And yet it moves: Recovery of volitional control after spinal cord injury

Review

article

And

yet

it

moves:

Recovery

of

volitional

control

after

spinal

cord

injury

G.

Taccola

,

D.

Sayenko

,

P.

Gad

,

Y.

Gerasimenko

,

V.R.

Edgerton

*

a_{DepartmentofIntegrativeBiologyandPhysiology,UniversityofCalifornia,LosAngeles,CA90095USA} b

NeuroscienceDepartment,InternationalSchoolforAdvancedStudies(SISSA),Bonomea265,Trieste,Italy

c

PavlovInstituteofPhysiology,St.Petersburg199034,Russia

d

DepartmentofNeurobiology,UniversityofCalifornia,LosAngeles,CA90095USA

e

DepartmentofNeurosurgery,UniversityofCalifornia,LosAngeles,CA90095USA

f

BrainResearchInstitute,UniversityofCalifornia,LosAngeles,CA90095USA

g_{TheCentreforNeuroscienceandRegenerativeMedicine,FacultyofScience,UniversityofTechnologySydney,Ultimo,2007NSW,Australia} h

InstitutGuttmann,HospitaldeNeurorehabilitació,InstitutUniversitariadscritalaUniversitatAutònomadeBarcelona,Barcelona,08916Badalona,Spain

ARTICLE INFO Articlehistory: Received9March2017

Receivedinrevisedform9October2017 Accepted21October2017

Availableonline2November2017 Keywords: Neuromodulation Electricalstimulation Spinalnetworks Motortraining ABSTRACT

Preclinical and clinical neurophysiological and neurorehabilitation research has generated rather surprisinglevelsofrecoveryofvolitionalsensory-motorfunctioninpersonswithchronicmotorparalysis followingaspinalcordinjury.Thekeyfactorinthisrecoveryislargelyactivity-dependentplasticityof spinalandsupraspinalnetworks.Thiskeyfactorcanbetriggeredbyneuromodulationofthesenetworks withelectricalandpharmacologicalinterventions.Thisreviewaddressessomeofthesystems-level physiologicalmechanismsthatmightexplaintheeffectsofelectricalmodulationandhowrepetitive trainingfacilitatestherecoveryofvolitionalmotorcontrol.Inparticular,wesubstantiatethehypotheses that:(1)inthemajorityofspinallesions,acriticalnumberandtypeofneuronsintheregionoftheinjury survive, but cannot conduct actionpotentials, and thus are electrically non-responsive;(2) these neuronalnetworkswithinthelesionedareacanbeneuromodulatedtoatransformedstateofelectrical competency;(3)thesetwofactorsenablethepotentialforextensiveactivity-dependentreorganization ofneuronalnetworksinthespinalcordandbrain,and(4)propriospinalnetworksplayacriticalrolein drivingthisactivity-dependentreorganization afterinjury.Real-timeproprioceptiveinputtospinal networksprovidesthetemplateforreorganizationofspinalnetworksthatplayaleadingroleinthelevel ofcoordinationofmotorpoolsrequiredtoperformagivenfunctionaltask.Repetitiveexposureof multi-segmental sensory-motor networks to thedynamics of task-specific sensoryinput as occurs with repetitivetrainingcanfunctionallyreshapespinalandsupraspinalconnectivitythusre-enablingoneto performcomplexmotortasks,evenyearspostinjury.

©2017TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Contents

1. Introduction ... 65

2. PrognosisofrecoveryfromaSCIinthe“chronicstate” ... 65

3. Greater-than-expectedrecoveryinthechronicstateinresponsetospinalneuromodulation ... 65

4. Theconceptualcoresofautomaticityofmovement ... 66

4.1. Centralpatterngeneration ... 66

4.2. Brainstemstimulation ... 66

4.3. Sensorycontrol ... 67

4.4. Propriospinalsystemasaninterfacebetweenthebrainandthespinalcord ... 67

4.5. Propriospinal“steppingstrip”thatcanfacilitatestepping ... 68

Abbreviations:CPG,centralpatterngenerator;DLF,dorsallateralfuniculus;DR,dorsalroot;eEmc,electricalEnablingmotorcontrol;ER,earlyresponse;MR,middle response;LR,lateresponse;SCI,spinalcordinjury;SCS,spinalcordstimulation;tSCS,transcutaneouselectricalspinalcordstimulation.

*Correspondingauthorat:DepartmentofIntegrativeBiologyandPhysiology,UniversityofCalifornia,LosAngeles,TerasakiLifeSciencesBuilding61CharlesE.YoungDrive East,LosAngeles,CA,90095-1527,USA.

E-mailaddress:vre@ucla.edu(V.R. Edgerton).

https://doi.org/10.1016/j.pneurobio.2017.10.004

0301-0082/©2017TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

ContentslistsavailableatScienceDirect

Progress

in

Neurobiology

5. Strategiesandmechanismsforelectricallyneuromodulatingspinalmotorfunction ... 68

5.1. Dorsalroots(DRs) ... 69

5.2. Longitudinalspinalfibersoftheposteriorcolumn ... 70

5.3. Propriospinalnetwork ... 70

5.4. Electricalstimulationmodulatesthechemicalenvironmentofthespinalcord ... 71

5.5. Othertargets ... 71

6. Subthresholdelectricalstimulation ... 72

7. Activity-dependentplasticity ... 72

7.1. Electricallyenablingnon-responsivespinalsensory-motornetworks ... 73

7.2. Reconnectivityofsupraspinal-spinalnetworksviatransformationofelectricallynon-responsivetofunctionallycompetentcellular elements ... 73

8. PossibleinterpretationofthevoluntaryrecoveryinducedbyeEmc ... 75

9. Synergismofneuromodulationandproprioception ... 75

10. Conclusion ... 76

Authordisclosurestatement ... 76

Acknowledgments ... 76

References ... 76

1.Introduction

A number of observations demonstrate the presence of potentiallyfunctioning fiberstravelling acrossthelesion in the majority(84%)ofpeoplewithclinicallydiagnosedcompletespinal cordinjury(SCI)(Sherwoodetal.,1992;Dimitrijevicetal.,1984, 1983).Inaddition,therehavebeenseveralreportsofpostmortem anatomicalevidenceofa significantnumberofaxonstraversing thelesionofindividualsthathavebeencompletelyparalyzedfor years prior to death(Kakulas, 1988). The significance of these observationshasbecomemoreapparentgiventherapidrecovery ofvoluntarycontrolofmuscles belowlesion inindividualsthat have been paralyzed for more than a year before receiving locomotor/standingtrainingandepiduralelectricalstimulation(a neurorehabilitative protocol overall named electrical Enabling motor control; eEmc). It seems imperative, based on these observations, that our current knowledge on the mechanisms underlying the recovery of voluntary movement after chronic, completeparalysismustbere-examined.

2.PrognosisofrecoveryfromaSCIinthe“chronicstate” Allpersonswithaspinalcordinjuryexperiencesomeextentof recoveryfromthefirstclinicalassessment.Thisnormallyoccursin the first three months after the lesion, but recoveries can be observedfor uptoone year,andoccasionallyaftereven longer periods(Watersetal.,1992;Kirshblumetal.,2004;Fawcettetal., 2007). These recoveries are broadly termed “spontaneous”, althoughpartofthem isprobablydue tocurrentearly surgical procedures,standardpharmacologicaltreatmentsandtraditional physicaltherapies.

Extension of recovery and consequent functional improve-mentswereinverselyrelatedtotheseverityofinjury(completevs. incomplete;Marinoetal.,1999).Indeed,theextentofspontaneous recovery was significantlygreaterfor incompletelesions,while onlyabout10%ofpersonswithaclinically-definedcompeteinjury within the first 15days post injury became motor incomplete within12monthspostinjury(Spiessetal.,2009).Amongthese,as apercentage,personswithtetraplegiademonstratedalmosttwice asmanyrecoveriesthanthosewithparaplegia(Kirshblumetal., 2004;Fisheretal., 2005).Moreover,these improvementsoften referonlytothesegmentslocatedimmediatelybelowthelesion, and can very rarely bring to real functional benefits, such as independentstandingandstepping(Fawcettetal.,2007).

Paraplegicsubjectswithacompletespinallesionhaveascarce probability of motor recovery after six months from injury, thereforeanextremelylowsamplesizewillhavesufficientpower

toexaminetheeffects ofa newexperimental interventionasa valuabletoolforrecovery(sixtoeightparticipantsaccordingtothe Guidelinesfortheconductofclinicaltrialsonspinalcordinjury,as developedbytheInternationalCampaignforCuresofSpinalCord InjuryParalysis–ICCPpanel;Fawcettetal.,2007).

