The physiological basis of neurorehabilitation - locomotor training after spinal cord injury.

0.43 MB·English

by  Hubli, Michele

HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 JNERJOURNAL OF NEUROENGINEERING
http://www.jneuroengrehab.com/content/10/1/5 AND REHABILITATION
REVIEW Open Access
The physiological basis of neurorehabilitation -
locomotor training after spinal cord injury
Michèle Hubli* and Volker Dietz
Abstract
Advances in our understanding of thephysiological basis of locomotion enable us to optimize the
neurorehabilitationof patients with lesions to the central nervous system, such as stroke or spinal cord injury (SCI).
It is generally accepted, based onwork inanimal models, that spinal neuronal machinery can produce a
stepping-like output. Inboth incomplete and complete SCI subjects spinal locomotor circuitries can be activated by
functional training which providesappropriate afferent feedback. In motor complete SCI subjects, however, motor
functions caudal to the spinal cord lesion are nolonger used resulting in neuronal dysfunction. In contrast, in
subjects with anincompleteSCI such training paradigms can lead to improved locomotor ability. Appropriate
functional training involves the facilitation and assistance ofstepping-like movements withthe subjects’ legs and
body weight support as far as is required. Inseverely affected subjects standardized assisted locomotor training is
provided by body weight supported treadmill training with leg movements either manually assisted or moved by a
driven gait orthosis. Load- and hip-joint related afferent input is of crucial importanceduring locomotor training as
it leads to appropriate leg muscle activation and thus increases the efficacy of therehabilitative training. Successful
recovery of locomotion after SCI relies on theability of spinal locomotor circuitriesto utilize specific multisensory
information to generate a locomotor pattern. It seems that a critical combination ofsensory cues is required to
generate and improve locomotor patternsafter SCI. Inaddition to functional locomotor training thereare numbers
ofother promising experimentalapproaches,such as tonic epidural electrical or magnetic stimulation of thespinal
cord, which both promote locomotor permissive states that lead to a coordinatedlocomotor output. Therefore, a
combination of functional training and activationof spinal locomotor circuitries, for example by epidural/flexor
reflexelectrical stimulation or drug application (e.g. noradrenergic agonists), might constitute aneffective strategy
to promote neuroplasticity after SCI in thefuture.
Keywords: Locomotion, Neurorehabilitation, Neuronal plasticity, Spinal cord injury
Introduction accepted that the central nervous system is able to re-
A spinal cordinjury(SCI) isadevastatingeventthat,de- cover locomotor function following incomplete SCI with
pending on the level and severity, impacts sensorimotor functional training on a treadmill combined with partial
and autonomous function. In affected subjects the goal body weight support [3-5]. However, the physiological
of rehabilitative interventions is the regaining of inde- requirements for training effects remain vague. A cen-
pendence and thus a good quality of life. From the tury of research into the organization of the neuronal
patientsperspectivethisisprobablybestachievedbytar- processes underlying the control of locomotion in mam-
geting restoration of bladder and bowel function, and in mals has demonstrated that the basic neuronal circuit-
tetraplegic subjects upper limb function [1]. However, ries responsible for generating efficient stepping patterns
recovery of locomotor ability is also of high priority by are embedded within the lumbosacral spinal cord [6,7].
SCI subjects independently from the severity, time after These spinal locomotor circuitries appear to play a cru-
injury and age at the time of injury [2]. It is now widely cial role in stepping ability in animal models and in
human SCI. Thisreview covers the physiological basis of
effective locomotor training after SCI. Several studies
*Correspondence:[email protected]
SpinalCordInjuryCenter,BalgristUniversityHospital,UniversityofZurich, in animal models, especially in rats and cats, have
Forchstrasse340,8008,Zurich,Switzerland
©2013HubliandDietz;licenseeBioMedCentralLtd.ThisisanOpenAccessarticledistributedunderthetermsofthe
CreativeCommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,
distribution,andreproductioninanymedium,providedtheoriginalworkisproperlycited.
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page2of8
http://www.jneuroengrehab.com/content/10/1/5
unraveled the neuronal principles underlying neuroreh- control, especially during swing phase of stepping in
abilitation of locomotion after SCI. The role of neuronal humans[16]andthecat[17].
plasticity will be discussed in this review. After SCI neu- Quadrupedal locomotion is characterized by the coord-
roplasticity in the cortex, brainstem and spinal cord can ination of forelimb and hindlimb rhythmic activities gen-
be exploited by rehabilitative approaches. This review erated by common spinal neuronal control mechanisms,
will mainly focus on studies of neuroplastic changes at i.e., long propriospinal neurons coupling the cervical and
the spinal level as the knowledge gained from these lumbar segments [18,19]. This neuronal coupling and co-
studies is used to create novel translatory neurorehabil- ordination of upper and lower limbs that is present in
itative approaches with the aim to restore and improve quadrupedal locomotion is preserved inbipedal gait.Uni-
locomotorability afterhumanSCI. lateraltibial nervestimulationduringlocomotion,butnot
during sitting or standing, leads to reflex responses in leg
musclesandintheproximalmusclesofbotharms[20,21].
