Development of human locomotion

Development

of

human

locomotion

Francesco

Lacquaniti

,

Yuri

P

Ivanenko

and

Myrka

Zago

Neuralcontroloflocomotioninhumanadultsinvolvesthe generationofasmallsetofbasicpatternedcommands directedtothelegmuscles.Thecommandsaregenerated sequentiallyintimeduringeachstepbyneuralnetworks locatedinthespinalcord,calledCentralPattern Generators.Thisreviewoutlinesrecentadvancesin understandinghowmotorcommandsareexpressedat differentstagesofhumandevelopment.Similarcommandsare foundinseveralothervertebrates,indicatingthatlocomotion developmentfollowscommonprinciplesoforganizationof thecontrolnetworks.Movementsshowahighdegreeof flexibilityatallstagesofdevelopment,whichisinstrumentalfor learningandexplorationofvariableinteractionswiththe environment.

Addresses

1_{DepartmentofSystemsMedicine,NeuroscienceSection,Universityof}

RomeTorVergata,ViaMontpellier1,00133Rome,Italy

2_{CentreofSpaceBio-medicine,UniversityofRomeTorVergata,Via}

OrazioRaimondo1,00173Rome,Italy

3

LaboratoryofNeuromotorPhysiology,IRCCSSantaLuciaFoundation, ViaArdeatina306,00179Rome,Italy

Correspondingauthor:Lacquaniti,Francesco(lacquaniti@caspur.it)

CurrentOpinioninNeurobiology2012,22:822–828

ThisreviewcomesfromathemedissueonNeurodevelopment anddisease

EditedbyJosephGleesonandFranckPolleux

ForacompleteoverviewseetheIssueandtheEditorial

Availableonline10thApril2012

0959-4388/$–seefrontmatter,#2012ElsevierLtd.Allrights reserved.

http://dx.doi.org/10.1016/j.conb.2012.03.012

Introduction

Human neonates exhibittransitory, primitivebehaviors that developin utero and disappear a few months after birth [1,2]. Some of these are critical for survival; for instance, rooting and sucking reflexes are essential for feeding.Butthesignificanceofotherprimitivebehaviors andtheirrelationshipwithmore maturebehaviorsarea long-standingriddle [3,4,5].Traditionally, itisthought thatprimitivebehaviorsaresuppressedasaresultofbrain maturation.However, there is now evidence that some basiccontrol principles are conservedthrough develop-ment,sothattheprimitivepatternscanbeconsideredas precursorsofthematurepatterns.

Locomotorbehaviorisacaseinpoint.Humannewborns exhibitasteppingreflexthattypically disappearsat2 months and reappears several months later when it

evolves into intentional walking. It was once thought thatthepatternsof musclecontrolinnewbornstepping arediscardedduring development,replaced byentirely new patterns of walking. Instead, it has recently been shown thattheprimitive steppingpatterns are retained andtuned, whilenew patterns are addedduring devel-opment [6]. Surprisingly similar patterns are observed also in several other animal species, suggesting that locomotionisbuiltstartingfromcommonelements, per-hapsrelatedtoancestralneuralnetworks[7].

Herewefirstreviewrecentfindingsontheprenataland postnataldevelopmentofmotor patternsinhuman chil-dren, and on a comparative analysis with other animal species. Next we consider the role of learning and explorationinhumanlocomotordevelopment.Inafinal section, we deal with abnormalities of motor develop-mentastypifiedbycerebralpalsy.

Prenatal

movements

Spontaneous movements begin as soon as there are functioning muscles and nervous system in developing humans and animals. In humans, small, slow, cyclic bending of the head and/or trunk are detected with 4D-ultrasonographyat5weekspost-conception[8]. Wax-ing and waning general movements can be observed slightly later, at 7 weeks, and persist throughout preg-nancyandthefirstmonthsaftertermbirth[9–11].They consist of complex, variable, flexion-extensions of the wholebodyandlimbs,theyarenottriggeredbyexternal stimuliandlackdistinctivesequencingofdifferentbody parts.Inaddition,humanfetusesexhibitarichrepertoire of leg movements that includes single leg kicks, sym-metrical double legs kicks, and symmetrical inter-limb alternation with variable phase [12,13]. Spontaneous movements of the limbs evolve toward an increased coordination between the arms and between the legs, at2–4monthsafterbirth[14].Abnormalmovementslack complexity, variation, and fluency, and are associated withanincreasedprobabilityofcerebralpalsy[10,15]. Whileonlykinematicanalysesarecurrentlyavailablefor humanfetuses[16],directrecordingsofelectricalmuscle activity(EMG)arepossibleinanimals.EMGreflectsthe outputofspinala-motoneurons,andthereforetheneural commandsfor movement.DetailedEMGrecordings in chick embryos during the final week of incubation showedthattheprofilesof EMGactivityduring repeti-tive limb movements resemble those of locomotion at hatching [17]. However, in contrast with mature loco-motoractivity, EMGburstduration doesnotscalewith movementcycledurationin chickembryos.