Moreover,whenconsideringevenlongerperiodselapsingfrom the time of injury, and thus stable clinical conditions and less hospitalization, the recoveries can be more properly defined “spontaneous”andtheirappearanceisevenrarer.Indeed,afterone yearlessthan2%ofcompletespinalcordinjurieswerereportedto have become incomplete by the fifth year after the lesion (Kirshblumetal.,2004).Consequently,thenumberofparticipants with a chronic and complete lesion necessary to show a statistically significant motor gain due to some experimental therapybecomesevensmaller(Fawcettetal.,2007).

3.Greater-than-expectedrecoveryinthechronicstatein responsetospinalneuromodulation

A first case report (Harkema et al., 2011) relates to an individual with a motor completeSCI at the level of the first thoracicsegmentsformorethanayear,whowasimplantedwith anelectrodearrayintheepiduralspaceatthelevelofthelumbar spinalsegments(L2–S1).Afterafewmonthsofstandtraining, the subject could maintain continuous minimally assisted full weight-bearingstandingforuptofourminutes.Furthermore,this person revealed significant levels of voluntary control of the lowerlimbsafteralmostsevenmonthsofdailystimulationand stand/step training.It was speculated thatthislong posttreat-mentperiodleftopenthepossibilityforapotentialregrowthof functional supraspinal-sublesional spinal connections attribut-able to axonal growth through the lesion. When this neuro-rehabilitative protocol eEmc was repeated in three other chronicallymotorcompletespinallyinjuredsubjects,the recov-ery of voluntary movement appeared within a few weeks of stimulationandtraining(Angelietal.,2014).Unlikethestanding and stepping recovery, which can be explained from the perspectiveoftheintrinsic spinalautomaticity, therecoveryof voluntary control requires a critical level of functional supra-spinal-sublesional spinal reconnections. This hypothesis is furthersupported bythefact thatparticipantsinthesestudies have continued to improve years after implantation. Similar results as for regaining of voluntary control during rhythmic locomotormovementshavealsobeenreportedinfiveindividuals with SCI, treated with transcutaneous electrical stimulation (Gerasimenkoet al.,2015).Morerecently,anadditionalperson with chronic paraplegia regained volitional control of

task-specificmuscleactivityafterfew sessionsof epiduralelectrical stimulation(Grahnetal.,2017).

Overall,alltensubjectstestedwithamotorparalysis,whowere treated with spinal electrical stimulation (average time since injuryof3.11.2years),regainedvoluntarymovementoflower limbs (Table 1). These results are far beyond the spontaneous recoveryreportedafterchronicSCI,pointingoutthatthisstrategy, basedonthedatareportedsofar,ranksamongthemostpromising onesfortherecoveryoflocomotioninspinalcordinjuredpersons. Asnewdevelopinginterventionssuchasbrainmachineinterfaces, stemcellimplantationandroboticsemerge,itseemslikelythat multipleapproaches can becombined toaccommodatea wide rangeofneedsfollowingawidearrayofdysfunctions.

4.Theconceptualcoresofautomaticityofmovement 4.1.Centralpatterngeneration

Decades ofexperimental workonanimalpreparationshave resulted in a greatly improved level of understanding of the functional organization of a locomotor spinal network called Central Pattern Generator (CPG; Grillner and Zangger, 1975; GrillnerandRossignol,1978;Lovelyetal.,1986;ButtandKiehn, 2003). A CPG is an ensemble of neurons that can generate a rhythmicmotoroutputintheabsenceofanyrhythmicexternal stimuli. This output is characterized by a highly coordinated activation pattern of motoneuronal pools. But the most impressivefeature of the output of whatare presumed tobe CPGnetworksistheirabilitytoprocess complexcombinations of proprioceptive and tactile input in real time in in vivo conditions.SomeobservationssuggestthattheCPGdevelops,at least to some degree, with a deterministic strategy in what seems to be a well-defined number of genetically-targeted interneurons,startingfromtheembryonicstagethatestablishes distinct connections within the network (Pierani et al., 2001; Kullanderetal.,2003;Lanuzaetal.,2004;Gosgnachetal.,2006; Kwanet al.,2009;Wilsonet al.,2010;Zhongetal., 2011).This complexandspecificconnectivityamongthedifferentelements of the network has been only partly elucidated. Nevertheless, even if the rhythmogenic source of the pattern is isolated in vitro, and thus deprived of its descending and afferent inputs,

the rhythmic pattern can still be triggered bya wide range of electrical stimuli (Atsuta et al., 1990; Magnuson and Trinder, 1997;Marchettietal.,2001;StraussandLev-Tov,2003;Taccola, 2011; Dose et al., 2013; Dose and Taccola, 2016) and pharmacological agents (Cazalets et al., 1992; Houssaini et al., 1993; Kiehn and Kjaerulff, 1996; Marchetti and Nistri, 2001; TaccolaandNistri,2006)andevenbyanincreaseinextracellular K+, which can lead to a broad rise in the overall neuronal network excitability (Bracci et al., 1998). Thus, while a more exact composition and organization of the CPG evolves, it is anticipated that more precise strategies in formulating inter-ventionswillfacilitateourabilitytoexploittheautomaticityof neuralnetworks and theirplasticity.

4.2.Brainstemstimulation

The origin of another source of automaticity has been demonstrated by a series of experiments indicating thattonic stimulation at selected sites within the mesencephalon can induce highly coordinated quadrupedal stepping in place ona stationarytreadmillbeltwithfullweight-bearing.Inaddition,it canfacilitatesteppingthatcanaccommodatearangeofspeeds onamovingbeltinanacutelydecerebratedcat(Orlovskiıetal., 1966;Moriet al.,1977).Later, Moriandcolleagueschronically implanted stimulating electrodes in a neurologically intact awakecat, butinslightly differentsites ofthe mesencephalon than previouslystimulated. Following thisstimulation, thecat wouldstandupandbeginwalking,andthen,whenstimulating afewmillimetersaway,theanimalwouldstopsteppingandsit quietly(Morietal.,1978,1989).Furthermore,theseresultsalso demonstrate that the activation patterns of motor pools and body muscles that control standing and stepping can be triggered by a very simple tonic stimulus applied to some unspecificgroupof neurons.Thus,a highlevelof automaticity is built within the locomotor circuitry. Combination of these observationsonthehighlycoordinated posturalandlocomotor performancethatcanbegeneratedandcontrolledbythespinal circuitry along with the proprioceptive and cutaneous input demonstrates that a major neural component of automaticity lieswithin thespinalcircuitrythatprocessescomplexsensory input andexecutes motor outputcontrols in real time.

Table1

Parametersofeachstimulatingprocedureandtheiroutcomesfordifferentindividualswithaspinalcordinjury. Person Level of injury Degreeof injury (ASIA scale) Modeof electrical stimulation Site (vertebrae) Stimulatingprotocol (frequency, amplitude,pulse width) Outcomes Refs. #1 C7-T1 B 16-electrode epiduralarray T11 L1 15–40Hz,0.5–10V, 210or450ms

Independentstanding(max4.25min),locomotor-likeEMG patternsinthelegswhilesteppingonthetreadmillwithbody weightsupportandmanualassistance,volitionalmovementsof bothlegs,autonomicgains.

Harkema etal.(2011)

#2 T5-T6 A T11 T12 25–40Hz,0.5–9V,

250,330or400ms

Locomotor-likeEMGpatternsinthelegswhilesteppingonthe treadmillwithbodyweightsupportandmanualassistance, volitionalmovementsofbothlegswithgradedlevelsofforcein responsetoaverbalcommand.

Angelietal. (2014) #3 C6-C7 B #4 T5 A #5 T3-T4 B tworound transcutaneous electrodes (2.5cm) T11+Co1 30Hz+5Hz(10kHz modulated),80– 180mA,1000ms

Volitionallegoscillationsinagravity-neutralposition. Gerasimenko etal.(2015) #6 C5-C6 B #7 C6 B #8 C7 B #9 T3-T4 B #10 T6 A 16-electrode epiduralarray T11 L1 15,25and40Hz,1.5– 5.5V,210ms

Volitionalcontroloftask-specificmuscleactivity,volitional controlofrhythmicmuscleactivitytoproducesteplike movementswhileside-lying,independentstandingandwhilein averticalpositionwithbodyweightpartiallysupported, voluntarycontrolofsteplikemovementsandrhythmicmuscle activity.

Grahnetal. (2017)

4.3.Sensorycontrol

Athird origin of automaticity reliesin the variousforms of sensoryinformationthatisconstantlyavailable,though dynami-callychangingsourcesparticularlyfromvision,hearing, proprio-ceptionand cutaneousinput, tothespinal cordand brain.The importance of sensory information for this control has been demonstrated in rats that have been paralyzed from the mid thoracicregion.Aftera few weeks ofrecoveryfrominjury, the animalscouldstepforwardandbackwardand graduallychange anglesofsteppingonatreadmillbelt(Shahetal.,2012),aswellas varyingdegreesofloadingandspeedsofstepping(Edgertonetal., 1991).Further,SCIratswerealsocapableofmodulatingthefiring patterns of flexor and extensor muscles to accommodate the changingspeedandweightbearingwhilesteppingonatreadmill (Gad et al., 2013a).Similar observations were made in human subjectssteppingonatreadmillwithbodyweightsupportand variableloadandspeed(Harkemaetal.,1997).