Neuronalbasisofhumanlocomotion Suchtask-dependentcouplingofthoraco-lumbarandcer-
Thequestion,howdoesthecentralnervoussystemcoord- vical locomotor centers is flexible and allows humans to
inate limb movements during locomotion in a seemingly usetheupperlimbforfine,skilledmovements,oralterna-
“simple”andautomaticmannerchallengesneuroscientists tively for locomotor tasks, such as swimming or crawling,
for morethanacentury. Atthebeginningofthelastcen- or for the control of body equilibrium during stepping,
tury (1911) Graham-Brown postulated his “half-center” e.g.,armswingasaresidualfunctionofquadrupedalloco-
hypothesis which demonstrated the intrinsic capacity of motion[22].
the mammalian spinal cord to generate rhythmic motor It is important that the neuronal mechanisms under-
patterns without descending or sensory input [8]. Subse- lying human locomotor control in the normal and
quently Grillner called these spinal neuronal circuitries pathophysiologicalconditionareunderstood,asitisonly
central pattern generators (CPGs). CPGs are embedded then that it is possible to maximize the recovery of loco-
withinthelumbosacralspinalsegmentsandarecapableof motion in patients following central nervous system
generating stepping-like activation patterns [6]. However, damage.
a CPG alone does not appear to be sufficient for over-
ground walking. Locomotion represents the interaction Focusingonneuronalplasticitybroughtaboutby
betweentheinnatepatternandanappropriatemodulation locomotortraining
of leg muscle activation which has to continuously adapt Modern neurorehabilitation no longer aims to simply
tothe present requirements, e.g., tothe over-ground con- compensate for disabilities inSCI subjects, rather it aims
ditions. Feedback from a variety of sources, e.g., visual, to functionally regain locomotor ability by exploiting
vestibular and proprioceptive systems, is interpreted by neural plasticity and/or neural repair. A first question to
and then integrated into the activity of the CPG [9]. The ask is, “What is neuronal plasticity?” An example of
CPGcanopenandclosereflexpathwaysinacontext-and neuronal plasticity is the spontaneous reorganization
task-dependent manner. The sensory feedback and the that is observed after SCI at the cortical level which can
context-specificrequirementsofthemotortaskdetermine occur over one year [23]. Next to spontaneously occur-
the mode of organization of muscle synergies [10]. Add- ring neuronal plasticity, it can also be induced by loco-
itionally,supraspinalcontrolisneededtoprovideboththe motor training at cortical level [24]. Plastic changes in
drive for locomotion as well as the coordination to inter- sensorimotorcortexactivityarerelatedtofunctionalloco-
act with a complex environment. Corticospinal access to motorrecoveryafteranSCI[25].Besidescorticalplasticity
locomotion control in humans is phase-dependent [11]. it seems that other supraspinal centers, such as the cere-
Brain centers can initiate CPG activity butthe fundamen- bellumandbrainstemarealsoimportantsitesofneuronal
tal rhythmicity is hard-wired. For example, in the cat, it plasticity in humans receiving locomotor training after
wasshownthatapplicationofclonidine,asubstancemim- SCI[24].Inhumans,itisassumedthat supraspinalplasti-
icking the action of long descending pathways, results in cityisassociatedwithplasticityofspinalneuronalcircuits,
distinct and consistent alternating bursts of electromyo- but the evidence for this is predominantly from animal
graphic (EMG) activity which induce spinal stepping [12]. modelsofSCI[26].
In humans, neuroimaging methods have revealed that Spinal neuronal circuits below the level of lesion can
distinct cortical areas, e.g., both the medial primary be activated by an appropriate afferent input, and this is
sensory-motor cortices and the supplementary motor consideredimportant tosustain functionalrecoveryafter
areas, become activated during locomotion [13,14] and an SCI [9]. In contrast, typical movement disorders after
the size of the activated areas is related to the subjects’ SCI, e.g., spastic movement disorder, are due to the de-
walking speed [15]. In addition, more demanding tasks, fective utilization of afferent input in combination with
such as walking over obstacles, require more cortical secondary compensatory mechanisms [27]. It has been
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page3of8
http://www.jneuroengrehab.com/content/10/1/5
shownthatneuronalnetworksunderlyingthe generation than on the completeness of the injury [43]. Conse-
of locomotor patterns of cats [28] and humans [29,30] quently, the better the stepping ability of SCI subjects,
have an impressively high level of flexibility after SCI. the less the neuronal dysfunction. The functional state
Rehabilitative interventions after SCI should therefore of spinal locomotor circuitries is not fixed after an SCI
focus on exploiting the plasticity of neuronal circuits, and neuronal dysfunction can be improved by intensive
i.e., at supraspinal and/or spinal level, rather than focus- locomotor training over one month, but only in incom-
ing on improving isolated clinical signs, such as muscle plete SCIsubjects[43].
toneorreflexexcitability.
The plasticity of spinal neuronal circuits is task- Locomotorcapacityafterspinalcordinjury
specific and use-dependent as shown in several earlier Locomotorpatterngeneration:fromanimaltohuman
experiments in cats with complete SCI. For example, Cats with a complete spinal cord transection at thoracic
after several months of daily step training, spinal cats segments gradually improve hindlimb locomotion on a
regained full weight-bearing locomotion on a treadmill treadmill following 2–3 weeks of daily locomotor train-
[31,32]. If a spinal cat is intensively trained to stand it ing [31,33]. The spinal cat can relearn walking with al-
developstheabilitytosupportitsbodyweightforuptoan ternating steps in the hindlimbs, body weight support
hour, but stepping ability on the treadmill remains poor and plantar foot placement. Under such circumstances
[33]. These findings suggest that spinal neuronal circuits the EMG activity of the hindlimbs was remarkably simi-
learn the sensorimotor task that is specifically practiced larbeforeandafterthespinalcordtransection.However,
and trained [34]. The repetitive activation of particular the EMG amplitudes of leg extensors were smaller,
sensorimotor pathways by task-specific training can whilst the timing between flexor muscles, such as hip
reinforce circuits and synapses used to successfully per- and knee joint flexors, was changed and lead to foot
form the practiced movement [35,36]. Therefore, the out- dragging early in the swing phase [44]. With ongoing
come of a neurorehabilitative approach strongly depends training it has been shown that the body support can be
uponthetype,therepetitionandthequalityofthetrained decreased and this is associated with improved loco-
motorfunction. motor capacity until a complete support of body weight
Neuronal plasticity is not always positive, anatomical and well-coordinated hindlimb stepping movements are
and neurophysiological observations in animals suggest possible [45]. Experiments in cats as well as non-human
that after SCI severed axons degenerate and create free primates with complete spinal cord lesions have shown
synaptic territories which could become re-occupied by that the isolated spinal cord has the capacity to produce
sprouting of intraspinal fibres [37]. The new neuronal steppingpatterns [46].
circuits may be aberrant and can lead to inappropriate There are several indications for the existence of spinal
movement patterns or pain in rats [38] and humans [39]. neuronal circuitries for locomotion generation in humans.
