As a result, the force data were
Runningcanbecomparedtoaspring-masssystem,withthelower noisier than data collected with our in-ground, instrumented
limb functioning as if it were a single compression spring (Farley treadmill.Toaccountfortheadditionalnoise,welow-passfiltered
andFerris,1998;McMahonandCheng,1990).Incontrast,walking the ground reaction force data using a cut-off frequency of 6Hz,
hasinvertedpendulumdynamicswithdifferentiationbyjointthatis ratherthanamoretraditionalcut-offfrequencyof25Hz.Weused
unlike running (Alexander, 1992; Farley and Ferris, 1998; Kuo, thesamefilteringtechniquesforbothsurfaceconditions.Totestthe
2001;McGeer,1990).Thesedifferencesinfundamentaldynamics validity of comparing running for these different experimental
suggestthateachgaithasdifferentbenefitsanddrawbackstojoint- setups, we compared a representative subject running on the in-
specificadaptationsonuneventerrain. groundinstrumentedtreadmillwiththesamesubjectrunningonthe
The decrease in ankle work seen during running on uneven regulartreadmillonsupportswithabarebelt.Averagepeakvertical
terrain compared with even ground is likely due to the high load and anterior–posterior ground reaction forces, as well as average
sensitivity of the ankle joint. Muscles at the distal joints rely on peak ankle moments, were within 4% of each other and highly
high-gain proprioceptive feedback and are often the first to correlated(i.e.R>0.997).Meancenterofpressuretrajectoriesonthe
encounter perturbations due to uneven terrain (Daley and regular treadmill were also within 6% of mean center of pressure
Biewener, 2006). In contrast, the more proximal knee and hip trajectoriesrecordedontheinstrumentedtreadmill.Fortheuneven y
joints are largely feed-forward controlled. This control strategy surface, the terrain attached to the supported treadmill likely g
o
improves runningstability by maintaining consistent limb cycling introduced additional variability to calculations of the center of l
o
buthasamorepronouncedeffectonthedistaljoints(Daleyetal., pressure, because foot orientation during ground contact is highly i
B
2007).Inaddition,wesawasmallreductionintheanklerangeof variable on uneven terrain. However, this error would have been l
a
motion,whichcouldhavealsoledtoareductioninjointwork.This multidirectional and likely did not affect the mean results of the t
n
reductioninanklemotionlikelystabilizedthejointinresponseto inverse dynamics calculations. Instead, the main effect on the e
theunpredictablerunningsurface.However,itcouldalsohaveled inversedynamicscalculationswouldhavebeenincreasedvariability m
toareductioninenergystorageandreturnintheAchillestendon, in parametersthroughout the stance phase. Examination of s.d. in ri
e
leading to reduced work at the ankle joint. Recently, a numberof theankleprofilesinFig.3showsslightlygreatervariabilityinthe p
research groups have demonstrated that ultrasound imaging can anklemomentcalculationsbutincreasedvariabilityisalsopresent x
E
track both muscle fiber and tendon displacements during human in the ankle angle data. The ankle angle data are independent of f
o
running. Future experiments using ultrasound imagining could centerofpressurecalculations,sosimilarchangesinvariabilityseen
provide greater insight into the muscle–tendon mechanics on inotherparameterssuggestthattheeffectofthecenterofpressure al
n
uneventerrain. variabilityfromtheuneventerrainsurfacewasnotlarge. r
u
Inconjunctionwithourhypotheses,subjectsexhibitedgreaterleg Anotherlimitationofthestudywasthatsubjectsranatonlyone o
stiffness when running on uneven terrain compared with even prescribed speed. They could not negotiate the terrain by altering J
e
terrain. Therewerechanges in muscle co-activation but they were their speed as is possible when running on natural surfaces. We h
small in magnitude and were likely not the major drivers of leg choseaslowrunningspeedtomaximizeoursubjects’comfortlevel T
716
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
anddidnottestarangeofspeeds.Running atfasterspeedscould fabric.Theothersideofthehook-and-loopfabricwasgluedontowooden
have resulted in more pronounced biomechanical differences. We blockswithawidthof2.55cmandlengthof15.2cm,andofthreevarying
also tested only one pattern of stepping areas and one range of heights(1.27,2.54and3.81cm).Then,weattachedtheblockstothebelt,
surfaceheightsfortheuneventerrainsurface.However,assubjects oriented lengthwise across the belt. As a result, the blocks could curve
around the treadmill rollers, because of their relatively short width. The
didnotappeartogetaccustomedtothesurface,wedonotbelieve
blocks comprised 15.2×15.2cm stepping areas (after Voloshina et al.,
the inherent pattern of the uneven terrain affected gait dynamics.