These results demonstrate that mainly the combination of proprioceptionandcutaneousinputsderivedfromthelimbscan serveasthesolesourceofcontrolforallmotorpoolsandmuscles neededtoperformverycomplexmovements.

4.4.Propriospinalsystemasaninterfacebetweenthebrainandthe spinalcord

Thespinalcordcontainsinterconnectingpropriospinalneurons andrelatedaxons,whichareknownoverallasthepropriospinal system(PSS;Jankowskaetal.,1974).ThePSSextendsalong the wholespinalcord,connectingventralanddorsalhorns,aswellas cervicalandlumbarenlargements,andprovidesbilateral projec-tions totheleft and rightsidesof the cord(Reed etal., 2009; Brockettetal.,2013).

Propriospinal neurons are numerically predominant in the spinalcord(Chungetal.,1984)andhaveanextremely heteroge-neous morphology (Saywell et al., 2011). Moreover, during development,theyareexposedtosuchavarietyofextracellular guidance cues (Jacobi et al., 2014) that minimize specific connectivity, thus providing a stochastic feature to the whole process of network formation. Indeed, on par with realistic neuronalcircuitrysimulationbasedsolelyontheproximityamong cells(vanOoyenetal.,2014),duringspinalcorddevelopment,a plethora of interneurons establish a great number of synaptic connections bysimplyfollowing the anatomical overlappingof theiraxonsanddendrites(Lietal.,2007).ThePSScanfunctionas an interface between cortical, subcortical and spinal neuronal networks.Indeed,propriospinalneuronscaninterconnect multi-plespinalsegmentsalongthelengthofthespinalcordviatheir axonsanddendrites(Skinneretal.,1979;Menétreyetal.,1985; Courtineetal.,2008;vandenBrandetal.,2012).Amongthese,the majorityofunmyelinatedpropriospinalaxonsarelocatedinthe dorsolateralfuniculus,whileinthedorsalfuniculiarefoundalarge number of descending myelinated propriospinal fibers (Chung etal.,1987;ChungandCoggeshall,1988).

Atlumbar level,this highly-recurrentdiffuse network func-tionallyinterfaceswithCPGneurons(CowleyandSchmidt,1997; Cazalets2005).Here,PSSfunctioningcanbeassimilatedtocortical neuronalnetworkswithabalancedratioofexcitation/inhibition (Shew et al., 2011). Indeed, just like in cortical networks, spontaneousactivityofpropriospinalnetworksmight rhythmical-lyemergeasirregular,isolatedpopulationburststhatsynaptically spreadthroughthewholenetwork,withouttheneedforaspecific wiring (Streit et al., 2001). This activity might synchronize anatomically dispersed neuronal ensembles (Pennet al., 2016) andcanbedescribedasaneuronalavalanche,namelyasequence of events initiated byexciting a network nodewhich, in turn,

triggers a subsequent cascade of excitation of nearby nodes (Larremoreetal.,2012).Furthermore,feedforwardinputstothe PSS can amplify small fluctuations into large population ava-lanches (Murphy and Miller, 2009; Benayoun et al., 2010), as demonstratedbymodelingstudies,whereexcitationand inhibi-tion in a network of thousands of neurons has been closely balanced, generating an irregularsynchronous burstingactivity (FriedmanandLandsberg,2013).

Experimentally,theactivityofthePSScanberecordedinthe dorsalspinalcordintheformoftravelingelectricalwaves(Bayev andKostyuk,1981;Nogaetal.,1995;Cuellaretal.,2009;Saltiel etal.,2016).Thesewavescanalsoberecordedatrest(Cuellaretal., 2009), suggesting that the propriospinal network continuously producesastablebackgroundpattern,reverberatingtheexcitation alongthewholespinalcord(Eblen-ZajjurandSandkühler,1997). Also in vivo multiple recordings from spinal interneurons demonstrated that someunitsare tonicallyactive even atrest, butbecomemodulatedonlyduringlocomotion(Bergetal.,2007). Thus, the backgroundactivity of propriospinal networks repre-sentsacontinuoussourceofsubthresholdfacilitatoryinputacting onspinal CPGs(Liand Moult, 2012;Warp etal.,2012).Indeed, spontaneoustonicdischargeshavebeenrecentlyrecorded(Husch et al., 2015)froma class ofexcitatory spinal interneurons that couplelocomotorcircuitsonbothsidesofthespinalcord(Crone etal.,2008).Backgroundexcitatorysynapticpotentialsmightkeep CPGexcitabilitynearthetriggeringthresholdandthusmakethe CPGmoreresponsivetovoluntaryinputsforlocomotorcontrol,but alsotoafferentphasicinputs,especiallyofproprioceptivenature. Fromafunctionalpointofview,thePSSprovidesamechanism for intersegmental coupling and coordination of motor pools withinandamongmultiplespinalsegmentsthatcontrolthelower limbs,andeventrunkmusculatureduringquadrupedal locomo-tion(Skinneretal.,1979;Rovainen,1985;KrutkiandMrówczynski, 2004;Juvinetal.,2005;Courtineetal.,2008;Shahetal.,2013; Beliez et al.,2015).For example,humanPSS contributestothe rhythmic coordination of the upper and lower limbs during locomotion (Dietz et al., 2001) and mediates some inter-limb neurological responses, such as modulation of the H-reflex amplitudeduring phasicvoluntarycontractionsofarm muscles (MazevetandPierrot-Deseilligny,1994).

PSSisalsoinvolvedinthephysiologicalcontrolofthelocomotor activity. Indeed, in addition tothe direct cortico-motoneuronal control, an indirect descending pathway establishes synaptic relays within the PSS (Kostyuk and Vasilenko, 1978; Pierrot-Deseilligny and Marchand-Pauvert, 2002; Cowley et al., 2010). Then,onceavolitionalmovementisselected,thesynapticrelays withinthePSScanamplifyandprolongthedescendingcommand throughoutthelumbosacralsegmentsofthespinalcord(Kostyuk andVasilenko,1978;Nogaetal.,1995;Rocheetal.,2012).

Since thePSSpresentsa significant levelofredundancy and plasticity,it hasbeenlinkedtosomefunctionalrecoveriesafter severe experimental lesionsof the spinal cord (de Leon et al., 1998a,1998b;Courtineetal.,2008;Shahetal.,2013;Fillietal., 2014),aswellastoanaberrantintersegmentalconnectivityalong thelevelofinjury.Forinstance,insubjectswithchroniccervical lesions,(Calancie,1991;Calancieetal.,2002)thePSSmightplaya roleinupperlimbreflexesafterstimulatingnervesontheirlower extremity. In addition, the PSS generates the propriospinal myoclonus,namelythoraco-abdominalspontaneousmusclejerks that propagate rostrally to upper intercostal muscles and to abdominalmuscles(Chokrovertyetal.,1992).

Moreover, after a spinal lesion, descending cortico-spinal pathwaysaremoreorless compromisedand unabletoactivate motorpatternswiththesufficientsupraspinally-derivedvoluntary stimuli.Atthesametime,acompromisedcortico-spinal connec-tionperturbsthedescendingdrivetowardsthePSS.Theresulting

musculararyflexiaandhypotoniacharacterizingthespinalshock phase following an acuteinjury, might also originate fromthe sudden lack of descending input, which would physiologically modulatethePSS.Inturn,toaccountforthelossofdescending inputs,thePSSundergoesaseriesofplasticrearrangementsthat canfurtheralteritsexcitability(reviewedbyFlynnetal.,2011).

At the same time, the PSS might also be affected bysome phenomena of dendrite regression, secondary to the stress of injury(MagariñosandMcEwen,1995;Chenetal.,2008),whilethe expression of several genes with putative roles in synaptic plasticityaredownregulatedfollowingSCI(Siebert etal.,2010), attheexpensesofsynapticef_ficacy.

For all these reasons,a spinal lesion may reduce PSS basal rhythmicityand,inturn,thepossibilityforsprouteddescending connections to re-establish the local propriospinal circuitry (Bareyreetal.,2004).AlowerlevelofPSSactivitymaycontribute to a lower level of CPG excitability, which would reduce the probabilityofresidualinputstofacilitatelocomotorpatterns.

Nevertheless,PSSanatomicallyandfunctionallysurviveseven after a complete transection of the spinal cord (Faganel and Dimitrijevic,1982;ContaandStelzner,2004)andcanthussupport functionalrecoveryfollowingaseverelesion,whenpropriospinal neuron excitability is modulated towards physiological values (Courtineetal.,2008;Cowleyetal.,2015).