This can largely be prevented by a combination of loco- For example, step-like leg movements are present at birth
motor training, electrical and pharmacological stimula- andcanbeinitiatedspontaneouslyorbyperipheralstimuli
tions of the region of the spinal cord that is deprived of [47].ItislikelythattheEMGactivityunderlyingthisnew-
supraspinal input which leads to a weight-bearing loco- bornsteppingisproduced atthespinallevel,ascorticosp-
motor capacity in spinal rats [38]. For human SCI, it will inal projections are not fully grown and myelinated in
beimportanttomakesurethatfutureneurorehabilitatitive newborns. In addition, rhythmic: coordinated leg move-
approaches direct the spontaneous and/or experimentally mentshavebeenobservedinmotorcompleteSCIsubjects,
induced (e.g., stem cells or Nogo-A antibodies) neuronal duringsleep[48]andafterspinalcordstimulation[49,50].
plasticitytowardsfunctionalsynapticconnectionsthatare
associatedwithanimprovedlocomotorperformance. EffectoflocomotortrainingafterSCI
The decline in supraspinal and peripheral input in In motor complete and incomplete SCI subjects a coor-
humans after a severe SCI is suggested to be responsible dinated leg muscle activation pattern in both legs can be
for the development of a neuronal dysfunction below induced following partial unloading standing on a mov-
the level of lesion [40,41]. Characteristic changes in ing treadmill (Figure 1) [4,51]. In complete SCI subjects
neuronal behavior occur up to approximately one year with no voluntary motor control below the level of le-
after SCI. At one year post-injury complete and severe sion leg movements have to be assisted manually or by a
but incomplete SCIs are characterized by an exhaustion robotic device during the whole training period. These
of leg muscle activity during assisted locomotion which subjects cannot relearn the ability to perform unsup-
is associated with changes in polysynaptic spinal reflexes ported stepping movements on solid ground, but follow-
[42]. Neuronal dysfunction after SCI is assumed to be ing such training they experience positive effects on
dependent more on the subjects’ loss of mobility, i.e., cardiovascular and musculo-skeletal systems, such as a
decreased appropriate afferent information to the CPG, reductionofmusclespasms.
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page4of8
http://www.jneuroengrehab.com/content/10/1/5
anterior activation decreases [53]. This leads to an im-
provement inweightbearing functionduringstancephase
and thus allows a decrease of body unloading during
assisted treadmill locomotion. The successive reloading of
theSCIsubjectsmightbeanimportantstimulusforexten-
sorloadreceptorswhichareessentialforlegextensoracti-
vationduringlocomotionincats[54]andhumans[55,56].
Thegeneral rehabilitation strategy to regain locomotor
ability after SCI or stroke is based on the principles of
motor learning, such as task-specificity, task-variability,
feedbackinformationandthe intensity of training. So far
training paradigms for robotic assisted treadmill training
in human SCI rely on suggestions by experts and single
observational studies. However, no study has so far
determined the essential components required for a
locomotor training setup in individuals with SCI or in
animal models of SCI. Several questions remain to be
answered,forexample,howearlyshouldthetrainingther-
apy start, how intensiveshould itbe, e.g.,30min ormore
than 1 hour and how many times a week? It is not yet
clear whether a longer duration of training results in an
improved outcome or if certain endpoints, in terms of
walking capacity, can be achieved within shorter periods
of time. This question has so far only been addressed in
strokesubjectswherealongerdurationofearlyonsetaug-
mented locomotor training, i.e., 16 hours within the first
6monthsafterinsult,wascorrelatedwithimprovedwalk-
Figure1Twoexamplesoflocomotoractivityduringassisted ing performance (positive dose–response relationship)
walkinginthedrivengaitorthosisLokomatin(A)anacute
(3monthsafterSCI)and(B)achronic(41monthsafterSCI) [57]. Whether this relationship is also applicable to the
paraplegicsubject.BothSCIsubjectssufferedamotorcomplete SCIcommunitywillbeinvestigatedbyacontrolledrando-
spinalcordlesionandlegmuscleactivitywasassessedinrectus mizedmulticenterstudy[58].Inaddition,itremainstobe
femoris(RF),bicepsfemoris(BF),tibialisanterior(TA)and
determinedwhichtrainingparadigmsfitthebesttowhich
gastrocnemiusmedialis(GM)(modifiedfrom[42]).
SCIsubjectsandhowthistrainingcouldbeadaptedindif-
ficulty, e.g., by the amount of body weight support, guid-
ance force, treadmill speed. In stroke subjects higher
In the early phase of rehabilitation after a severeincom- treadmill speeds lead to better locomotor recovery than
plete SCI, the main limitations for overground ambulation lowertreadmillspeeds[59].Itisassumedthatthesetrain-
areusuallyreducedcoordination,legparesis,andimpaired ing principles which were applied in stroke subjects are
balance. Therefore, at the beginning of the training pro- also reasonable for successful training in SCI subjects. In
gram locomotor training has to be assisted either by phy- general, locomotor training should always be challenging
siotherapists or a robotic device (e.g. Lokomat). During forpatientswithonlyminimalsupportprovidedbythera-
assisted locomotion in incomplete SCI subjects the pistsorroboticdevices.