2013),inapatternthatwasdifficultforsubjectstoadopt.Wethenplacedthe
Largersurfaceheightvariabilitywouldhavelikelycausedamplified
unevensurfacebeltontopoftheregulartreadmillbeltandconnectedthe
biomechanicalandenergeticeffects.
ends of the second belt using zip-ties to form one continuous surface.
Additionallimitationsrelatetosubjecttrainingandtheinherent Subjects also ran on a separate, custom-built in-ground instrumented
difference between treadmill and overground running. For one, treadmill (Collins et al., 2009). The even surface served as the control
subjectswerealsonotallowedmultipledaystotrainandadjusttothe condition and allowed us to determine the biomechanical effects of the
terrain.Itmaybehelpfulinfuturestudiestodeterminewhetherthere unevensurfaceongaitbiomechanicsandenergetics.
are long-term adaptation effects. In addition, subjects ran with a Forbothsurfaces,subjectsranataspeedof2.3ms−1andwiththetrial
limitedvisibilityoftheterrainbecauseofthelengthofthetreadmill. orderrandomizedforeveryparticipant.Subjectsparticipatedinjustonetrial
foreachcondition,witheachtriallasting10minandwithaminimumof
Allsubjectswerecomfortablerunningonthetreadmillanddidnot
5minrestallowedbetweentrials.Onbothsurfaces,subjectswereinstructed
appear to be affected by limited visibility. However, it is possible
to run normally and not look down at their feet unless they felt it was
thatsubjectsmaynegotiatetheterraindifferentlyifmorevisibility
necessary.Generally,subjectschosenottolookattheirfeetduringrunning.
were allowed. For overground locomotion, runners typically have Aftertherunningtrialswerecompleted,subjectswalkedat1.0ms−1foran
extended visual feedback on the terrain and may choose different additional10minoneachsurface,whilewecollectedmetabolicdatatouse
pathsandfootplacementsinresponsetoterrainproperties. as a comparison with our previous study on walking on uneven terrain
In summary, we found that changes in mechanical work can (Voloshinaetal.,2013).Subjectsworerunningshoesoftheirchoiceforthe
explain approximately half of the energetic cost increase when experiments.
running on uneven terrain surfaces compared with flat, smooth
Metabolicrate
surfaces.Theotherhalfoftheenergeticcostincreasemayberelated Wemeasuredtherateofoxygenconsumption(V̇ )forallrunningtrials
to less efficient energy storage and return in elastic tendons and O2
using an open-circuit respirometry system (CareFusion Oxycon Mobile,
ligaments. Future studies using ultrasound imaging could provide
Hoechberg,Germany).Werecordedrespirometrydataforall10minofthe
greater insight into muscle fiber and tendon dynamics on various
runningandwalkingtrialsandalsofor7minduringquietstandingpriorto
terrains. We did find thathuman runners did not vary mechanical each datacollection. We allowed subjectsthe first 7.5min of the trial to
work done at their knee and hip joints when running on uneven reachsteady-stateenergyexpenditureandonlyusedthelast2.5minofdata
terraincomparedwithsmoothterrain.Instead,subjectsreducedlimb tocalculatethemetabolicenergyexpenditurerateofeachsubject.Tofind
mechanicalworkdoneattheanklejointwhenrunningontheuneven the metabolic rate, Ė (W), we used standard empirical equations as
met
surface. Using a similar control approach for legged robots with describedelsewhere(Brockway,1987;Weir,1949).Thenetmetabolicrate
biomimeticlimbarchitecturesmighthavebenefitsinincreasingthe wasfoundbysubtractingthestandingmetabolicpowerfromthemetabolic
power of all running conditions. All net metabolic power data were
relativestabilityofrunningasitaltersthelimbbiomechanicsclosest
tothefoot–groundinterface(Daleyetal.,2007). normalizedbysubjectbodymass(kg).