4.5.Propriospinal“steppingstrip”thatcanfacilitatestepping Shik and colleaguesina series of experimentson mesence-phalic catsdescribed a “stepping strip” extending through the dorsallateral funiculus (DLF)from thecervical tolumbar level (Kazennikovetal.,1983a).Ithasbeensuggestedthatthe“stepping strip” is a continuation of pontomedullary locomotor strip (Kazennikovetal.,1983b).Stimulationofthe“steppingstrip”at 60Hz canevokesteppingmovements initially intheipsilateral limb, and with increased intensity of stimulation, in the contralateralhindlimbandforelimbsaswell.Usinglocallesions of DLF at different levels of the spinal cord, they found that stimulationoftheDLFcanelicitsteppingonlywhenatleast8to12

segmentswereintact(Kazennikovetal.,1983b).Basedonthese findings, it was suggested that propriospinal neurons must be activatedtofacilitatesteppingduringstimulationofthe“stepping strip” (Kazennikovet al.,1990).Interestingly,stepping couldbe inducedduringDLFstimulationatthoraciclevelevenafterdamage ofDLFatcervicalandupperlumbarlevels,suggestinganindirect mechanismofactivationofthelocomotorcircuitry(Kazennikov andShik,1988).Indeed,itwasfoundthattheneuronsresponding toDLFstimulationwerelocalizedataborderbetweenlateraland ventralfuniculus(Kazennikovetal.,1985).

Shik (1997) hypothesized that stimulation of the “stepping strip_” distinctivelyexcitesthespinal neuronsthatprojecttothe locomotornetwork,bysendingtheiraxonstotheVLF,andforthis reasonaretermedVneuron(Fig.1).Criticalforsuchahypothesis could beexperiments with verification of V neurons and with demonstrationoftheirdirectinfluenceonlocomotornetwork. 5.Strategiesandmechanismsforelectricallyneuromodulating spinalmotorfunction

At present, individuals with a complete paralysis due to a severe SCI are considered to have virtually no potential of functional recoveryafter severalmonths post-injury.However, this dogma is challenged by animal and human experiments demonstrating the recovery of sensory-motor and autonomic functionswithelectricalspinalcordstimulation(SCS;Gadetal., 2013b, 2015; Gerasimenko et al., 2015). These same observa-tionsprioritizetheneed tounderstandthemechanismsof the neuromodulatory strategies that are so far largely based on modeling, eventually corroborated by experimental studies (Struijk et al., 1993; Capogrosso et al., 2013). In turn, these theoreticalstudiesweremainlyconceivedfromtheassumption thatneuralnetworksrespondtoelectricalstimulationsolelyby inducingaction potentials,and that the biophysicalproperties of the axons follow Ohms law, i.e., the larger axons have the lowestexcitatorythresholds.Nevertheless,recentexperimental results challenge these assumptions, indicating thatcontrol of neuralnetworksin theinjuredsystemmaybeequallyrelevant

Fig.1.Wiringdiagramsuggestingtheroleofpropriospinalandreticulospinalsystemsininitiationofstepping.

Spinalcordincludesthedorsalandventralpartsofthelateralfuniculus(DLFandVLF)andgraymatter(GM).ThespinalneuronsthatsendtheiraxonstoDLFaretermedD neurons,whileVneuronssendtheiraxonstoVLF.MLR,mesencephalic“locomotorregion”;HB,hindbrain;thoracic(Th)andlumbar(L)segmentsofthespinalcord.;IS, interneuronsassemblingstepping;M,motoneurons;RS,reticulospinalneurons;St,steppingstrip(modifiedfromShik,MotorControl,1997).

in the uninjured system (Gerasimenko et al., 2006; Bergquist et al., 2011;Sayenko et al., 2015).

Giventhatthetermneuromodulationisusedtorefertoawide rangeofneuralphenomena,inthepresentmanuscriptweusethis termtoindicateachangeinthefunctionalstateofexcitabilityofa neuronorcombinationofneuronswithinneuronalnetworks.

There also seems to be largely similar mechanisms in the different methods of neuromodulation of spinal networks. For example, transcranial magnetic stimulation is a tool used to noninvasivelystimulatethehumanbrainandspinalcord(Barker et al., 1985; Hallett, 2000, 2007; Gerasimenko et al., 2010). Transcutaneouselectricalspinalcordstimulation(tSCS)isanother non-invasivetechniqueforengaginglocomotor-relatedcircuitries in human. This method utilizes unique painless stimulation waveforms,which aretransmittedviaelectrodes placed onthe skinoverthespineandpresumablytravelthroughDRstoactivate thespinalcircuitry(Fig.2).Observationsfromtheseexperiments are consistent with the concept of specific modulation of the networks projecting to distinct combinations of interneurons coordinatingthemotorpools’recruitmentduringastepcycle.A third neuromodulatory intervention which shows considerable potentialinrecoveringfunctionafterparalysisispharmacological (Pearson and Rossignol 1991; Rossignol et al., 2001; Musienko etal.,2011).Althoughthereareundoubtedlyuniquefeaturesof eachofthesetechniques,themaindifferenceappearstobewhich neurons,axons and networksare more readily modulated.The focusofthepresentreviewisonthebasicbiophysicalpropertiesof membranes,axons,dendrites, cells andneuronal networksthat underlie neuromodulatory interventions. We propose that the main topics of neuromodulation of the spinal cord are the identificationofthemechanismsthatdriveresponses.

In the following sections, we have examined some of the elementsthatareimportantinunderstandingthemechanismsof

electricalneuromodulationofspinalneuronalnetworksandtheir interactionwithsupraspinalandperipheralsensoryinputs.One crucialquestiontoaddressis:Whatsub-cellular,cellular,networks components of the neural networks respond to the numerous combinationsofneuromodulatoryparametersused?

5.1.Dorsalroots(DRs)

Electrical SCS, witheithertranscutaneous, epiduralor intra-spinal electrodes, facilitatesfunctionalstanding and walkingin individualswithachronicSCI.Inmanycases,thiseffecthasbeen almostexclusivelyascribedtothedirectmodulationofposterior roots(Struijketal.,1993;Murgetal.,2000;Rattayetal.,2000; Danneretal.,2011;Capogrossoetal.,2013;Minassianetal.,2016). DRactivationduringepiduralelectrical stimulationhasbeen effectively proven by antidromic potentials recorded from peripheralnerves(Suetal.,1992).Onpar,alsotheantidromically driven cutaneousvasodilationprovides an indirectproofof DR involvement(Linderothetal.,1989),althoughinthislattercase, alsotheactivationofsympatheticefferentsplaysan activerole (Croometal.,1997).

Afferent inputs, as the ones generated by posterior root stimulation, reach the locomotor networks (Hultborn et al., 1998; Rybaketal.,2006;Buiand Brownstone,2015), asclearly demonstratedbyepochsoflocomotorpatternselicitedinvitroby theselectiveelectricalstimulationofsingleormultipleDRs,using tightsuctionelectrodes(Marchettietal.,2001;Straussand Lev-Tov,2003;Taccola,2011;Doseetal.,2016;DoseandTaccola,2016). Nevertheless,electricalSCSdoesnotonlyactivateDRs.Indeed, the epiduralstimulation technology currentlyused tofacilitate locomotionandstandingafterlesioncomespartlyfromthefieldof neuropathicpainmanagement,whichhoweveraimsat minimiz-ing DR stimulation (Molnar and Barolat, 2014). This is

Fig.2.Facilitationofstepping-likevolitionaloscillationsusingnon-invasivetranscutaneouselectricalspinalcordstimulationinSCIsubject.

(A)Positionoftheparticipantinthegravity-neutralapparatus.(B)Biphasicelectricalstimulationwasdeliveredusinguniquewaveformsconsistingof0.3–1.0msburstsfilled by10kHzfrequencythatwereadministeredat5–40Hz.(C)EMGactivityofrightsoleus(RSol),righttibialisanterior(RTA),rightmedialgastrocnemius(RMG),right hamstrings(RHam),rightvastuslateralis(RVL),rightrectusfemoris(RRF)andangulardisplacementinthekneeandhipjointsofbothlegsduringlegoscillationswitha voluntaryeffortalone(Vol),stimulationatT11(Stim),andVol+Stimareshown.(D)Schematicsdemonstratingtheapproximatelocationoftranscutaneouselectrodesabove thelumbosacralenlargement,inrelationtothelocationofthemotorpoolsbasedonKendalletal.(1993)andSharrard(1964).

fundamentally antithetical to the purpose of finding the best simulationpatterntogeneratepostureandlocomotion(Sayenko etal.,2014;Rejcetal.,2015).