Opticalimagingofspontaneousactivityinventralspinal neurons of the zebrafish embryo showed a rapid (few hours)transitionfromuncorrelated,sporadicslowactivity to ipsilaterally correlated and contralaterally anticorre-lated fast activity involving several adjacent somites [18]. The transition to correlated activity may depend on electrical connections initially coupling nearby neurons in local microcircuits and then merging to in-clude the majority of active ipsilateral neurons into a single coupled network [18]. Recurrently connected excitatorynetworkswithinthespinalcordaretransiently silencedbyactivity-dependentdepression[19].Inthese networks, motoneurons generate large, slow depolariz-ations crestedbyburstsof actionpotentials,resultingin the correlated discharge across a population of neurons withaperiodicityintheorderofminutes,afiringpattern that drives spontaneousembryonic movements[20,21]. Thus, theepisodesof spontaneousactivityare presum-ably triggered by motoneurons, but the periodicity of activityis setbyrecurrentexcitatoryinteractionsin the network[21].Burstingactivityoccurswhilemotoneurons arestillmigratingandprolongingtheiraxonstowardthe baseofthelimbs,sothatcorrectmotoraxonpath-finding iscontingentonnormalburstingactivity[22].Inaddition, spontaneous motor activity at an early developmental stagemayfacilitatetheself-organizationofneuralcircuits atboth spinaland supra-spinal levels[21].Thus,motor activity modulates the spinal circuits of central pattern generators (CPG) and those of nociceptive withdrawal reflexes[23],anditalsomodulatescorticalsomatosensory maps in a somatotopic manner [24]. Once established, spinal CPGs underlie fetal movements[22], but devel-oping supra-spinal structures (suchas thetransient cor-tical subplate) presumably also play a role in more complex sequences of general movements, as demon-strated by the abnormality of general movements in humanfetuseswithbrain disorders[10].

Postnatal

development

of

locomotion

Inadditiontothespontaneousgeneralmovements,human newbornsalsoexhibitsteppingmovements.Thesecanbe elicitedinaninfantsupportedunderthearmsinanupright, slightlytiltedforwardposture,aftercontactinggroundwith thefeetsoles[6,25].Reflexsteppinghasbeenreported alsoinprematureinfantsat30+post-conceptionweeks[26] and anencephalicnewborns[27].Thissuggests a predo-minantroleofspinalandbrainstemmechanisms,owingto immature cerebral connections to the spinal cord [28]. Whilebuoyancyoftheamnioticfluidcounteractsgravity effectsinthehumanfetus(andotheramniotes),newborns must deal with gravityto movetheirlimbs andsupport their body weight. They cansupport 30–40% of their weight,theremainingbeingsupportedfromtheoutside. Infant stepping is irregular, variable, and lacks several features of more mature walking, most notably postural control[6,25,29–31].Itlookslikeanexaggerated march-ingwithflexedlegs,highfootlift,andflat-foottouch-down

(insteadofheel-contactasinadults).Groundforcesarenot accurately controlled: newborns typically exert vertical forcessupportingpartoftheirweight,butonlynegligible horizontal(shear)forces.However,theangularmotionof thelowerlimbssegmentsisalreadycoordinated,resulting inaplanarinter-segmentalcovarianceroughlycomparable tothatseeninadults[6].

Uprightsteppingoftenbecomesverydifficulttoelicitin infants between2 and 7 months of age, whilesupine kicking (which shares some features of stepping) con-tinues [29].The disappearanceof stepping presumably dependsonbothneuralchangesintheCNS(supraspinal inhibitoryinfluencesonspinalCPGs)andbiomechanical factors (thelegs maybecometoo heavy for the muscle strength [29]). However, stepping can still be evoked during thatperiodwith dailypractice[29,30,32] or sup-porting the legs’ weight by means of water immersion [29].Practiceincreasestheincidenceofalternatingsteps, but doesnotappreciably affectthemuscleactivity pro-files [30]. Infant stepping shows sensory adaptation to imposed loading and perturbations [33]. Longitudinal studiescarriedoutoverthefirstyearshowaprogressive reduction of muscle co-contraction and increasingly se-lective activationpatterns [31,34]. However,there is a persistentlylargeinter-stepvariabilityofEMGactivities compared with more repeatable kinematics. The latter findingisreminiscentofadultlocomotion, anddepends on a prominent role of passive biomechanics in loco-motionin bothinfantsandadults[35,36].

Theneuralpatternsofmusclecontrolcanberevealedby factorization ofEMGactivity[36].Inhumannewborns, two patterns are sinusoidally modulated over the step cycle:onepattern(#2inFigure1)helpsprovidingbody support during stance, while the other one (#4) helps driving the limb during swing, but there is no specific activation pattern at either touch-down or lift-off [6]. EMGfactorizationintoddlers(1-year-old)attheirfirst unsupportedstepsfindsthesametwopatterns(#2and#4 inFigure1)ofthenewborn,plustwonewpatternstimed attouch-down(#1)andlift-off(#3)thatcontributeshear forces necessaryto decelerateand accelerate thebody, respectively [6]. In preschoolers (2–4-years), all four patterns show transitional shapes. Thus, the waveform shiftsintimerelativetothestepcycleprogressivelywith age(seelowerpanelinFigure1):theolderthechild,the closerthewaveformtotheadult.Insum,twobasiccontrol patternsareretainedfromthestageofnewbornstepping, whiletwootherpatternsdevelopafterthatstage. Transi-tionalpatternsindicateacontinuousdevelopmentofthe correspondingmotorcontrolmodules.