Kineticsandkinematics
MATERIALSANDMETHODS
For both even and uneven conditions, we recorded the positions of 31
Wemodifiedaregularexercisetreadmillbyattachinganadditionalbeltwith
reflectivemarkersusinga10-cameramotioncapturesystem(framerate:
wooden blocks of varying heights to the original treadmill surface. The
100Hz;Vicon,Oxford,UK).Weplacedmarkersonthepelvisandlower
blocks simulated an uneven surface on which subjects could run
limbsasdescribedpreviously(Voloshinaetal.,2013)andtapedthemon
continuously while we collected biomechanical and metabolic data.
totheskinorspandexshortswornbythesubjects.Althoughtrialslasted
Subjectsalsoranonaseparate,smoothtreadmillsurface,resultingintwo
for 10min, the first 7.5min allowed subjects to reach steady-state
testing conditions termed ‘uneven’ and ‘even’. For both surfaces, the
dynamics and we only used the last 2.5min of data to calculate step
runningspeedwasmaintainedat2.3ms−1,whilewecollectedkinematic,
parameters such as step width, length and height. For each subject and
kinetic,metabolicandEMGdata. y
trial,the2.5minofdataanalyzedconsistedofaminimumof250stepsand g
upto400steps.Toreducemotionartifact,welow-passfilteredallmarker o
l
Subjects dataat6Hz(fourth-orderButterworthfilter,zero-lag).Wedefinedstep o
Twelve young, healthy subjects participated in the study (seven males, width,lengthandheightasthedistancebetweenthelateral,fore–aftand Bi
fivefemales;mean±s.d.:age24.3±4.0years,mass68.6±7.1kgandheight verticaldistancesbetweenthecalcaneousmarkersonthetwofeetattheir al
175.5±7.1cm).Subjectsranontheevenandunevensurfacesduringthe respectiveheel-strikeinstances.Stepheightmeasurementswereusedonly t
n
same datacollection. For running on both surfaces,we collected oxygen to determine changes in step height variability caused by the uneven e
consumption (N=11), step parameter (N=12), EMG (N=10) and joint surface. We calculated the effective leg angle as the angle relative to m
kinematicsandkineticsdata(N=11).Becauseoftechnicalissues,somedata horizontalmadebythestraight-linedistancefromthegreatertrochanter ri
werenotcollectedforparticularsubjects,resultinginavalueofNlessthan markertothecalcaneousmarkerofthestancefoot.Theeffectivelegangle pe
12insomedatasubsets. andthepositionofthecalcaneousmarkersactedasameanstodetermine x
Priortotheexperiment,allsubjectsprovidedwritteninformedconsent. the time of heel-strike. This method of determining heel contact agreed E
Experimental procedures were approved by the University of Michigan wellwiththeonsetoftheverticalgroundreactionforce.Inaddition,we of
HealthSciencesInstitutionalReviewBoard. normalizedallmeasurementstosubjectleglengthmeasuredpriortoeach l
a
data collection and defined as the mean distance between the greater n
Runningsurfacesandtrialprocedures trochanterandcalcaneousmarkersofbothlegs. ur
Wecreatedanunevensurfacetreadmillbelt(Fig.1)thatweattachedtothe We recorded ground reaction forces using a custom-built in-ground o
J
regulartreadmillbeltofamodifiedexercisetreadmill(JASFitnessSystems, instrumentedtreadmill(Collinsetal.,2009)fortheevenconditionandtwo
e
TrackmasterTMX22,Dallas,TX,USA).Tocreatetheunevensurfacebelt, in-ground force platforms for the uneven condition. For the uneven h
we sewed one side of the hook-and-loop fabric on to thick, non-stretch condition,weplacedthetreadmillontopoftwosupports,eachofwhich T
717
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
restedsolelyonanin-groundforceplatform(samplerate:1000Hz;AMTI, integralsovertheentirestrideandin1%increments,todeterminewherein
Watertown,MA,USA).Toobtaintotalgroundreactionforces,weaddedthe the stride cycle mutual contraction occurred. The purpose of calculating
recorded forces from each of the force plates. For the even condition, mutualcontractionwasnottodeterminetheamountofco-activationrelative
subjects ran on one belt of the split-belt instrumented treadmill and we towhateachmusclepaircouldhavedoneduringrunning,buttotestfor
recordedonlyonesetofforces(samplerate:1000Hz).Forceplatformswere differencesinco-activationstrategiesbetweenthetwosurfaces.