induced leg muscle EMG activity is modulated over the
step cycles in a similar way to as observed in healthy sub- Appropriatesensorycues
jects.However,theamplitudeoflegmuscleactivityiscon- AfteranincompleteSCI,sparedcorticospinaland/orpro-
siderably smaller and corresponds to the degree of paresis priospinalpathwayscanplayanactiveroleintherecovery
probablyduetoadiminishednoradrenergicinputtospinal of locomotion. However, under these circumstances the
locomotor circuitries [45]. Providing full body unloading intrinsic capacity of spinal locomotor circuitries and the
duringroboticassistedwalkingincompleteparaplegicsub- sensoryfeedbackinformationstillremainsasthebasisfor
jectsdoesnotleadtosignificantlegmuscleactivation[52], generatingalocomotorpattern.Thespinallocomotorcir-
i.e., ground contact is essential for leg muscle activation. cuitries interact dynamically with specific afferent inputs
During the course of an assisted locomotor training pro- from receptors located in muscles, joints, and skin, and
gram, the EMG amplitudes of gastrocnemius activity in- this interaction shapes the locomotor output (for review
crease during the stance phase, while inappropriate tibialis see [9]). The sensory input most relevant for locomotion
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page5of8
http://www.jneuroengrehab.com/content/10/1/5
comesprimarilyfromstretch-andload-sensitivemechan- duringthe stancephase inhibitsthe flexor activity.This is
oreceptorslocatedinthemusclesandskin.Afferentinfor- functionallymeaningfulbecauseloadingofthestancelimb
mationfromhipflexorsandankleextensorsisparticularly has to decrease before swing can be initiated and leg ex-
important for the stance to swing transition phase in cats tensor activity is reinforced during the stance phase by
[60].Furthermore,sensoryinputfromtheskinmechanor- positive feedback [68]. In addition, the role of hip joint
eceptors of the paw are involved in positioning the paws afferents is assumed to control phase transition and
and, therefore, become increasingly important during reinforce ongoingactivity as it has been shown indecere-
complexlocomotortasks in whichprecisepaw placement bratedcats[65].Theobservationsthatinmotorcomplete
isrequired,e.g.,ladderwalkinginthecat[61].Skinrecep- paraplegic subjects assisted stepping movements within a
torson the dorsal foot play a role during the swingphase driven gait orthosis and restricted movements of the hips
of walking over obstacles in both cats [62] and humans (blocked knees) induces a patterned leg muscle EMG ac-
[63] and cutaneous input from the plantar surface of the tivity, highlights the significance of hip joint receptors in
pawreinforcesextensoractivityindecerebratedcatswalk- the generation of locomotor activity [52]. Such that
ing on a treadmill [64]. Group II hip flexor afferents, assisted stepping movements restricted to imposed ankle
group I ankle extensor afferents, and low-threshold cuta- joints were followed by no, or only focal reflex responses
neousafferentsfromthecatpawareassumedtohavedir- inthestretchedmuscles[52].
ect access to spinal locomotor circuitries with the ability
toresetorentrainfictivelocomotioninadultdecerebrated Modulationofspinalneuronalexcitability
cats[65]. The importance of appropriate afferent information from
Afferentinputsfromloadandhipjointreceptorsarees- peripheralreceptorsasasourceofcontrollinglocomotion
sentialfortheactivationoflegmuscleactivityduringloco- became obvious from experiments, during the last dec-
motioninparaplegicsubjectsinacorrespondingwaytoin ades, in completely transected animals. The inability to
cats,[52,66].Loadinformationisprovidedforpropriocep- producelocomotorpatternsafterasevereSCIis,toalarge
tive input from leg extensor muscles, namely Ib afferent extent, due to the depressed functional state of spinal
signals fromGolgitendonorgans, and probably alsofrom locomotor circuitries [69]. It is believed that reduced sen-
mechanoreceptorsinthefootsole(seeFigure2)[67].This sory feedback after SCI has a negative impact on loco-
information is thought to be integrated into polysynaptic motor recovery. Therefore, there is a need to develop
spinal reflex pathways that adapt the autonomous loco- tools to artificially activate spinal cord pattern generators.
motor pattern to the actual ground condition and it is Besides the essential sensory cues provided during loco-
assumed that the Ib afferent input from leg extensors motor training, i.e., from body loading and hip joint
Figure2InfluenceofbodyloadingonlegmuscleEMGactivity.Loading/unloadingeffectsontheEMGactivityofthelefttibialisanterior(TA,
ankleflexor)andgastrocnemiusmedialis(GM,ankleextensor).EMGsaretheaveragesfromsevenhealthysubjects.Threewalkingconditionsare
shown:normalbodyloading(NBL),30%bodyloading(BL),and60%bodyunloading(BU).Inallsubjects,therewasastrongreductionofGM
EMGactivityduringBUcomparedwithNBLand30%BLwalking(modifiedfrom[67]).
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page6of8
http://www.jneuroengrehab.com/content/10/1/5
afferents,therehasbeenmucheffort spentindeveloping (between 35 and 40 cm2) these stimulation parameters re-
additional strategies to increase the excitability of spinal sult in a total delivered charge of 0.056 and 0.064 C/cm2
neuronal circuitries in order to tune the physiological respectively [81-83,86,87]. The dispersion and density of
state of these circuitries to a level that leads to a facili- the current is, however, hard to predict due to the large
tation of the locomotor patterns in humans. These volume of the conductor surrounding the target tissue. So
approachesincludecontinuousvibrationofthequadriceps far tsDCS has only been applied in healthy subjects as a
and hamstring muscle groups [70], continuous electrical tool to modulate trans-synaptic efficacy in monosynaptic
stimulation of the peroneal or sural nerve [71], and mag- [82,86] and polysynaptic reflex pathways [83], dorsal col-
netic stimulation of the spinal cord [72]. The latter ap- umn function [81] or pain thresholds [87]. The effect on
proach of repetitive electromagnetic stimulation shows theexcitabilityofspinallocomotorcircuitries,especiallyin
positive effects on spastic muscle tone in patients with di- SCIsubject,needstobeaddressedinthefuture.
verse neurological diseases [73-75]. So far only one study
investigated the effect of repetitive magnetic stimulation
Conclusion
applied at the thoracolumbar vertebral level on spinal lo-
Designing effective neurorehabilitation after SCI depends
comotor circuitries in healthy subjects [72]. The most
on having knowledge about the neuronal mechanisms
favourable stimulation parameters to induce stepping-like
involved in normal and pathological movement condi-
activityinthelegsplacedinagravity-neutralpositionwere
tions, such as the interactions between central programs
observedtobe3Hz,1.3-1.82teslaattheT11-12vertebrae
and afferentfeedback as well as the quadrupedal coordin-
[72].However,thereisnoproofthatthiskindofmagnetic
ationof humanlocomotion.Task-specificrehabilitationto
stimulationisabletoevokesufficientlegmuscleactivation
improve locomotor function after SCI is well established
toleadtoload-bearinglocomotioninSCIsubjects.