Furtherevidencethat electricalSCSrecruitsotherstructures apartfromDRscomesfrompreviousreports,whereepiduralSCSat lowamplitudeappliedtohigherthoracicvertebraeantidromically activated ascending fibers in the dorsal columns, with the appearance of DR potentials from the underlying DRs. On the otherhand,athigherintensities,alsothedescendingmotortracts wereactivated, generating responses, from the lumbar ventral roots and fromhindlimb muscles, which did not change after dorsallumbarrhizotomy(Haghighietal.,1994).

AclearindicationoftheinvolvementofothertargetsofSCSin thespinalcordcomesfromthemeasurementofintraspinalcurrent density during epiduralstimulation in isolated human cadaver cords,demonstratingthat,despitearapidcurrentdecreaseaway fromelectrodelocation,asubstantialamountofappliedcurrent inevitablypassesthroughtheentirespinalcord(Swionteketal., 1976).Althoughthisstudycannotexplainwhatneuralstructures areexcited,thisobservationprovides insightsontheshape and distributionoftheelectricalfieldgeneratedbyelectricalSCSofthe spinal cord, showing that many neural structures are directly impactedby theelectrical field: axons, synapses, neuronal cell bodies,glialcellsandcerebrospinalfluid.

Amongthese,electricalSCSmainlyrecruitstype-Amyelinated fibersandaxons.(Ranck,1975;Holsheimer,2002).Nevertheless, whenaxonsareorientedinthesamedirectionasthefield,they becomepolarizedand theirsynapticextremities,once depolar-ized,canmodulatethe integration ofincoming synapticinputs (Jankowska,2017).Forexample,polarizingcurrents,passingacross thecatspinalcordinadorsal-ventraldirection,increasedsynaptic excitatorypotentialselicitedinmotoneuronsbyafferent stimula-tion(Ecclesetal.,1962).Accordingtothisselectivityprovidedby fibers’ orientation within spinal networks, both excitatory and inhibitorysynapsescanbeequallydepolarized.Nevertheless,the physiologicalcontributionofexcitatoryinputsprevailsatthelevel oftherhythmogeniccoreoftheCPG(Grillner,2006)and,thus,the neteffectwouldbeanincreasedexcitabilityofthewholenetwork, withtheconsequencethatevenweakstimuli(e.g.afferent),can activatelocomotorpatterns,especiallycutaneousand propriocep-tiveinput,aswellasascending(Etlinetal.,2010;Gadetal.,2013a) anddescendinginput.

5.2.Longitudinalspinalfibersoftheposteriorcolumn

Stimulationoftheresidual,longitudinalaxonsofthespinalcord belowinjurymightfacilitatespinallocomotornetworks. Never-theless, activation of longitudinal pathways, ascending and descending,isquitecontroversial.Indeed,sometheoreticalstudies indicatethatthethresholdforactivatingDRsthroughelectricalSCS islowerthantheonerequiredtoactivatelongitudinalpathways (Danneretal.,2011).However,electricalSCSdoesseemtomainly recruitlongitudinalpathwaysrunningalongtheposteriorcolumn fibers,accordingtoadditionalcomputationalstudiesthatconsider fibers' curved trajectory, inhomogeneity and anisotropy of the conductivemediatheytraverse,andtheirorientationrelativeto theplacementofthestimulationleads(Coburn,1985;Struijketal., 1993).Moreover,theneuromodulatoryeffectofelectricalSCSin alleviatingneuropathicpain,withoutinducinganyuncomfortable motorreflexes,hasbeenmainlyascribedtotheselectiveactivation oflongitudinaldorsalspinalpathways,withaminimalrecruitment ofDRs(MolnarandBarolat,2014).

Similarly, stimulation of residual, longitudinal fibers below injuryisconsideredasthemainmechanismforelectricalSCSto modulatespasticity (Gybels and van Roost,1985; Waltz, 1998; Minassianetal.,2012), which alsocorrespondstoanimproved

locomotor performance in persons with a partial injury ( Hof-stoetteretal.,2014).

Itisstillpossibletoactivatedescendingfibers,bymodifyingthe stimulationsettingsappliedtothespinalcord.Forexample,this couldbedoneby modifyingtheelectrode geometryutilizing a tripole witha centralcathodeor bylongitudinallyaligning the electrical field with the spinal cord, that is, by reducing the distance between the anode and cathode (Holsheimer and Wesselink,1997).Duetothis partial selectivity,it seemslikely that electrical SCSprotocols used for motor recovery after SCI wouldinevitablyrecruitotherresidualdescendingfibers.Among these,electricalSCSshouldactivatethedescending_fibers,witha motorsignificance,runningalongthepostero-lateraltract,suchas thelateral corticospinaltract andthe rubrospinaltract.In post mortem samples collected from SCI survivors, the preserved continuityofaportionoflateralandposteriorwhitematterwas widely reported (Kakulas, 1984), even in lesions diagnosed as ‘clinically'complete(Kakulas,1988).

5.3.Propriospinalnetwork

Activationofpropriospinalneurons,elicitedbyboth descend-ing commands of voluntary movement attempts and electrical stimulationofthemorerostralspinalsegments,canbepropagated caudal toa spinal lesion and activate the locomotor networks belowthelesion(Yakovenkoetal.,2007).Infact,inadditiontothe directactivationofthedescendingpropriospinaltractprojecting topostural and locomotor networks,SCSfacilitateslocomotion alsothroughanonspecificincreaseintheoverallexcitabilityofa widerangeofspinalnetworks,whichallowotherinputstoactually initiateandcontrolthemovement,thusbringingtotheenabling “concept”,namelyenablingspinalcircuitriestoprocess proprio-ceptiveanddescendingsupra-spinalinput.Theseotherinputsare generatedbytheremaining,oralternativelyreorganizing, supra-spinaldescendinginput,aswellasbytheextensiveproprioceptive andcutaneousinputtriggeredbymovements.Indeed,relatively prolonged periodsofSCSatlow intensityprovedtoenablethe spinalcomponentsthathadbeenelectricallynon-responsivefor over a year. Nevertheless, it is not determined whether the restoration of functional supraspinal-spinalconnectivitycan be strictlyattributedtopropriospinalnetworks.Still,itisclearthat theactivationpatternsnecessarytoperformawiderangeofhighly coordinatedmotortaskscanberestoredbyanincreaseinthenet excitabilitylevelsofoneormoretypesofspinalnetworks(Shah etal.,2012,2013;Gadetal.,2013b;Gerasimenkoetal.,2015).This netincreaseinexcitabilityoftheposturalandlocomotornetworks augments the probability of complementary supraspinal and peripherallyderivedsensoryinputstoexceedthethresholdforthe motor pools to generate movement, whether it is intentional (supraspinally)ormore“automatically”derivedfrom propriocep-tiveandcutaneousinputs.Thismodulatorystrategyenablesthe locomotornetworkstogeneratehighlycoordinatedandcomplex movements because spinal networks can accurately process constantlychangingvolleys ofafferent inputsand generatethe motorpattern‘intended’bythatsensoryensemble.

From this point of view, the highly diffused divergence of sensoryfiberstolargepopulationsofspinalneuronsonmultiple spinal segments, provides a comprehensive mechanism for monitoring the kinetics and kinematics of micro and macro mechanicalevents,whichenableshighlysystemiclevelsofcontrol (Gerasimenkoetal.,2015).Forinstance,whenwestand,virtually allsensorimotorsystemsareactivatedandhighlycoordinated.Asa consequence,adynamicandminimallyassistedstandingposture of a subject with compromised sensory feedback and motor controlofthetrunkandlowerlimbsmaybeamoredynamicand task-specific experience(Sayenko et al.,2010, 2012)than even

locomotor training. Indeed, a sudden postural perturbation triggersmultipleattempts torecoverbalance,which can result inanextensivesourceofcompellingdescendinginputsandina higher modulation frequency compared to inputs evoked by stereotypedsteppingonatreadmill,withpassiveweightbearing. 5.4.Electricalstimulationmodulatesthechemicalenvironmentofthe spinalcord

Neurotransmitterand neuromodulatorrelease from neurons (Raiteri,2006)andglia(Tawfiketal.,2010)isinducedbyelectrical stimulation.Indeed,someexamplesofelectricalSCS(30_–90min) modifying the concentrations of neurotransmitters and neuro-modulatorsinthespinalcord,areglycine(Simpsonetal.,1991), GABA (Linderoth et al.,1994), serotonin(Linderoth etal.,1992; Francketal.,1993; Songet al.,2009)and catecholaminelevels (LevinandHubschmann,1980;Liuetal.,2008).