re-zeroedpriortoeachtrial.Forcedatausedtodeterminethetimeofheel
contactandtocalculatetheaverageverticalgroundreactionforcesforeach Legstiffness
subjectwerelow-passfilteredat25Hz(fourth-orderButterworthfilter,zero Inordertocomparethesensitivityofourresults,wecalculatedlegstiffness
lag). However, forces synced to kinematic data and used for inverse valuesusingtwomethods.First,wereducedsubjectdynamicstothatofa
dynamicsanalysiswerelow-passfilteredat6Hz(fourth-orderButterworth spring-massmodel(McMahonandCheng,1990)anddefinedlegstiffness
filter,zerolag)becauseofthehighnoisesensitivityoftheinversedynamics tobetheratiobetweentheverticalgroundreactionforceandthechangein
calculations.WeusedVisual-3D(C-Motion,Germantown,MD,USA)to effectiveleglength:
conductinversedynamicsanalysisandtodeterminejointangles,moments
F
and powers in the sagittal plane for the stance limb. For both running k ¼ max ; ð2Þ
conditions,wedeterminedanypositionoffsetoftheforcevectorrelativeto max DL
max
therunningsurfaceandcorrectedforitslocationinVisual-3D.Fortheeven
wheretheeffectiveleglengthwasthestraight-linedistancefromthegreater
condition,thiscorrectionwasminimal.However,fortheunevencondition,
trochanter marker to the fifth metatarsal marker of the stance foot,
the ground reaction force was measured across two force plates located
normalized to subject leg length. We defined ΔL asthe difference in leg
substantiallybelowthesurfaceofthetreadmill.Tocompensateforthis,we
lengthatheel-strikeandanyotherpointduringstance,suchthatmaximum
transformedthemeasuredforcesandtorquesattheforceplatformsintoa
leglengthdeflection,ΔL ,occurrednearmid-stancewhenleglengthwas
commonreferenceframewithanoriginattheheightofthetreadmillsurface, max
shortest. In addition, we defined F as the maximum vertical ground
and added them together. This introduced a small error into the inverse max
reactionforceaftertheimpactforce.Wethencalculatedlegstiffness,k ,
dynamicscalculations,astheactualgroundreactionforcewasappliedonthe max
foreachstrideastheratiobetweenthepeakverticalgroundreactionforce
terrain,whichwasupto2.54cmhigherthanthesurfaceofthetreadmill.We
andthemaximumleglengthdeflection(GüntherandBlickhan,2002).