Thegradualdevelopmentofadultgaitfrominfant step-pingisgenerallybelievedtostemfromagrowing integ-ration of supraspinal, intraspinal and sensory control [3,30]. The lackof activation patterns correspondingto

footcontact in theneonate coulddepend on immature sensory and/or descending modulation of stepping. Indeed,intheabsenceofsensorymodulation(e.g.during fictivemotortasks),thespinalcircuitryofanimalstendsto producesinusoidal-likepatterns[37,38],similartothose observed in the human neonate. The addition of basic patterns in thefirst months of life impliesa functional reorganizationofinter-neuronalconnectivity,the appear-anceofadditional functionallayersintheCPGs,and/or more powerful descending and sensory influences on CPGs. In particular, there is increasing consensus that motorcentersinthebrainplayanimportantandgreater role in human adult walking than in quadrupeds [39]. Indeed,thereisevidenceformaturationofcortico-spinal driveonlegmusclesduringlocomotordevelopment[40].

Comparative

aspects

Inpostembryonictadpoles,motoneuronsinitially innerv-ate most of the dorso-ventral extent of the swimming muscles,butduringearlylarvallifetheinnervationfields becomerestrictedtoalimitedsectorofeachmuscleblock [41].Thisdevelopmentaltrendleads tomoreselective andflexiblecontrolofthemuscles.Justashumans,ratsdo nothaveamatureneuralcontroloflocomotionatbirth, and they walk only several days later. The CPGs of neonatalratspinalcordareintrinsicallyflexible,inasmuch as different patterns of hindlimb muscle activation are evoked dependingonwhether pharmacological (seroto-ninandN-methyl-D-asparticacid)modulationorsensory

afferent stimulationis applied [42,43]. Locomotor-like

oscillatoryactivitycanberecordedfromthelumbarand sacralventralrootsoftheisolatedspinalcordofneonatal rats,bathedwithdopamineplusNMDAorserotonin[37]. Factorization of the electroneurograms associated with thisfictivewalkingrevealstwopatternsessentially iden-ticaltothoseofhumannewborns[6].Factorizationofthe EMG of adult rats, cats, macaques, and guinea fowls shows four patterns, closely resembling those found in humantoddlers[6].

Theseresultsare consistentwithcomparativestudies in vertebrates based on genetic and electrophysiological approacheswhichdemonstratethat,despitetheexistence of species-specific features, there are several common principles in the organization and regulation of CPGs [44,45]. In particular, the core premotor components of locomotorcircuitrymainlyderivefromasetofembryonic interneuronsthatareremarkablyconservedacrossdifferent species [46]. Grillner [7] hypothesizes that the neural controlsystem for locomotioncan betracedbackto the oldest known vertebrate, the lamprey, which appeared morethan500millionyearsago,beforeanyleggedanimal had evolved yet. Evolutionary conservation of develop-mentalpatterns[6]andneuralcorecontrolnetworks[44– 46,47]pointstothecomparativeapproachasamostfruitful oneforthestudyoflocomotordevelopment[47,48].

Human development shows commonalities with other animalspecies,butalsoimportantidiosyncraticfeatures, as demonstrated by the distinct motor patterns of the Figure1

Neonates Toddlers Preschoolers Adults

0 % cycle 100 Pattern 1 0 1 Pattern 2 Pattern 3 Pattern 4 42 mo 25 33

Current Opinion in Neurobiology

Basicpatternsofmusclecontrolatdifferentages.Top.Stickdiagramsdepictingonestepcyclestartingwithstanceonsetinneonates(2–7-days-old), toddlers(11–14-months-old),preschoolers(24–48-months),andadults.Middle.Basicactivationpatternsobtainedbynon-negativematrix factorizationofaveragedEMGprofilesof24bilaterallegmusclesineachagegroup.Patternsareplottedversusnormalizedgaitcycle(alignedwith stanceonsetintherightleg).Bottom.Pattern#4wasaveragedseparatelyin3differentsubgroupsofpreschoolerswiththeindicatedmeanage. Noticetheshiftofthewaveformwithincreasingage.

adults[6].Thus,wearetheonlyanimalstouse habitu-allyan erectbipedal locomotionwitha heel-strikewell ahead of the body. The long time required to develop independentlocomotioninhumansisprobablyrelatedto the overall complexity of neural wiring in our species. Consistentwiththisview,ithasbeenshown[49]thatthe time fromconceptionto independentlocomotionis lin-early relatedto theadult brain massacross 24different mammalianspecies: thebiggerthebrain,thelongerthe timetostartwalking.Thissuggeststhatthedevelopment of independent locomotiondepends ontheduration of overallneural development,presumablybecauseof the needtodevelopstance,balanceandorientationcontrolin parallel with locomotor control; these diverse functions requirematurationof largepartsof theCNS.

Learning

and

exploration

Motor patterns of locomotion are not fixed but highly flexible. Variability and versatility of behavior may be instrumental for learning and exploration of different solutions indifferentenvironmentalcontexts[3,5,50].

Infantmovementsdisplayahighdegreeofvariabilityat allstagesofdevelopment,startingfromfetalstages.After birth,steppingremainsnon-functionaluntilastableerect posturecanbemaintained,andinfantsadoptavarietyof different crawling styles to move around, although a significant proportion (30%) never crawls and walks upright directly [5,51]. Infants can crawl on hands-and-knees, hands-and-feet, hands-and-buttocks, or on the belly. Inter-limb coupling may involve diagonal trot-likegait,oripsilateralpace-likegait.Thisversatility reflectsaflexible couplingbetweencervicaland lumbo-sacralCPGs(controlling upperandlowerlimbs respect-ively)thatpersiststilladulthood[52].