[51,88]. In addition, there is growing evidence that the
Incontrasttothelowfrequency(3Hz)forthoracolum-
enhanced application of afferent input to the spinal cord,
barlevels,higherfrequencies(20–30Hz)arenecessaryfor
forexamplebyelectricalepiduralstimulationofthespinal
peripheral nerve stimulation to decrease spastic muscle
cord, can lead to a facilitation of standing and stepping-
tone [74] and also for electrical epidural stimulation
like activity during assisted leg movements in human SCI
(~40Hz)atthelumbarspinal cordincompleteSCIsub-
[77]. Consequently, it is argued that specific multi-
jectstoinducestepping-likeactivity[49,76,77].Inaddition
pronged neurorehabilitative approaches will pave the way
to electrical or magnetic stimulation approaches, diverse
for more efficient therapeutic strategies to improve loco-
pharmacologicalagents,suchasserotonergicandnoradre-
motor functionin severely affected SCI subjects [89].The
nergicagonists,canactivatespinallocomotorcircuitriesin
aim of new neurorehabilitative approaches should be to
rats,catsandmice[38,78,79].However,sofarthereisonly
optimize the use of task-specific sensory cues in order to
limited evidence about whether pharmacological agents
facilitate locomotor pattern generation involving arm
facilitatelocomotorrecoveryfollowinghumanSCI(forre-
movementsandtheprovisionofapproachesfavouringthe
view see [80]). Recently, two new noninvasive methods,
recruitment of both spinal circuitries and spared suprasp-
namely transcutaneous spinal direct current stimulation
inalconnectionsduringrehabilitation.
(tsDCS) [81-83] and paired spinal associative stimulations
of H-reflexes and transcranial magnetic stimulation [84],
Competinginterests
have been applied to modulate spinal excitability and Theauthorsdeclarethattheyhavenocompetinginterests.
might be a promising tool for the modulation of sensori-
motorpathwaysunderlyingfunctionalmovements. Authors’contributions
MHstructuredthecontentofthereviewandwrotethedraftofthe
TsDCS is a method derived from transcranial direct
manuscript.VDmadetherevisionsofthewholearticle.Bothauthorsread
current stimulation and it influences neuronal excitability andapprovedthefinalversionofthemanuscript.
byanodalorcathodalpolarization.Anodalstimulationtyp-
ically increases neuronal firing rates in stimulated cortical Acknowledgements
areas, whereas cathodal stimulation evokes the opposite ThisreviewhasbeencarriedoutwithfinancialsupportfromtheWingsfor
Lifefoundation.AspecialthankgoestoMichelleStarkeyforhereditorial
effect, i.e., inhibition. Depending on the duration and
help.
strengthofpolarization,thesechangescanpersistformore
than one hour after stimulation [85]. For the spinal ap- Received:27March2012Accepted:7January2013
Published:21January2013
proach the electrode positions are changed from the scalp
to one electrode at T11 vertebral level and the other elec-
References
trode on a shoulder region [81-83,86,88]. The polarization 1. AndersonKD:Targetingrecovery:prioritiesofthespinalcord-injured
of the stimulation (anodal or cathodal) refers to the spinal population.JNeurotrauma2004,21(10):1371–1383.
2. DitunnoPL,PatrickM,StinemanM,DitunnoJF:Whowantstowalk?
electrode. TsDCS is applied for 15 min with a stimulation
PreferencesforrecoveryafterSCI:alongitudinalandcross-sectional
intensity of 2.5 mA. Depending on the electrode size study.SpinalCord2008,46(7):500–506.
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page7of8
http://www.jneuroengrehab.com/content/10/1/5
3. BarbeauH,McCreaDA,O'DonovanMJ,RossignolS,GrillWM,LemayMA: 29. DietzV:Proprioceptionandlocomotordisorders.NatRevNeurosci2002,
Tappingintospinalcircuitstorestoremotorfunction.BrainResRev1999, 3(10):781–790.
30(1):27–51. 30. EdgertonVR,TillakaratneNJ,BigbeeAJ,deLeonRD,RoyRR:Plasticityof
4. DietzV,ColomboG,JensenL:Locomotoractivityinspinalman. thespinalneuralcircuitryafterinjury.AnnuRevNeurosci2004,27:145–167.
Lancet1994,344(8932):1260–1263. 31. BarbeauH,RossignolS:Recoveryoflocomotionafterchronic
5. WernigA,MullerS:Laufbandlocomotionwithbodyweightsupport spinalizationintheadultcat.BrainRes1987,412(1):84–95.
improvedwalkinginpersonswithseverespinalcordinjuries.Paraplegia 32. LovelyRG,GregorRJ,RoyRR,EdgertonVR:Effectsoftrainingonthe
1992,30(4):229–238. recoveryoffull-weight-bearingsteppingintheadultspinalcat.Exp
6. GrillnerS:Neurobiologicalbasesofrhythmicmotoractsinvertebrates. Neurol1986,92(2):421–435.
Science1985,228(4696):143–149. 33. DeLeonRD,HodgsonJA,RoyRR,EdgertonVR:Fullweight-bearing
7. SherringtonCS:Flexion-reflexofthelimb,crossedextensionreflex,and hindlimbstandingfollowingstandtrainingintheadultspinalcat.
reflexsteppingandstanding.JPhysiol(Lond)1910,40:28–121. JNeurophysiol1998,80(1):83–91.