estimatethecenterofpressureerrorintroducedbythissimplificationtobe
Weusedanalternativemethodforcalculatinglegstiffnessasameansto
lessthan1.5cmbasedonprojectingaforcefromthehighestpointonthe
ensure that methodology did not alter the conclusions of the study. We
terraintothesurfaceofthetreadmill.Meansubjectforceswerenormalized
computedthesecondlegstiffness,k ,byfindingalinearfittothecurve
tosubjectweightandthenaveragedoversubjects. fit
producedbyplottingeffectiveleglengthagainsttheverticalgroundreaction
forceoverastride.Theslopeofthislinearfitdefinedtheapproximateleg
Electromyography
stiffness(GüntherandBlickhan,2002).Forbothcalculations,wefoundleg
For all trials, we recorded and processed EMG signals as previously
stiffnessateachstepforeverysubjectandthenaveragedthestridestofind
described(Voloshinaetal.,2013).Bipolarsurfaceelectrodes(samplerate:
subject means for both running conditions. The mean inter-subject leg
1000Hz;Biometrics,Ladysmith,VA,USA)wereplacedoverthebellyof
stiffness for each running surface was the average leg stiffness across
fourlowerlegandfourthighmuscles.Inparticular,werecordedEMGdata
subjects.
fromthetibialis anterior(TA),soleus(SO),medialgastrocnemius(MG),
lateral gastrocnemius (LG), rectus femoris (RF), vastus medialis (VM),
Dataandstatisticalanalyses
vastuslateralis(VL)andthesemitendinosusofthemedialhamstring(MH)
Wedefinedvariabilityforstepparameters,jointparameters(consistingof
muscles of the right leg only. The surface electrodes had a diameter of
jointangles,torquesandpowers)andEMGdataastheaverages.d.ofeach
1.0cm, an inter-electrode distance of 2.0cm, and an EMG amplifier
bandwidth of 20–460Hz. Only the last 2.5min of data were used for parameteracrosssubjects,perrunningtrial.Forexample,forstepparameter
data, variability was calculated by averaging the s.d. of consecutive step
analysis. EMG data were first high-pass filtered at 20Hz (fourth-order
distancesorperiodsovertimeforeachsubject,acrosssubjects.Similarly,
Butterworthfilter,zerolag)andthenfull-waverectified.Foreachsubject,
jointparameterandEMGvariabilitywerefoundbyaveragingthes.d.ofthe
wethenaveragedthedataoverstepstocreateEMGmeansforeachmuscle
parameter at each time point, per condition, across subjects. We then
and normalized these means to the maximum mean value for the two
reportedthemeanvariability(andthes.d.ofthevariabilityoversubjects)for
runningconditionstominimizeinter-subjectvariability(YangandWinter,
eachcondition.Weusedrepeated-measuresANOVAtoassessdifferences
1984). These signals were then averaged over subjects to create
betweenconditions.Thesignificancelevel,α,wassetat0.05,andposthoc
representativeEMGprofiles.Wealsofoundthes.d.oftheEMGsignalat
Holm–Sidakmultiplecomparisontestswereconductedwhereappropriate.
each time point, for each subject. These s.d. were also averaged over
y
subjects,tocreatemeans.d.envelopesforeachrunningcondition.Although g
Acknowledgements o
variabilityinmuscleactivitycannotdirectlyberelatedtochangesinenergy
expenditure,itcandemonstratetheamountofperturbationexperienceddue TheauthorsthankDrJohnRebulaforhelpfuldiscussionsandassistancewithdata ol
analysis,aswellasBryanSchlinkandmembersoftheHumanNeuromechanics i
touneventerrain.Inaddition,wequantifiedchangesinmuscleactivationby B
Laboratoryforassistanceincollectingandprocessingthedata.
averaging, over the stride time, the normalized subject EMG profiles for al
eachsubjectandcondition.Thesesubjectaveragevalueswerethenaveraged Competinginterests nt
oversubjectstoproduceonemeanvalue,foreachmuscleandcondition, Theauthorsdeclarenocompetingorfinancialinterests. e
m
indicativeofmuscleactivity.WealsousedthemeansubjectEMGprofiles
to calculate muscle mutual contraction (MC), or ‘wasted’ contraction as Authorcontributions ri
e
defined by Thoroughman and Shadmehr (1999), for three pairs of A.S.V.andD.P.F.developedtheconceptsanddesignoftheexperiments, p
antagonisticmuscles(SO/TA,MH/VMandMH/VL): interpretedtheresultsandpreparedandrevisedthearticle.A.S.V.executedthe x
E
experimentsandperformeddataanalysis.