Alsouprightwalkingbeforeindependentwalkingshows great flexibility.Infants mayfirst cruise sideways while graspingfurniturewithbothhandsforsupport,thenturn their bodyto face forwards holding furniture with one hand only [5,53]. Different developmentalstages (for instance,crawlingandcruising)maybeconcurrentrather than serial, and there may even be a reversal of order (cruising before crawling). Moreover, often there is no transfer of learning environmental risks from one loco-motormodetothenext:experiencedcrawlersorcruisers discriminate very precisely affordable versus unafford-ablesupportsurfaces,butwhentheystartwalking inde-pendentlytheymayfallbecausetheydonotdiscriminate anymore[53].

Infants start walking independently around 12-months (median, 9–18months range),but cultural child-rearing habitsmayanticipateordelaythistime[5].Unsupported walkingisjerkyandvariable,withpoorbalanceoverthe singlesupportleg(whileswingingthecontralateralleg), the armsraised above thewaist (as balancepoles), legs

splayedwideapart,andshortvariablesteps[5,25,54–59]. Double support is relatively prolonged, while swing is brief. Touch-down is with flat-foot or toes-first. Some idiosyncratic features of toddlers gait maybe useful to cope with initial unstable conditions of unsupported walking, suchas theincreased base of support and the flailing arms[58].Also,non-plantigradegait withahigh footlift represents asimplestrategy toavoid stumbling and falling, while reducing foot drag owing to limited dorsiflexoractivity.However,energyrecoveryby exchan-ging forward kinetic energy and gravitational potential energyof thecenterofbodymassisvery limited[55]. It is often assumed that infants cannot walk indepen-dentlyuntiltheyachievebalancecontrol,butithasbeen shown that step variability and several other gait parameters of toddlers remain unchanged even when balanceisaugmentedwiththehelpofaparentor exper-imenterhand[56].Moreover,walkingexperiencerather thanchronologicalageexplainsimprovementin perform-ance[5].Indeed,onsetofunsupportedlocomotion trig-gerstheimprovementofseveralgaitparameters(speed, inter-steprepeatability,trunkoscillations,tuningof pla-narcovariance,energyrecovery)relativetotheprevious supported locomotion [55,58]. These changes occur rapidly over the first6 months after the onset of inde-pendent walking.Afterwards,gait continues to develop moreslowlyuntil8–10yearsofage,asshownbychanges in several parameters, such as stride length, cadence, coordinationtiming, andenergyrecovery[5,54,59].

Whentoddlersmuststepacrossanobstacleorwalkona staircase, they donotadapt theinter-segmental coordi-nationtothesurfaceinclinationandheightasadultsdo, buttheykeepconstantphaserelationships[60].Thisis consistentwiththehypothesisproposeddecadesagoby Nikolai Bernstein that, when humans start learning a skill, they restrict the number of controlled degrees of freedom to reduce the size of the search space and simplify the coordination. Toddlers often place a foot ontheobstacleorontheedgesofthestairs,presumablyas part of an exploratory strategy of the environment [5,60,61].Naturalisticobservationsattheinfants’home showthatmosttoddlersspontaneouslycarryobjectswhile walking,combininglocomotorandmanualskills.Despite the additional biomechanical constraints, carrying an object isactually associated withimproved upright bal-ance,as demonstratedbysmallerprobabilitiesof falling withtheobjectthan without[62].

Split-belts treadmills can impose a different direction and/orspeed to themotionundereach legduring loco-motion.They are especiallysuitedfor studying sensor-imotor adaptation and learning mechanisms. Young infants (7–12 months of age) show the ability to adapt to asynchronous split-belt motion [63]. However, the mechanisms controllingtemporal and spatialadaptation

totheseconditionsaredifferentandmatureatdifferent times,withspatialparametersadaptingmoreslowlythan temporalones[64,65].

Body size and proportions change dramatically during development. Locomotor commands must take these changesintoaccounttokeeplimbsegmentmotion cali-bratedwithbodysize.Theimportanceofabodyscheme incorporatinglimbandbodyparametersisdemonstrated bytheobservationthatan11-years-oldchild,who under-wentsurgicalelongationoftheshanksby>50%,walked as if on the pre-surgery shorter legs, just as do adults walkingonstilts[66].

Cerebral

palsy

Cerebralpalsy (CP)is one ofthemost common devel-opmental motor disorders. It is a non-progressive syn-dromeinvolvingpoormotorcontrol,spasticity,paralysis, andotherneurologicalproblemsresultingfromperinatal brain injury [67]. It may be hypothesized that neural control patterns in CP children are closer to those of younger, normally developing children, and this would reflect relative immaturity of the locomotor networks. ManyCPchildrenstartwalkingmuchlaterthannormal. Tilladulthood,theycontinuewalkingontheirtoeswith kneehyper-flexionduringstanceandankledorsiflexion duringswing[68].Thefoottrajectoryisundulatingowing to poor control ofankletorque. Muscle co-activationis greater than in healthy children of the same age. Hip flexors lackphasicactivity,hipextensorsand adductors are hyper-active, while gastrocnemius is hypoactive at push-off[68].Thenormaltonicdepressionofsoleus H-reflex during gait is absent in CP, reflecting a lack of maturationofthecorticospinaltract[69].Reflexbehavior andwalkingspeedimproveswithtreadmilltraining[68], although the long-term effectiveness of this protocol remainstobevalidated [70]alsousingenergy expendi-turemonitoring[71].