On the other hand, the application of several exogenous pharmacologicalagentswas abletoactivatelocomotorpatterns andcanprovidearathervariegatedneuralmodulatoryeffecton spinalsegments,tolargelycontrolstandingandstepping(Douglas etal.,1993;deLeonetal.,1999;Rossignoletal.,2001;reviewedby

Alfordetal.,2003;Gerasimenkoetal.,2007;Ichiyamaetal.,2008; Fouadet al., 2010; Musienko et al.,2011; Weiet al., 2014). As mentionedabove,thelocomotorpatternscanbeactivatedbymany anddifferentsubstances,andthisscenarioisfurthercomplicated byexperimentsshowingthataspecificpharmacologicalblockage of distinct receptor subtypes was able to halt the hindlimb stepping induced by epidural stimulation, although the same specificpharmacologicalblockwasreadilybypassedsimplybythe passivemovementofforelimbsinasteplikemotion(Gerasimenko etal.,2009).Thisisacleardemonstrationthatthesamemotortask canbegeneratedbydifferentandindependentreceptorsystems. Nevertheless,inhigherprimates,includingman,ithasbeenharder to evoke similar patterns with the sole application of neuro-chemicals(Fedirchuketal.,1998).Pharmacologicalfacilitationof locomotioninindividuals withanincompletespinalinjury was obtainedthroughpharmacologicalcocktails,suchasclonidine+ cyproheptadine.Consequent improvements in walkingabilities wereassessedbothasanincreasedmaximalbeltspeedatwhich individuals manage treadmill stepping associated with more phasickinematicandEMGpatterns,aswellasareducedspasticity and an ability for some of them to functionally ambulate overground (Funget al.,1990; Normanet al.,1998). Moreover, the combination of serotoninergic and dopaminergic agents (buspirone+levodopa+cardidopa) is currently under clinical investigation (NCT01484184, clinicaltrials.gov). In addition, the cocktailswiththemostclinicalinterestforfacilitatinglocomotion often include pharmacological agents, such as clonidine and cyproheptadine, relatively non-selective for a single receptor subtype(Alexanderetal.,2015).Thismightpossiblyindicatethat eitherthereceptortargetshavenotbeencompletelyidentifiedor thatawiderrangeofreceptorsneedstobeactivatedinthespinal cord.

Nevertheless,thesystemicadministrationof neurochemicals seemstobelesseffectivethanelectricalSCS.Thismightbethecase forseveralreasons.Firstofall,electricalSCSisnotsingle-agent selective, but simultaneously facilitates the release of a great varietyofneurotransmittersandneuromodulatorsbecauseofthe differentsynapsesencounteredbytheelectricalfieldinthespinal cord. Secondly, electrical SCS induces a distinct increase of transmitters specifically at the level of the synaptic milieu of multiple network sites. Thirdly, as opposed to the stable extracellularconcentrationsobtainedwiththeadditionof exoge-nous agents, release induced by phasic electrical SCS can be patterned (Hentall et al., 2006) and may be able to exploit

mechanisms of synaptic potentiation, also limiting fatigue or receptordesensitization.

Throughthisorchestratedtuningofmultipleneurotransmitters and neuromodulators, each one with its specific release site, electrical SCS might mimic more accurately the physiological changes inthechemical environmentof thespinalcordduring locomotionthantheexogenousapplicationofarelativelysimple cocktailofpharmaceuticalagents.

However,alargenumberofanimalstudiessuggesttheneedfor athoroughexaminationofthepotentialofdifferentcombinations ofinhibitoryandexcitatoryreceptorsystems,toprovideevenmore evidenceontheef_ficacyofcombiningapharmacologicalapproach withtheneuralmodulatorystrategiesthatuseelectrical stimula-tion.Thispathseemstobeonlyatthebeginning,althoughinvitro studiessuggestthatpharmacologicalneuromodulationofspinal locomotorcircuitsismuchmorevariegatedthantestedtodatein preclinicalmodels.

5.5.Othertargets

Althoughlocalizedfarfromthedorsalstimulationsiteawide range of spinal interneurons can be reached by electrical stimulation(Kjaerulffetal.,1994;Cazaletsetal.,1995;Kjaerulff andKiehn,1996;CowleyandSchmidt,1997;KremerandLev-Tov, 1997),andtheirmembranescanbeprogressivelydepolarizedby electricalfieldswithouttriggeringactionpotentials.Nevertheless, spinal interneurons canstill providea higher level of network excitabilitythatcanfacilitateorenablearhythmicpatternwhen reached by further excitation from proprioception or by a voluntaryattempttoperformaspecificmotortask.

WhileelectricalSCSdeliveredatlowerintensitiescanengage thespinalcircuitrythroughDRs,whichgenerallyhavethelower activation thresholds, SCS at higher levels of stimulation can directlyrecruitmotoneurons,andtheirexcitationmaybemainly associated with the activation of longitudinally oriented fiber systems in the dorsal column (Hunter and Ashby, 1994). Furthermore,highvoltagestimulationofthelowerlumbar cord canbeusedforassessingventralrootfunctionality,forthedirect activationofmotorfibresastheyexitthespinalcord(Royetal., 2012).It isworthnoting thatDRactivationathighstimulation intensitiesisinevitableandshouldbeconsideredwhendiscussing potentialmechanismsandtargetedneuralstructuresduringSCS. Electricalstimulationcanalsorecruitglialcells(Tawfiketal., 2010; reviewed by Vedam-Mai et al., 2012). Among these, astrocytes represent a numerous component within thespinal cord, each of them tightly connected to one another by gap junctions toformanextendednetwork(Fellin,2009).Although glialcellsaretraditionallyconsiderednon-excitable,i.e.,unableto generate actionpotentials, theirdepolarization is coupledwith variationsintheirintracellularcalciumconcentrations(Kimetal., 1994). Thus, calcium waves propagate rapidly within the astrocytary network (Newman and Zahs, 1997; Tawfik et al., 2010; Fleischer et al., 2015), alsoin response toelectrical SCS. Propagation of calcium waves regulates synaptic transmission alongtheneuronalnetwork,mainlythroughthreemechanisms: theneurotransmitterreuptake(Contietal.,1998;Olietetal.,2001; Pascualetal.,2005),thebufferingofextracellularionic concen-trations (Carmignoto and Haydon, 2012) and the release of gliotransmitters(Liangetal.,2006;Jourdainetal.,2007;Panatier et al., 2011; Kanget al., 2013; Tang et al., 2014).For example, electrical stimulation of astrocytes releases adenosine (Caciagli et al.,1988) and, in turn,endogenous adenosine of glial origin (ActonandMiles,2015)modulatesrhythmicmotorpatterns(Dale andGilday,1996).

Otherneuralfactorscancontributetofacilitatingthelocomotor rhythm.Inparticular,electricalSCSinevitablyspreadstothespinal

canal, where the cerebrospinal fluid is contained. This is characterized by an electrolytic composition that makes it significantly conducive at body temperature (Baumann et al., 1997).Electricfieldsdistributedthroughthecerebrospinalfluid mightreachtheventrolaterallaminaeadjacenttothecentralcanal (Hoppenstein,1975),evenofsegmentsrostralandcaudaltothe stimulationsite.Noteworthy,aroundthecentralcanalarelocalized aclassofinterneuronscrucialfortheexpressionofthelocomotor pattern(Huangetal.,2000;Tillakaratneetal.,2014;Duruetal., 2015;Jalalvandetal.,2016).

6.Subthresholdelectricalstimulation

Atrest,themembraneofneuronsiselectricallypolarizedasa consequenceoftheunbalancedconcentrationofionsacrossthe membrane, which acts as a selective barrier. These variations generateamembranepotentialthatisaffectedbyanyelectrical field ableto reduce(depolarize) or increase (hyperpolarize) its magnitude(KufflerandNicholls,1976).

Inexcitablecells,depolarizationandhyperpolarization corre-spondtotwooppositeshiftsofthemembranepotential,bringingit closer(orfarther,respectively)tothethresholdfortriggeringan actionpotential.However,evensmalldepolarizations,namelythe ones unable to reach the threshold for triggering an action potential(i.e.subthreshold),canaffectthebiophysicalproperties ofneuronalmembraneandmodulatesynaptictransmission(Katz andMiledi,1967;MartinandRingham,1975).

Electrical SCS delivered at intensities unable to trigger any motor action potentials interacts with the ongoing network activity(Ozenetal.,2010).Theefficacyofnetwork recruitment using subthreshold electrical pulses is not due to a mere summationofweakstimuli,butalsoexploitsdistinctandprobably significantly underestimated mechanisms that do not parallel thoseof“threshold”stimulation.Asamatteroffact,networksare more sensitive to subthreshold electrical stimuli than single neurons,since weakelectric fieldscanactuallysynchronizethe neuronscomposingacircuit(Francisetal.,2003;Selverstonetal., 2009).