ð f
o
MC¼ minðf1;f2Þdt; ð1Þ Funding l
a
ThisresearchwassupportedbygrantsfromtheArmyResearchLaboratory
n
such that f1 and f2 are the mean EMG profiles of the two antagonistic [MWic9h1ig1NanF-R0a9c-1k-h0a1m39GtroaDdu.Pa.tFe.,SWtu9d1en1tNFFe-l1lo0w-2s-h0ip02to2tAo.SD..VP..F.]andtheUniversityof ur
musclesandmin(f ,f )istheminimumofthetwoprofilesateachtimepoint o
1 2 J
(Voloshinaetal.,2013).Inotherwords,whenmeanEMGprofilesfromtwo
References e
musclesdonotoverlap,wecanexpectzeromutualcontraction,whereasany Alexander,R.M.(1992).Amodelofbipedallocomotiononcompliantlegs.Philos. h
overlapwouldproduceanon-zero,sharedactivitylevel.Wecomputedthe Trans.R.Soc.Lond.BBiol.Sci.338,189-198. T
718
RESEARCHARTICLE TheJournalofExperimentalBiology(2015)218,711-719doi:10.1242/jeb.106518
Arellano,C.J.andKram,R.(2011).Theeffectsofstepwidthandarmswingon preserve gait stability in response to balance perturbations. Gait Posture 36,
energeticcostandlateralbalanceduringrunning.J.Biomech.44,1291-1295. 260-264.
Arellano, C. J. and Kram, R. (2012). The energetic cost ofmaintaining lateral Hof,A.L.,Geelen,B.A.andVandenBerg,J.W.(1983).Calfmusclemoment,
balanceduringhumanrunning.J.Appl.Physiol.112,427-434. work and efficiency in level walking; role of series elasticity. J. Biomech. 16,
Bauby,C.E.andKuo,A.D.(2000).Activecontroloflateralbalanceinhuman 523-537.
walking.J.Biomech.33,1433-1440. Jensen,K.,Johansen,L.andKarkkainen,O.-P.(1999).Economyintrackrunners
Blum,Y.,Birn-Jeffery,A.,Daley,M.A.andSeyfarth,A.(2011).Doesacrouched andorienteersduringpathandterrainrunning.J.Sport.Sci.17,945-950.
legpostureenhancerunningstabilityandrobustness?J.Theor.Biol.281,97-106. Kuo,A.D.(2001).Asimplemodelofbipedalwalkingpredictsthepreferredspeed–
Brockway,J.(1987).Derivationofformulaeusedtocalculateenergyexpenditurein steplengthrelationship.J.Biomech.Eng.123,264-269.
man.Hum.Nutr.Clin.Nutr.41,463-471. Lejeune,T.,Willems,P.andHeglund,N.(1998).Mechanicsandenergeticsof
Burdet,E.,Osu,R.,Franklin,D.W.,Milner,T.E.andKawato,M.(2001).The humanlocomotiononsand.J.Exp.Biol.201,2071.
central nervous system stabilizes unstable dynamics by learning optimal Margaria,R.(1968).Positiveandnegativeworkperformancesandtheirefficiencies
impedance.Nature414,446-449. inhumanlocomotion.Int.Z.Angew.Physiol.25,339-351.
Cavagna, G. A., Saibene, F. P. and Margaria, R. (1964). Mechanical work in Margaria,R.,Cerretelli,P.,Aghemo,P.andSassi,G.(1963).Energycostof
running.J.Appl.Physiol.19,249-256. running.J.Appl.Physiol.18,367-370.