Conclusions

We argued that the neural control patterns underlying maturelocomotionaretightlyrelatedtothoseinvolvedin primitivemovements.Inaddition,developmentofmotor patterns shows variability and versatility of behavior presumably as a means to learn and explore different solutions. Notice that this holds truenot only for loco-motion,but evenforbehaviors–suchasvocalization in birds – where mature neural substrates are definitely distinctfromtheimmatureones[50].Wealsohighlighted theremarkablesimilaritiesinmotorpatternsacross differ-ent animal species, despite gross morpho-functional differencesin themusculo-skeletal architecture. These similarities probably reflect common principles in the underlyingcontrolmechanisms,as wellascommon bio-mechanical constraints related to stability, kinematics, kinetics,andenergy-efficiency[72,73].Anemergingview is thatthe co-ordination of limb and body segments in

mature locomotion arises from the coupling of neural oscillatorsbetween eachotherand withlimb mechan-ical oscillators [35,36]. Muscle activations intervene at discrete times to re-excite the intrinsic oscillations of thesystemwhenenergyislost.Developmentofmotor patterns,then,requiresprogressivelytuningthetiming andamplitudeofmuscleactivitytotheintrinsicmodes ofmechanicalbehaviorresultingfromtheinteractionof thelimbs/body partsbetweeneach otherandwiththe environment,alsotakingintoaccountthegrowingbody ofthe child.

Thecurrentcoarsepictureofthedevelopmentofhuman locomotor patterns needs now to berefined in order to understand howthelocomotor networks are configured precisely at different developmental stages. Moreover, the functional abnormalities associated with perinatal motordisorderssuchas CPneedto beunderstood,also takingadvantageoftheapplicationofmodern quantitat-iveanalysesofthemotor patterns.

Acknowledgements

OurworkwassupportedbytheItalianMinistryofHealth,ItalianMinistry ofUniversityandResearch(PRINgrant),ItalianSpaceAgency(DCMC andCRUSOEgrants),andEuropeanUnionFP7-ICTprogram(AMARSi grant#248311).Fundingagencieshadnoinvolvementintheconductofour research,inthewritingofthereport,andinthedecisiontosubmitthe articleforpublication.

References

and

recommended

reading

Papersofparticularinterest,publishedwithintheperiodofreview, havebeenhighlightedas:

 ofspecialinterest ofoutstandinginterest

1. Milani-ComparettiA,GidoniEA:Routinedevelopmental examinationinnormalandretardedchildren.DevMedChild Neurol1967,9:631-638.

2. ZafeiriouDI:Primitivereflexesandposturalreactionsinthe neurodevelopmentalexamination.PediatrNeurol2004,31:1-8.

3. ForssbergH:Neuralcontrolofhumanmotordevelopment.Curr OpinNeurobiol1999,9:676-682.

4. KonczakJ:Onthenotionofmotorprimitivesinhumansand robots.In In ProceedingsoftheFifthInternationalWorkshopon EpigeneticRobotics:ModelingCognitiveDevelopmentinRobotic Systems.EditedbyBerthouzeLetal.: ProceedingsoftheFifth InternationalWorkshoponEpigeneticRobotics:Modeling CognitiveDevelopmentinRoboticSystemsLundUniversity CognitiveStudies;2005:47-53.

5. 

AdolphKE,RobinsonSR:Theroadtowalking:Whatlearningto walktellsusaboutdevelopment.InOxfordhandbookof developmentalpsychology.EditedbyZelazoP.NY:Oxford UniversityPress;2012,inpress.

Inthisenjoyablereview,theauthorsargueforflexibilityandversatilityas commondevelopmentalthreadsprovidingplentyinsightfulobservations fromdevelopmentalpsychology.

6. 

DominiciN,IvanenkoYP,CappelliniG,d’AvellaA,Mondı` V, CiccheseM,FabianoA,SileiT,DiPaoloA,GianniniCetal.: Locomotorprimitivesinnewbornbabiesandtheir development.Science2011,334:997-999.

Thispapercomparestheneuralpatternsofmusclecontrolinhuman neonates,toddlers,preschoolers,adults,andseveralanimalspecies.It showsthatbothhumanandratnewbornshavetwophasesofactivation, out-of-phasebetweeneachother.Humantoddlersandadultrats,cats, macaquesandguineafowlshavethesameactivationpatternsas neo-nates,andinadditiontwoothershorteractivationsattouchdownand

liftoff.Thefouractivationpatterns inhuman adultsbecomepulsatile. Thesefindingssuggestthatlocomotioninhumansandotherspeciesis built starting fromcommon elements,andadding fewnewelements duringdevelopment.

7. GrillnerS:Neuroscience.Humanlocomotorcircuitsconform. Science2011,334:912-913.

8. FeltRH,MulderEJ,Lu¨chingerAB,vanKanCM,TaverneMA, deVriesJI:Spontaneouscyclicembryonicmovementsin humansandguineapigs.DevNeurobiol2011http://dx.doi.org/ 10.1002/dneu.20945.

9. deVriesJI,VisserGH,PrechtlHF:Theemergenceoffetal behaviour.I.Qualitativeaspects.EarlyHumDev1982, 7:301-322.

10. Hadders-AlgraM:Putativeneuralsubstrateofnormaland abnormalgeneralmovements.NeurosciBiobehavRev2007, 31:1181-1190.

11. Lu¨chingerAB,Hadders-AlgraM,vanKanCM,deVriesJI:Fetal onsetofgeneralmovements.PediatrRes2008,63:191-195.

12. PiontelliA:DevelopmentofNormalFetalMovements:TheFirst25 WeeksofGestation. Italy:Springer-Verlag;2010.

13. StanojevicM,KurjakA,Salihagic´-Kadic´ A,VasiljO,MiskovicB, ShaddadAN,AhmedB,Tomasovic´ S:Neurobehavioral continuityfromfetustoneonate.JPerinatMed2011, 39:171-177.