8. Graham-BrownT:Theintrinsicfactorsintheactofprogressioninthe 34. TillakaratneNJ,deLeonRD,HoangTX,RoyRR,EdgertonVR,TobinAJ:Use-
mammal.ProcRSopcLondBBiolSci1911,84:308–319. dependentmodulationofinhibitorycapacityinthefelinelumbarspinal
9. RossignolS,DubucR,GossardJP:Dynamicsensorimotorinteractionsin cord.JNeurosci2002,22(8):3130–3143.
locomotion.PhysiolRev2006,86(1):89–154. 35. EdgertonVR,CourtineG,GerasimenkoYP,LavrovI,IchiyamaRM,FongAJ,
10. HorakFB,NashnerLM:Centralprogrammingofposturalmovements: CaiLL,OtoshiCK,TillakaratneNJ,BurdickJW,RoyRR:Traininglocomotor
adaptationtoalteredsupport-surfaceconfigurations.JNeurophysiol1986, networks.BrainResRev2008,57(1):241–254.
55(6):1369–1381. 36. CaiLL,CourtineG,FongAJ,BurdickJW,RoyRR,EdgertonVR:Plasticityof
11. SchubertM,CurtA,JensenL,DietzV:Corticospinalinputinhumangait: functionalconnectivityintheadultspinalcord.PhilosTransRSocLondB
modulationofmagneticallyevokedmotorresponses.ExpBrainRes1997, BiolSci2006,361(1473):1635–1646.
115(2):234–246. 37. HouS,DualeH,CameronAA,AbshireSM,LyttleTS,RabchevskyAG:
12. EdgertonVR,RoyRR,HodgsonJA,ProberRJ,deGuzmanCP,deLeonR: Plasticityoflumbosacralpropriospinalneuronsisassociatedwiththe
Potentialofadultmammalianlumbosacralspinalcordtoexecuteand developmentofautonomicdysreflexiaafterthoracicspinalcord
acquireimprovedlocomotionintheabsenceofsupraspinalinput. transection.JCompNeurol2008,509(4):382–399.
JNeurotrauma1992,9(1):119–128. 38. CourtineG,GerasimenkoY,vandenBrandR,YewA,MusienkoP,ZhongH,
13. MiyaiI,TanabeHC,SaseI,EdaH,OdaI,KonishiI,TsunazawaY,SuzukiT, SongB,AoY,IchiyamaRM,LavrovI,etal:Transformationofnonfunctional
YanagidaT,KubotaK:Corticalmappingofgaitinhumans:anear-infrared spinalcircuitsintofunctionalstatesafterthelossofbraininput.
spectroscopictopographystudy.Neuroimage2001,14(5):1186–1192. NatNeurosci2009,12(10):1333–1342.
14. FukuyamaH,OuchiY,MatsuzakiS,NagahamaY,YamauchiH,OgawaM, 39. MaegeleM,MullerS,WernigA,EdgertonVR,HarkemaSJ:Recruitmentof
KimuraJ,ShibasakiH:Brainfunctionalactivityduringgaitinnormal spinalmotorpoolsduringvoluntarymovementsversussteppingafter
subjects:aSPECTstudy.NeurosciLett1997,228(3):183–186. humanspinalcordinjury.JNeurotrauma2002,19(10):1217–1229.
15. SuzukiM,MiyaiI,OnoT,OdaI,KonishiI,KochiyamaT,KubotaK:Prefrontal 40. HubliM,BolligerM,DietzV:Neuronaldysfunctioninchronicspinalcord
andpremotorcorticesareinvolvedinadaptingwalkingandrunning injury.SpinalCord2011,49(5):582–587.
speedonthetreadmill:anopticalimagingstudy.Neuroimage2004, 41. DietzV:Behaviorofspinalneuronsdeprivedofsupraspinalinput.Nat
23(3):1020–1026. RevNeurol2010,6(3):167–174.
16. SchubertM,CurtA,ColomboG,BergerW,DietzV:Voluntarycontrolof 42. DietzV,GrillnerS,TreppA,HubliM,BolligerM:Changesinspinalreflex
humangait:conditioningofmagneticallyevokedmotorresponsesina andlocomotoractivityafteracompletespinalcordinjury:acommon
precisionsteppingtask.ExpBrainRes1999,126(4):583–588. mechanism?Brain2009,132(8):2196–2205.
17. ArmstrongDM:Thesupraspinalcontrolofmammalianlocomotion. 43. HubliM,DietzV,BolligerM:Spinalreflexactivity:amarkerforneuronal
JPhysiol1988,405:1–37. functionalityafterspinalcordinjury.NeurorehabilNeuralRepair2012,
18. CazaletsJR,BertrandS:Couplingbetweenlumbarandsacralmotor 26(2):188–196.
networksintheneonatalratspinalcord.EurJNeurosci2000, 44. BelangerM,DrewT,ProvencherJ,RossignolS:Acomparisonoftreadmill
12(8):2993–3002. locomotioninadultcatsbeforeandafterspinaltransection.
19. NathanPW,SmithM,DeaconP:Vestibulospinal,reticulospinaland JNeurophysiol1996,76(1):471–491.
descendingpropriospinalnervefibresinman.Brain1996,119:1809–1833. 45. BarbeauH,RossignolS:Enhancementoflocomotorrecoveryfollowing
20. MichelJ,vanHedelHJ,DietzV:Obstaclesteppinginvolvesspinal spinalcordinjury.CurrOpinNeurol1994,7(6):517–524.
anticipatoryactivityassociatedwithquadrupedallimbcoordination. 46. VilenskyJA,O'ConnorBL:Steppinginnonhumanprimateswitha
EurJNeurosci2008,27(7):1867–1875. completespinalcordtransection:oldandnewdata,andimplicationsfor
21. DietzV,FouadK,BastiaanseCM:Neuronalcoordinationofarmandleg humans.AnnNYAcadSci1998,860:528–530.
movementsduringhumanlocomotion.EurJNeurosci2001,14(11):1906–1914. 47. ForssbergH:Ontogenyofhumanlocomotorcontrol.I.Infantstepping,
22. DietzV:Dohumanbipedsusequadrupedalcoordination?TrendsNeurosci supportedlocomotionandtransitiontoindependentlocomotion.Exp
2002,25(9):462–467. BrainRes1985,57(3):480–493.