Cavanagh,P.R.(1987).Thebiomechanicsoflowerextremityactionindistance McAndrew,P.M.,Dingwell,J.B.andWilken,J.M.(2010).Walkingvariability
running.FootAnkleInt.7,197-217. duringcontinuouspseudo-randomoscillationsofthesupportsurfaceandvisual
Collins,S.H.,Adamczyk,P.G.,Ferris,D.P.andKuo,A.D.(2009).Asimple field.J.Biomech.43,1470-1475.
methodforcalibratingforceplatesandforcetreadmillsusinganinstrumented McGeer,T.(1990).Passivedynamicwalking.Int.J.Robot.Res.9,62-82.
pole.GaitPosture29,59-64. McMahon,T.A.andCheng,G.C.(1990).Themechanicsofrunning:howdoes
Daley,M.A.andBiewener,A.A.(2006).Runningoverroughterrainrevealslimb stiffnesscouplewithspeed?J.Biomech.23,65-78.
controlforintrinsicstability.Proc.Natl.Acad.Sci.USA103,15681-15686. Pinnington,H.C.andDawson,B.(2001).Theenergycostofrunningongrass
Daley,M.A.,Felix,G.andBiewener,A.A.(2007).Runningstabilityisenhancedby comparedtosoftdrybeachsand.J.Sci.Med.Sport4,416-430.
a proximo-distal gradient in joint neuromechanical control. J. Exp. Biol. 210, Pinnington,H.C.,Lloyd,D.G.,Besier,T.F.andDawson,B.(2005).Kinematic
383-394. and electromyography analysis of submaximal differences running on a firm
Donelan,J.M.,Kram,R.andArthur,D.K.(2001).Mechanicalandmetabolic surfacecomparedwithsoft,drysand.Eur.J.Appl.Physiol.94,242-253.
determinantsofthepreferredstepwidthinhumanwalking.Proc.R.Soc.BBiol. Thies,S.B.,Richardson,J.K.andAshton-Miller,J.A.(2005).Effectsofsurface
Sci.268,1985. irregularityandlightingonstepvariabilityduringgait:astudyinhealthyyoungand
Farley,C.T.andFerris,D.P.(1998).Biomechanicsofwalkingandrunning:center olderwomen.GaitPosture22,26-31.
ofmassmovementstomuscleaction.ExerciseSport.Sci.R.26,253-286. Thoroughman,K.A.andShadmehr,R.(1999).Electromyographiccorrelatesof
Franklin,D.W.,Liaw,G.,Milner,T.E.,Osu,R.,Burdet,E.andKawato,M.(2007). learninganinternalmodelofreachingmovements.J.Neurosci.19,8573-8588.
Endpointstiffnessofthearmisdirectionallytunedtoinstabilityintheenvironment. Voloshina,A.S.,Kuo,A.D.,Daley,M.A.andFerris,D.P.(2013).Biomechanics
J.Neurosci.27,7705-7716. andenergeticsofwalkingonuneventerrain.J.Exp.Biol.216,3963-3970.
Grimmer,S.,Ernst,M.,Günther,M.andBlickhan,R.(2008).Runningonuneven Weir,J.B.d.V.(1949).Newmethodsforcalculatingmetabolicratewithspecial
ground: leg adjustment to vertical steps and self-stability. J. Exp. Biol. 211, referencetoproteinmetabolism.J.Physiol.109,1-9.
2989-3000. Yang, J. and Winter, D. (1984). Electromyographic amplitude normalization
Günther,M.andBlickhan,R.(2002).Jointstiffnessoftheankleandthekneein methods: improving their sensitivityas diagnostic tools in gait analysis. Arch.
running.J.Biomech.35,1459-1474. Phys.Med.Rehabil.65,517.
Hak,L.,Houdijk,H.,Steenbrink,F.,Mert,A.,vanderWurff,P.,Beek,P.J.and Zamparo,P.,Perini,R.,Orizio,C.,Sacher,M.andFerretti,G.(1992).Theenergy
vanDieën,J.H.(2012).Speedinguporslowingdown?:Gaitadaptationsto costofwalkingorrunningonsand.Eur.J.Appl.Physiol.O.65,183-187.
y
g
o
l
o
i
B
l
a
t
n
e
m
i
r
e
p
x
E
f
o
l
a
n
r
u
o
J
e
h
T
719