14. 

KanemaruN,WatanabeH,TagaG:Increasingselectivityof interlimbcoordinationduringspontaneousmovementsin 2-to4-month-oldinfants.ExpBrainRes2012,218:49-61. Longitudinal examination ofspontaneous movementsin 2–4-months infants showedan age-related increase in the velocity and position correlationbetweenthearmsandbetweenthelegs.Thiscorresponds toaprogressionfromageneralactivityinvolvingallthelimbstoanactivity involvingmoreselectiveinterlimbcoordination.

15. HamerEG,BosAF,Hadders-AlgraM:Assessmentofspecific characteristicsofabnormalgeneralmovements:doesit enhancethepredictionofcerebralpalsy? DevMedChild Neurol2011,53:751-756.

16. CastielloU,BecchioC,ZoiaS,NeliniC,SartoriL,BlasonL, D’OttavioG,BulgheroniM,GalleseV:Wiredtobesocial:the ontogenyofhumaninteraction.PLoSONE2010,5:e13199.

17. RyuYU,BradleyNS:Precociouslocomotorbehaviorbeginsin theegg:developmentoflegmusclepatternsforsteppingin thechick.PLoSONE2009,4:e6111.

18. WarpE,AgarwalG,WyartC,FriedmannD,OldfieldCS,ConnerA, DelBeneF,ArrenbergAB,BaierH,IsacoffEY:Emergenceof patternedactivityinthedevelopingzebrafishspinalcord. CurrBiol2012,22:93-102.

19. O’DonovanMJ:Theoriginofspontaneousactivityin developingnetworksofthevertebratenervoussystem. CurrOpinNeurobiol1999,9:94-104.

20. FellerMB:Spontaneouscorrelatedactivityindeveloping neuralcircuits.Neuron1999,22:653-656.

21. BlankenshipAG,FellerMB:Mechanismsunderlying

spontaneouspatternedactivityindevelopingneuralcircuits. NatRevNeurosci2010,11:18-29.

22. HansonMG,MilnerLD,LandmesserLT:Spontaneousrhythmic activityinearlychickspinalcordinfluencesdistinctmotor axonpathfindingdecisions.BrainResRev2008,57:77-85.

23. PeterssonP,Waldenstro¨mA,Fa˚hraeusC,SchouenborgJ: Spontaneousmuscletwitchesduringsleepguidespinal self-organization.Nature2003,424:72-75.

24. KhazipovR,SirotaA,LeinekugelX,HolmesGL,Ben-AriY, Buzsa´kiG:Earlymotoractivitydrivesspindleburstsinthe developingsomatosensorycortex.Nature2004,

432:758-761.

25. ForssbergH:Ontogenyofhumanlocomotorcontrol.I.Infant stepping,supportedlocomotionandtransitiontoindependent locomotion.ExpBrainRes1985,57:480-493.

26. AllenMC,CaputeAJ:Theevolutionofprimitivereflexesin extremelyprematureinfants.PediatrRes1986,20:1284-1289.

27. PeiperA:CerebralFunctioninInfancyandChildhood.NewYork: ConsultantsBureau;1961.

28. MartinJH:Thecorticospinalsystem:fromdevelopmentto motorcontrol.Neuroscientist2005,11:161-173.

29. ThelenE,CookeDW:Relationshipbetweennewbornstepping andlaterwalking:anewinterpretation.DevMedChildNeurol 1987,29:380-393.

30. YangJF,StephensMJ,VishramR:Infantstepping:amethodto studythesensorycontrolofhumanwalking.JPhysiol1998, 507:927-937.

31. OkamotoT,OkamotoK:DevelopmentofGaitby Electromyography.WalkingDevelopmentGroup.2007.

32. ZelazoPR,ZelazoNA,KolbS:‘‘Walking’’inthenewborn. Science1972,176:314-315.

33. PangMYC,LamT,YangJF:Infantsadapttheirsteppingto repeatedtrip-inducingstimuli.JNeurophysiol2003, 90:2731-2740.

34. 

TeulierC,SansomJK,MuraszkoK,UlrichBD:Longitudinal changesinmuscleactivityduringinfants’treadmillstepping. JNeurophysiol2012,inpress.

This paper presents a longitudinal comparison of muscle activation profilesinchildrensteppingonatreadmillat1, 6and12months of age.Althoughagonist-antagonistmuscleco-activationdecreasedwith age,inter-stepvariabilityremainedhighthroughoutthefollow-upstudy.

35. LacquanitiF,GrassoR,ZagoM:Motorpatternsinwalking. NewsPhysiolSci1999,14:168-174.

36. LacquanitiF,IvanenkoYP,ZagoM:Patternedcontrolofhuman locomotion.JPhysiol2012,http://dx.doi.org/10.1113/ jphysiol.2011.215137.

37. FalgairolleM,CazaletsJR:Metachronalcouplingbetween spinalneuronalnetworksduringlocomotoractivityin newbornrat.JPhysiol2007,580:87-102.

38. CuellarCA,TapiaJA,Jua´rezV,QuevedoJ,LinaresP,Martı´nezL, ManjarrezE:Propagationofsinusoidalelectricalwavesalong thespinalcordduringafictivemotortask.JNeurosci2009, 29:798-810.