23. JurkiewiczMT,MikulisDJ,McIlroyWE,FehlingsMG,VerrierMC: 48. LeeMS,ChoiYC,LeeSH,LeeSB:Sleep-relatedperiodiclegmovements
Sensorimotorcorticalplasticityduringrecoveryfollowingspinalcord associatedwithspinalcordlesions.MovDisord1996,11(6):719–722.
injury:alongitudinalfMRIstudy.NeurorehabilNeuralRepair2007, 49. DimitrijevicMR,GerasimenkoY,PinterMM:Evidenceforaspinal
21(6):527–538. centralpatterngeneratorinhumans.AnnNYAcadSci1998,
24. WinchesterP,McCollR,QuerryR,ForemanN,MosbyJ,TanseyK,Williamson 860:360–376.
J:Changesinsupraspinalactivationpatternsfollowingrobotic 50. RosenfeldJE,SherwoodAM,HalterJA,DimitrijevicMR:Evidenceofa
locomotortherapyinmotor-incompletespinalcordinjury.Neurorehabil patterngeneratorinparalyzedsubjectwithspinalcordstimulation.
NeuralRepair2005,19(4):313–324. SocNeurosciAbstr1995,21:688.
25. LundellH,ChristensenMS,BarthelemyD,Willerslev-OlsenM,Biering- 51. DietzV,HarkemaSJ:Locomotoractivityinspinalcord-injuredpersons.
SorensenF,NielsenJB:Cerebralactivationiscorrelatedtoregional JApplPhysiol2004,96(5):1954–1960.
atrophyofthespinalcordandfunctionalmotordisabilityinspinalcord 52. DietzV,MullerR,ColomboG:Locomotoractivityinspinalman:
injuredindividuals.Neuroimage2011,54(2):1254–1261. significanceofafferentinputfromjointandloadreceptors.Brain2002,
26. TanseyKE:Neuralplasticityandlocomotorrecoveryafterspinalcord 125(12):2626–2634.
injury.PmR2010,2(12):220–226. 53. DietzV,ColomboG,JensenL,BaumgartnerL:Locomotorcapacityof
27. DietzV,SinkjaerT:Spasticmovementdisorder:impairedreflexfunction spinalcordinparaplegicpatients.AnnNeurol1995,37(5):574–582.
andalteredmusclemechanics.LancetNeurol2007,6(8):725–733. 54. PearsonKG,CollinsDF:ReversaloftheinfluenceofgroupIbafferents
28. PearsonKG:Neuraladaptationinthegenerationofrhythmicbehavior. fromplantarisonactivityinmedialgastrocnemiusmuscleduring
AnnuRevPhysiol2000,62:723–753. locomotoractivity.JNeurophysiol1993,70(3):1009–1017.
HubliandDietzJournalofNeuroEngineeringandRehabilitation2013,10:5 Page8of8
http://www.jneuroengrehab.com/content/10/1/5
55. DietzV,ColomboG:Effectsofbodyimmersiononposturaladjustments 79. LapointeNP,GuertinPA:SynergisticeffectsofD1/5and5-HT1A/7
tovoluntaryarmmovementsinhumans:roleofloadreceptorinput. receptoragonistsonlocomotormovementinductionincompletespinal
JPhysiol1996,497:849–856. cord-transectedmice.JNeurophysiol2008,100(1):160–168.
56. DietzV,GollhoferA,KleiberM,TrippelM:Regulationofbipedalstance: 80. DomingoA,Al-YahyaAA,AsiriYA,EngJJ,LamT:Asystematicreviewon
dependencyon"load"receptors.ExpBrainRes1992,89(1):229–231. theeffectsofpharmacologicalagentsonwalkingfunctioninpeople
57. KwakkelG,vanPeppenR,WagenaarRC,WoodDauphineeS,RichardsC, withspinalcordinjury.JNeurotrauma2012,29(5):865–879.
AshburnA,MillerK,LincolnN,PartridgeC,WellwoodI,LanghorneP:Effects 81. CogiamanianF,VergariM,PulecchiF,MarcegliaS,PrioriA:Effectofspinal
ofaugmentedexercisetherapytimeafterstroke:ameta-analysis.Stroke transcutaneousdirectcurrentstimulationonsomatosensoryevoked
2004,35(11):2529–2539. potentialsinhumans.ClinNeurophysiol2008,119(11):2636–2640.
58. WirzM,BastiaenenC,deBieR,DietzV:Effectivenessofautomated 82. WinklerT,HeringP,StraubeA:SpinalDCstimulationinhumans
locomotortraininginpatientswithacuteincompletespinalcordinjury: modulatespost-activationdepressionoftheH-reflexdependingon
arandomizedcontrolledmulticentertrial.BMCNeurol2011,11:60. currentpolarity.ClinNeurophysiol2010,121(6):957–961.
59. PohlM,MehrholzJ,RitschelC,RuckriemS:Speed-dependenttreadmill 83. CogiamanianF,VergariM,SchiaffiE,MarcegliaS,ArdolinoG,BarbieriS,
traininginambulatoryhemipareticstrokepatients:arandomized PrioriA:Transcutaneousspinalcorddirectcurrentstimulationinhibits
controlledtrial.Stroke2002,33(2):553–558. thelowerlimbnociceptiveflexionreflexinhumanbeings.Pain2010,
60. PearsonKG:Roleofsensoryfeedbackinthecontrolofstancedurationin 152(2):370–375.
walkingcats.BrainResRev2008,57(1):222–227. 84. CortesM,ThickbroomGW,Valls-SoleJ,Pascual-LeoneA,EdwardsDJ:Spinal
61. BouyerLJ,RossignolS:Contributionofcutaneousinputsfromthe associativestimulation:anon-invasivestimulationparadigmto
hindpawtothecontroloflocomotion.IIntactcatsJNeurophysiol2003, modulatespinalexcitability.ClinNeurophysiol2011,122(11):2254–2259.