Furthermore,subthresholdimpulsesmightselectivelyactivate discreteinterneuronalpopulations,potentiallydifferentfromthe onesrecruitedbyathresholdstimulation.Asamatteroffact,the sensitivitytosubthresholdelectricfieldsisdefinedinpartbycell morphology (Radman et al., 2009). Likewise, subthreshold electrical stimuli might alsoflexibly reconfigure the functional couplingamongtheelementsofthecircuitandmodulatedistinct patternregimes(Berzhanskayaetal.,2013).

Subthreshold electrical stimulation might also contribute to amplifyspontaneousvoltagefluctuationsofneuronalmembranes, which, eventually, can also affect the tonic activity of the propriospinal network and the integration of sensory inputs. Thecombination of allthese subthreshold contributions might exploitphenomenaofspiketiming-dependentplasticitytobring someelementsofthespinallocomotornetworktothreshold(Dan andPoo,2004).

As a consequence, circuit excitability is modulated in a subthresholdmannerbysmallelectricalfields.Inturn,thesmall fieldsproducedbynetworkactivitycanmodulateexcitabilityof neighborcells,eventhosenotsynapticallyconnectedtothefiring neuronsthat producetheelectrical fields,orthose thatarenot spikingatthemoment(Francisetal.,2003).

Nevertheless,constitutivelynon-spikinginterneurons,withthe ability of sensing and transmitting any changes of the local electrical field, have been predominantly identified in inverte-brates,in partbecausetheyare technically moreaccessible for microelectrode recordings from different sites of the same interneuronorfromarelatedmotoneuron(PearsonandFourtner,

1975;BurrowsandSiegler,1978;Burrows,1980;Dickinsonetal., 1981; Wilson, 1981; DiCaprio, 2004; Smarandache-Wellmann et al., 2013; Berg et al., 2015). Networks of non-spiking interneuronsininvertebrateswerereportedtogenerateagraded modulationof synapticneurotransmitterrelease inresponse to slightchangesintheirmembranepotential(Mendelson,1971).The possibility toconvert a small depolarizationof the presynaptic terminalinto aproportional amountoftransmitter released,as opposedtothestereotypedall-or-noneresponseoftransmission based solely on action potentials, should provide an efficient meanstofinelymodulatetheactivityofevenasubsetofaxonalor dendriticbranchesofneurons.

Ascrucial elementsof invertebrate CPGs,non-spiking inter-neurons convey numerous incoming input, resulting in a high backgroundsynapticnoise(PearsonandFourtner,1975).Fortheir unique properties, non-spiking neurons represent decisional nodesofneuronalnetworks,astheyareinterposedbetweenthe source of rhythmic activityand motoneurons,thus filtering or amplifyingthepatternedCPGoutput(PearsonandFourtner,1975; Burrows,1980).

Thepresenceofneuronsthatdonotproduceanyspikesmight represent a ubiquitous principle of CNS functioning also in mammals.Thereareinfactmanyexamplesofgradedreleasein mammalsinspecificcellsoftheretinaandolfactorybulb(Hartveit, 1999;Charpaketal.,2001;Panetal.,2001;SterlingandMatthews, 2005;Snellmanetal.,2011),neverthelessnon-spiking interneur-onsinmammalianspinalcordhaveseldombeenreportedsofar (KingandLopez-Garcia,1994;Darbonetal.,2004).Regardlessof thepresenceorabsenceofnon-spikinginterneuronsinmammal SNC,asimilargradedmodulationofnetworkactivitycanbeplayed alsobyspikingneurons,asatitratedslightdepolarizationoftheir presynaptic terminals determines a graded increase in neuro-transmitterrelease(KatzandMiledi,1967;MartinandRingham, 1975; Graubard et al., 1980). Indeed, the resting membrane potential recorded immediately before an invasion of action potentialsmodulatesneurotransmitterrelease(Awatramanietal., 2005).Inmoredetail,asubthresholddepolarizingelectricalinput atpresynapticlevelmightaugmentneurotransmitterrelease,only ifoccurringsoonafteranadequatereleasinginput(Dudel,1984). Suchpotentiationofthereleaseinducedbyaweakdepolarization of the presynaptic terminal does not seem to depend upon variationsinintracellularcalcium(Dudel,1984).Rather,itmight derive from conformational changes of membrane proteins affecting the exocytotic machinery (Castro and Urban, 2009; HuangandTrussell,2011).

Insummary,intheneuromotorsystem,amechanismbasedon digital input (action potentials) suits well the rapidity and reliabilityrequiredtoimpartcommandstostartorstopgait.On the other hand, a continuum of analogue input (subthreshold depolarizations)frommultiplesensorystimuliandpropriospinal sources,anditsintegrationinrealtimewouldbeadvantageousto finelymodulateoutgoinglocomotorpatterns.

In this scenario,diffused subthreshold electrical stimulation exploitsthepropertiesofanaloguemodulationatcellularlevel. Theneteffectwouldbetofacilitatetheintrinsicrhythmogenicity oflocomotornetworks,whichcannowactivateawiderangeof physiologicalpatterns,eveninresponsetoaweaktriggeringinput. 7.Activity-dependentplasticity

Inspiteofthecomplexityofthespinalcord(frommoleculesto synapses, cells, networks and systems), there appears to be a commonunderlyingmechanism,whichallowsarelativelysimple state of neuromodulation to enable ambulatory animals and humanswithSCItorecoversensorymotorfunctions.Forexample, the application of the same common stimulating parameters,

consistinginatonicstimulationatthesamestereotypedfrequency andintensity,generatesverycomplexmovements(Ichiyamaetal., 2005).

Itisalsoevidentthatsomeofthese“plastic”eventscanoccur veryacutelyinresponsetothefirstapplicationofasingleburstof tonicstimulation,evenifadditionalchangesoccurafterexposure torepetitivestimulatingperiodswithinasingletrainingsession,as wellas furtheradaptiveeventsoccurringoveraperiod ofeven months(Gadetal.,2013b).

Whiletheroleofplasticityisfarfromclear,weusetheterm “plasticity” to embrace the reconfiguration of networks, the ancillary recruitmentof otherrhythmogenic sources aswell as theexploitationofsilentornewlysproutedconnections,withthe understandingthatweknowlittleaboutthemechanismsofanyof theseadaptivephenomena.Thesensory-motoractivitycandrive all these events by generating peripheral afferent inputs that supportand refine movements withfeedback and feedforward loopsinrealtime(Gerasimenkoetal.,2016b).

7.1.Electricallyenablingnon-responsivespinalsensory-motor networks

Alldatacollectedtodateinourlaboratoryhavebeenconsistent withthehypothesisthat,followingasevereSCI,thebaselinelevel of spinal network excitability in many cases is substantially reduced(Harkemaetal.,2011; Angelietal.,2014;Gerasimenko etal.,2015;Grahnetal.,2017).Thus,thebasicconceptunderlying neuromodulation isthat ourtonic electrical stimulationand/or pharmacologicalmodulationmovesthisbaselineexcitabilitylevel closer to the motor threshold of the networks generating the sensory-motorandautonomicfunctionsofinterest(Gerasimenko et al., 2015).Thischanges thephysiological stateof the spinal circuitry below the lesion thus requiring smaller levels of excitationfromcutaneousandproprioceptiveinputfrom descend-ingmotor pathways (Harkema et al.,2011; Grahn et al., 2017). Thesetwoconditions(elevationofthebaselineexcitabilitylevel combined with adding relatively small amounts of sensory excitationand\orsupraspinalinput)allowthesubjecttogenerate amovementbecausethelevelofexcitabilitycannowbeelevated abovemotorthreshold(Angelietal.,2014;Rejcetal.,2015;Grahn etal.,2017).Thus,spinalcordnetworkscanbeconvertedfroma non-responsivestatetoonethatcangenerateandconductaction potentialswhenthemembranesaresufficientlyexcited.Thus,the finalcomponentoftheneuralmodulatorystrategyistoengagethe networksthathavebeenelevatedtoalevelthatexceedsthemotor thresholdof theappropriate motor poolswith theappropriate timingthatwillgenerateawell-coordinatedmotortask,suchas stepping,standingorreaching,foranimprovedtrunkcontrol.This engagementof thenetworkswithelevated excitabilityprovides the means for the networks to relearn how to translate proprioceptiveandcutaneousinputduringmotortasks,tolearn thenecessarytimingamongagivensetofsynapsesthroughwhich sensory information is translated to a motor event. When the performance of these motor tasks is practiced periodically, extensivesensory-motor learningoccurswithinand among the networksthatbecomeengagedandtheimprovedfunctioninthe presenceofneuromodulationinonetrainingsessionpersiststothe next trainingsession. Thispersistence reflects thelearning and memorycapability within the spinal networks. This combined series of events provides the basis of the effectiveness of rehabilitation.Forexample,therewouldbelittletonoprogression in improvement if there was not some persistent effect, i.e. memory of the events that occurred in the previous training session.Thecrucialpointisthatthespinalnetworksneedtobe engagedina motortaskthatgeneratescutaneousand proprio-ceptive inputthat can drive thereorganization of thesynaptic

efficaciesofselectedcombinationsofinterneuronsand motoneur-onstogenerateacoordinatedmovement.