39. NielsenJB:Howwewalk:centralcontrolofmuscleactivity duringhumanwalking.Neuroscientist2003,9:195-204.

40. 

PetersenTH,Kliim-DueM,FarmerSF,NielsenJB:Childhood developmentofcommondrivetoahumanlegmuscleduring ankledorsiflexionandgait.JPhysiol2010,588:4387-4400. Thispaperassessesthecortico-spinaldrivein4–15-years-oldchildren usingEMGcoherenceandsynchrony.Asignificantincreaseofcoherence withincreasingagewasfoundinthe30–45Hzfrequencyband(gamma) duringwalkingandduringstaticankledorsiflexion.Thechanges con-tinuedtilladolescence,presumablyindicatingthatmaturationof cortico-spinalcontrolisnotcompletetillthatage.

41. 

ZhangHY,IssbernerJ,SillarKT:Developmentofaspinal locomotorrheostat.ProcNatlAcadSciUSA2011, 108:11674-11679.

Theauthorsshowthatswimmotorneuronsoffrogtadpolesinitiallyforma homogenouspooldischargingisolatedactionpotentialsandinnervating diffuselythedorsoventralextentoftheswimmingmuscles.Subsequently, theinnervationfieldsbecomerestrictedtoalimitedsectorofeachmuscle block.

42. 

KleinDA,PatinoA,TreschMC:Flexibilityofmotorpattern generationacrossstimulationconditionsbytheneonatalrat spinalcord.JNeurophysiol2010,103:1580-1590.

Thisstudyusestheinvitroneonatalratspinalcordwithattachedhindlimb toexaminethemuscleactivationpatternsevokedbydifferenttypesof stimulation. They found that therectus femoris and semitendinosus muscleswereactivateddifferentiallydepending onwhether5-HTand NMDAorelectricalstimulationofthecaudaequinawereapplied.

43. KleinDA,TreschMC:Specificityofintramuscularactivation duringrhythmsproducedbyspinalpatterningsystemsinthe invitroneonatalratwithhindlimbattachedpreparation. JNeurophysiol2010,104:2158-2168.

44. Garcia-CampmanyL,StamFJ,GouldingM:Fromcircuitsto behaviour:motornetworksinvertebrates.CurrOpinNeurobiol 2010,20:116-125.

45. KiehnO:Developmentandfunctionalorganizationofspinal locomotorcircuits.CurrOpinNeurobiol2011,21:100-109. 46. GouldingM:Circuitscontrollingvertebratelocomotion:

movinginanewdirection.NatRevNeurosci2009,10:507-518.

47. 

Stephenson-JonesM,SamuelssonE,EricssonJ,RobertsonB, GrillnerS:Evolutionaryconservationofthebasalgangliaasa commonvertebratemechanismforactionselection.CurrBiol 2011,21:1081-1091.

Thispapershowsthatthebasalgangliacircuitryispresentinthelamprey, thephylogeneticallyoldestvertebrate,andhasbeenconserved asa mechanismforactionselectionusedbyallvertebratesfor>560million years. All sectors of the mammalian basal ganglia (striatum,globus pallidus,andsubthalamicnucleus)arepresentinthelampreyforebrain. Moreover, the circuit features, molecular markers, and physiological activitypatternsareconserved.

48. ClancyB,FinlayBL,DarlingtonRB,AnandKJ:Extrapolating braindevelopmentfromexperimentalspeciestohumans. Neurotoxicology2007,28:931-937.

49. GarwiczM,ChristenssonM,PsouniE:Aunifyingmodelfor timingofwalkingonsetinhumansandothermammals. ProcNatlAcadSciUSA2009,106:21889-21893.

50. AronovD,AndalmanAS,FeeMS:Aspecializedforebraincircuit forvocalbabblinginthejuvenilesongbird.Science2008, 320:630-634.

51. PatrickSK,NoahJA,YangJF:Interlimbcoordinationinhuman crawlingrevealssimilaritiesindevelopmentandneural controlwithquadrupeds.JNeurophysiol2009,101:603-613.

52. MaclellanMJ,IvanenkoYP,CappelliniG,SylosLabiniF, LacquanitiF:Featuresofhand-footcrawlingbehaviorinhuman adults.JNeurophysiol2012,107:114-125.

53. 

AdolphKE,BergerSE,LeoAJ:Developmentalcontinuity? Crawling,cruising,andwalking.DevSci2011,14:306-318. Beforewalking,mostinfantscruisebygraspingstableobjectssuchas furniture.Cruisinginfantsperceiveaffordancesforlocomotionoveran adjustablegapinahandrailusedformanualsupport.However,when theystartwalkingindependently,despiteweeksofcruisingexperience, theyaresurprisinglyoblivioustothedangersofgapsinthehandrail.

54. CheronG,BouillotE,DanB,BengoetxeaA,DrayeJP, LacquanitiF:Developmentofakinematiccoordinationpattern intoddlerlocomotion:planarcovariation.ExpBrainRes2001, 137:455-466.

55. IvanenkoYP,DominiciN,CappelliniG,DanB,CheronG, LacquanitiF:Developmentofpendulummechanismand kinematiccoordinationfromthefirstunsupportedstepsin toddlers.JExpBiol2004,207:3797-3810.

56. IvanenkoYP,DominiciN,CappelliniG,LacquanitiF:Kinematics innewlywalkingtoddlersdoesnotdependuponpostural stability.JNeurophysiol2005,94:754-763.