90(6):3625–3639. 85. NitscheMA,FrickeK,HenschkeU,SchlitterlauA,LiebetanzD,LangN,
62. ForssbergH:Stumblingcorrectivereaction:aphase-dependent HenningS,TergauF,PaulusW:Pharmacologicalmodulationofcortical
compensatoryreactionduringlocomotion.JNeurophysiol1979, excitabilityshiftsinducedbytranscranialdirectcurrentstimulationin
42(4):936–953. humans.JPhysiol2003,553(1):293–301.
63. SchillingsAM,vanWezelBM,MulderT,DuysensJ:Muscularresponsesand 86. LamyJC,HoC,BadelA,ArrigoRT,BoakyeM:ModulationofSoleus
movementstrategiesduringstumblingoverobstacles.JNeurophysiol H-reflexbyspinalDCstimulationinhumans.JNeurophysiol2012,
2000,83(4):2093–2102. doi:10.1152/jn.10898.2011.
64. DuysensJ,PearsonKG:Theroleofcutaneousafferentsfromthedistal 87. TruiniA,VergariM,BiasiottaA,LaCesaS,GabrieleM,DiStefanoG,Cambieri
hindlimbintheregulationofthestepcycleofthalamiccats.ExpBrain C,CruccuG,InghilleriM,PrioriA:Transcutaneousspinaldirectcurrent
Res1976,24:245–255. stimulationinhibitsnociceptivespinalpathwayconductionand
65. FrigonA,SiroisJ,GossardJP:Effectsofankleandhipmuscleafferent increasespaintoleranceinhumans.EurJPain2011,15(10):1023–1027.
inputsonrhythmgenerationduringfictivelocomotion.JNeurophysiol 88. WirzM,ZemonDH,RuppR,ScheelA,ColomboG,DietzV,HornbyTG:
2010,103(3):1591–1605. Effectivenessofautomatedlocomotortraininginpatientswithchronic
66. HarkemaSJ,HurleySL,PatelUK,RequejoPS,DobkinBH,EdgertonVR: incompletespinalcordinjury:amulticentertrial.ArchPhysMedRehabil
Humanlumbosacralspinalcordinterpretsloadingduringstepping. 2005,86(4):672–680.
JNeurophysiol1997,77(2):797–811. 89. MusienkoP,HeutschiJ,FriedliL,vandenBrandR,CourtineG:Multi-system
neurorehabilitativestrategiestorestoremotorfunctionsfollowing
67. BastiaanseCM,DuysensJ,DietzV:Modulationofcutaneousreflexesby
severespinalcordinjury.ExpNeurol2012,235(1):100–109.
loadreceptorinputduringhumanwalking.ExpBrainRes2000,
135(2):189–198.
68. ProchazkaA,GillardD,BennettDJ:Positiveforcefeedbackcontrolof doi:10.1186/1743-0003-10-5
muscles.JNeurophysiol1997,77(6):3226–3236. Citethisarticleas:HubliandDietz:Thephysiologicalbasisof
neurorehabilitation-locomotortrainingafterspinalcordinjury.Journal
69. HarkemaSJ:Plasticityofinterneuronalnetworksofthefunctionally
isolatedhumanspinalcord.BrainResRev2008,57(1):255–264. ofNeuroEngineeringandRehabilitation201310:5.
70. GurfinkelVS,LevikYS,KazennikovOV,SelionovVA:Locomotor-like
movementsevokedbylegmusclevibrationinhumans.EurJNeurosci
1998,10(5):1608–1612.
71. SelionovVA,IvanenkoYP,SolopovaIA,GurfinkelVS:Toniccentraland
sensorystimulifacilitateinvoluntaryair-steppinginhumans.
JNeurophysiol2009,101(6):2847–2858.
72. GerasimenkoY,GorodnichevR,MachuevaE,PivovarovaE,SemyenovD,
SavochinA,RoyRR,EdgertonVR:Novelanddirectaccesstothehuman
locomotorspinalcircuitry.JNeurosci2010,30(10):3700–3708.
73. NielsenJF,SinkjaerT,JakobsenJ:Treatmentofspasticitywithrepetitive
magneticstimulation;adouble-blindplacebo-controlledstudy.MultScler
1996,2(5):227–232.
74. KrauseP,EdrichT,StraubeA:Lumbarrepetitivemagneticstimulation
reducesspastictoneincreaseofthelowerlimbs.SpinalCord2004,
42(2):67–72.
75. NielsenJF,SinkjaerT:Long-lastingdepressionofsoleusmotoneurons Submit your next manuscript to BioMed Central
excitabilityfollowingrepetitivemagneticstimuliofthespinalcordin and take full advantage of:
multiplesclerosispatients.MultScler1997,3(1):18–30.
76. MinassianK,PersyI,RattayF,PinterMM,KernH,DimitrijevicMR:Human
• Convenient online submission
lumbarcordcircuitriescanbeactivatedbyextrinsictonicinputto
generatelocomotor-likeactivity.HumMovSci2007,26(2):275–295. • Thorough peer review
77. HarkemaS,GerasimenkoY,HodesJ,BurdickJ,AngeliC,ChenY,FerreiraC, • No space constraints or color figure charges
WillhiteA,RejcE,GrossmanRG,EdgertonVR:Effectofepiduralstimulation
ofthelumbosacralspinalcordonvoluntarymovement,standing,and • Immediate publication on acceptance
assistedsteppingaftermotorcompleteparaplegia:acasestudy. • Inclusion in PubMed, CAS, Scopus and Google Scholar
Lancet2011,377(9781):1938–1947.
• Research which is freely available for redistribution
78. RossignolS,GirouxN,ChauC,MarcouxJ,BrusteinE,ReaderTA:
Pharmacologicalaidstolocomotortrainingafterspinalinjuryinthecat.
JPhysiol2001,533(1):65–74. Submit your manuscript at
www.biomedcentral.com/submit