57. DominiciN,IvanenkoYP,LacquanitiF:Controloffoottrajectory inwalkingtoddlers:adaptationtoloadchanges.JNeurophysiol 2007,97:2790-2801.

58. IvanenkoYP,DominiciN,LacquanitiF:Developmentof independentwalkingintoddlers.ExercSportSciRev2007, 35:67-73.

59. SutherlandDH,OlshenR,CooperL,WooSL:Thedevelopment ofmaturegait.JBoneJointSurgAm1980,62:336-353.

60. 

DominiciN,IvanenkoYP,CappelliniG,ZampagniML, LacquanitiF:Kinematicstrategiesinnewlywalkingtoddlers steppingoverdifferentsupportsurfaces.JNeurophysiol2010, 103:1673-1684.

Thispaperdescribesthestrategiesusedbytoddlerssteppingoveran obstacle,walkingonaninclinedsurfaceorastaircase.Inadults,the covarianceplane describing the inter-segmental coordinationrotates

systematicallyacross tasks,whereas itsorientation remainsconstant intoddlers.Thisbehaviorrestrictsthenumberofdegreesoffreedomand simplifiesthecoordination.Moreover,thetoddlersoftenplacethefoot ontotheobstacleoracrosstheedgesofthestairs,aspartofahaptic exploratorystrategy.

61. GallowayJC,ThelenE:Feetfirst:objectexplorationinyoung infants.InfantBehavDev2004,27:107-112.

62. KarasikLB,AdolphKE,Tamis-LemondaCS,ZuckermanAL:Carry on:spontaneousobjectcarryingin13-month-oldcrawlingand walkinginfants.DevPsychol2011http://dx.doi.org/10.1037/ a0026040.

63. YangJF,LamontEV,PangMY:Split-belttreadmillsteppingin infantssuggestsautonomouspatterngeneratorsfortheleft andrightleginhumans.JNeurosci2005,25:6869-6876.

64. 

MusselmanKE,PatrickSK,VasudevanEV,BastianAJ,YangJF: Uniquecharacteristicsofmotoradaptationduringwalkingin youngchildren.JNeurophysiol2011,105:2195-2203.

Thisandthefollowingstudyusesplit-belttreadmillwalkingasameansto probemotoradaptationandlearninginchildren.Inthispaper, 8–36-monthschildrenwerestudied.Mostlearntthenovelconditionasshown bytheslow,progressivereductionofasymmetriesintemporal coordina-tionwithpractice,andbythepresenceofaftereffectsintheearly post-splitperiod.Instead,asymmetriesinspatialcoordinationpersistedduring split-beltwalkingandnoaftereffectwasseen.Theseresultsare con-sistentwiththehypothesisthatthemechanismscontrollingtemporaland spatialadaptationaredifferentandmatureatdifferenttimes.

65. 

VasudevanEV,Torres-OviedoG,MortonSM,YangJF,BastianAJ: Youngerisnotalwaysbetter:developmentoflocomotor adaptationfromchildhoodtoadulthood.JNeurosci2011, 31:3055-3065.

Here3–18-yearschildrenandadultswithcerebellardamagewere stu-died.Allhealthychildrenandadultslearntthenewtimingatthesamerate andshowedsignificantaftereffects.However,childrenyoungerthan 6-yearsdidnotlearnthenewspatialcoordination.Youngchildrenshowed patterns similarto cerebellar patients, withgreaterdeficits inspatial versustemporaladaptation.Takentogether,thisandthepreviouspaper showthatthematurationoflocomotoradaptationfollowsatleasttwo timecourses,onefortemporalandtheotherforspatialcoordination,and thismaydependonthedevelopmentalstateofthecerebellum.

66. DominiciN,DapratiE,NicoD,CappelliniG,IvanenkoYP, LacquanitiF:Changesinthelimbkinematicsandwalking distanceestimationaftershankelongation.Evidencefora locomotorbodyschema?JNeurophysiol2009,101:1419-1429.

67. AisenML,KerkovichD,MastJ,MulroyS,WrenTA,KayRM, RethlefsenSA:Cerebralpalsy:clinicalcareandneurological rehabilitation.LancetNeurol2011,10:844-852.

68. BergerW:Characteristicsoflocomotorcontrolinchildrenwith cerebralpalsy.NeurosciBiobehavRev1998,22:579-582.

69. HodappM,VryJ,MallV,FaistM:ChangesinsoleusH-reflex modulationaftertreadmilltraininginchildrenwithcerebral palsy.Brain2009,132:37-44.

70. Valentin-GudiolM,Mattern-BaxterK,Girabent-Farre´sM, Bagur-CalafatC,Hadders-AlgraM,Angulo-BarrosoRM: Treadmillinterventionswithpartialbodyweightsupportin childrenundersixyearsofageatriskofneuromotordelay. CochraneDatabaseSystRev2011,12:CD009242.

71. VandeWalleP,HallemansA,SchwartzM,TruijenS,GosselinkR, DesloovereK:Mechanicalenergyestimationduringwalking: validityandsensitivityintypicalgaitandinchildrenwith cerebralpalsy.GaitPosture2012,35:231-237.

72. DickinsonMH,FarleyCT,FullRJ,KoehlMAR,KramR,LehmanS: Howanimalsmove:anintegrativeview.Science2000, 288:100-106.

73. MausHM,LipfertSW,GrossM,RummelJ,SeyfarthA:Upright humangaitdidnotprovideamajormechanicalchallengefor ourancestors.NatCommun2010http://dx.doi.org/10.1038/ ncomms1073.