A relatively common observation during the process of improvingmotorfunctionwithneuromodulationisthatthelevel ofexcitationthatmustbeprovidedforthespinalnetworkstoreach motorthresholdisgraduallyreduced(Shahetal.,2012).Thiswas demonstrated in subjects with complete paralysis receiving transcutaneouseSCSonceaweekforeightweeks,withthemotor taskbeingthe generationofrhythmic stepping in anunloaded condition(Gerasimenkoetal.,2015).Attheendofthestudythere wasasignificantincreaseintherangeofmovementsthatcouldbe generatedduringthestep-likecyclewhenthesubjectattempteda stepeitherinthepresenceorabsenceofanystimulation.Butthe mostimportantobservationinthis casewastheabsenceofany significantdifferenceinthevoluntarilygeneratedmovementswith orwithoutstimulation.Thus,itsuggeststhatboththesupraspinal and spinal networks had reorganized sufficiently to become equally effective with a voluntary effort alone compared to a voluntary effortgenerated in the presence of electrical neuro-modulation.Asimilarphenomenonwasdemonstratedinratswith acompleteSCIthatweretrainedtostepsideways,i.e.significantly lowerstimulationintensitywasneededattheendof28training sessions(Shahetal.,2012).Further,withinasingledayoftesting, wheneitherquipazineand/orstrychninewereadministeredtoa spinalrat,theintensityofstimulationneededtoinducebipedal steppingwaslowerthanwhenneitherwereadministered.

To gain somemechanistic insight into thecharacteristics of spinalnetworksthatarebeingmodulatedbyelectricalstimulation, there has been an extensive focus on the effect of different intensities ofstimulation onthe characteristicsof the simulta-neous spinally evoked potentials in multiple muscles. The characteristics of these potentials are highly dependent on multiple factors, particularly on the intensity of stimulation. Evokedpotentialshavebeencharacterized asearly,middle and lateresponses(ER,MR,LR,respectively,Gerasimenkoetal.,2006). ERcanoccuronlyathighstimulationintensitiesanditislikelyto betheresultofthedirectrecruitmentofmotoneuronsorventral roots. The presence of a time-linked MR over all stimulation intensitiesintheflexorandextensormuscleshassomecomponent consistentwithamonosynapticreflex.Ontheotherhand,theLRs almost certainlyreflectcombinationsof polysynapticresponses generatedthroughspinalinterneuronnetworks.TheseLRscanbe much more readily generated in the presence of different pharmacologicalcocktails(Gadetal.,2013a,2015)orofchronic subthreshold stimulation (Gad et al., 2013b). The lowered dependence on the time-linked MR and increased LRs during chronic subthreshold stimulation suggests that the spinal net-works arenot beinginducedtogeneratea movement,but are placedinastateofreadinesstoenablemovementsbasedonthe appropriateproprioceptiveinformationofsupraspinalcommands. 7.2.Reconnectivityofsupraspinal-spinalnetworksviatransformation ofelectricallynon-responsivetofunctionallycompetentcellular elements

FirstthestudyofHarkemaetal.(2011)ononemotorcomplete paralyzedsubjectfollowedbythestudyofAngelietal.(2014)on threeadditionalindividualswithchroniccompletemotorparalysis demonstratedclearevidenceoftheabilitytorecoverjoint-specific, coordinated voluntary movements in the presence of epidural stimulation.Theresultsinthethreeindividualswhoweretested after implantation, but before repetitive training suggest that descendingconnectionstothespinalcordcircuitryemergedasa result of stimulation of networks wellbelow the spinal lesion transforming cellular components within the more proximal lesionedsegmentsfromelectricallynon-responsivetoresponsive.

Unlikeinthefirstsubjectinwhichthisobservationoccurredafter almost7monthsofdailystimulationsessionswithtraining,the latterthreesubjectsregainedthesupraspinal-spinalconnectivity withinonlyafewstimulationsessions.Sucharapid transforma-tionmakesit highlyunlikelythatthemechanism for reconnec-tivitycanbeattributabletoasignificantdegreeofaxonalgrowth. Also,evenwhenthisreconnectivityemerged,itcouldbeobserved inthepresenceofacritical levelofstimulation.Dailyexercises using epidural stimulation in conjunction with standing and voluntary activity resulted in the generation of higher forces duringvolitional movementsand lower stimulationvoltages to reachthemotor thresholds.Interestingly, theparticipantswere abletomorereadilygenerateflexioncomparedtoextensioninthe lowerlimbs(Angelietal.,2014).Althoughtheseresults demon-stratethatthespinalcircuitriesbelowthelesionhaveapotential thatexceedswhathasbeenreportedpreviously,criticalquestions to address are (1) To what degree is activity-based training necessaryorbeneficialtorecovervoluntarymotorcontrol?and(2) What are the critical training parameters for regaining more effectivevoluntarymotorcontrol?Whilethereislittleinformation availabletoanswerthesetwoquestionsastohowtheypertainto electricalorpharmacologicalneuromodulation,thereisabundant

evidenceofplasticitywithinthesensory-motornetworkswithin brainandspinalnetworksinresponsetotrainingalone.Thereis littleclarity,however,onaconceptualprinciplethatdefinesthe degreeofspecificityamongdifferentsensory-motortasks neces-sarytogain somecross-over effectsamongdifferenttasks.This importanttopic,however,iswellbeyondthescopeofthisreview. In considering the questions noted above, however, several pointsareappropriatetokeepinmind.Significantlevelsofcortical reorganizationarecommoninhumansafterSCI(Oni-Orisanetal., 2016).Mappingstudiesusingtranscranialmagnetic stimulation revealanexpansion of corticalsensorimotor areasrepresenting musclesabovetheleveloflesioninindividualswithtetraplegia (Levyetal.,1990)andenhanced excitabilityof motorpathways targetingmusclesrostraltothelevelofaspinallesioninthosewith paraplegia(Topkaetal.,1991).Corticalreorganizationafterdorsal columnlesionisnotlimitedtotheprimarysomatosensoryarea, but also extends to the secondary somatosensory cortex and parietalventralarea(Tandonetal.,2009).Multiplecombinations ofadaptivenetworkreorganizationsundoubtedlyoccurandsome of these changes are likely to lead tosome recovery of motor control. It seems probable that functional reorganization of cortico-brainstem-spinal,corticospinaland spinal networkswill

Fig.3.NeuromodulatorymechanismsfortherecoveryofvolitionalcontrolafterSCI(simplifieddepiction).

Thiscartoonillustratessomeoftheproposedtargetsandmechanismsofneuromodulationofthespinalcord,asdiscussedinthebodyofthepresentreview. Physiologically(A),aplethoraofpropriospinalneurons(reddots)establishesadiffuseddynamicnetworkofsynapticcontactsalongthegreymatteroftheentire thoracolumbarspinalcord.

CPGsresponsibleforthegenesisandthecontrolofdifferentmotortasksarerepresentedbycircumscribedareasofthenetworksasfunctionalsubgroupsofinterneuronsthat projectinacontinuouslychangingpatterntowarddifferentcombinationsofmotorpools.

Fibersdescendingfromsupraspinalcenters(topblacklines)reachthepropriospinalnetworkandeventuallythespinalmotoneurons(purple).Descendingvolitional commandisrepresentedbyasingledepolarizinginputthatreachesthemotorthresholdfornetworkactivation(yellowboxwithblackline).

Afferentinputfromtheperipheryalone(greenline)canreachthemotorpoolsviatheinterneuronalnetworksystemsnotedabove,whichiscapableofconveying subthresholdandsuprathresholdphasicoutput(yellowboxwithgreenline).

Forexample,duringlocomotionasynchronousdischargesofsubthresholdandsuprathresholdinput(yellowboxwithredlines)converges(whitearrows)ontointerneuronal networkswithCPGqualities(highlyorderedactivationofnetworksthatprocessproprioceptiveinputinrealtime)withsufficientprecisiontocontrolmovementsascomplex aslocomotion(locoCPG;paleyellowcircle).

ThelocomotorCPGusesthesubthresholdbackgroundnoisefromthepropriospinalnetworkandthesubthresholdphasicrhythmfromafferentstoorganizerhythmicand alternatedpatterns(locomotorpattern).Thelocomotorpatternpassesthethresholdinordertosequentiallyactivatemotoneuronalpools(purplelines)andthusmusclesina highlytime-dependentactitionpatternofmuscles(notshown).

Forthesakeofsimplicity,weonlyrepresentedthemotorpoolsononeside,buttheoutputfromtheinterneuronalnetworkshavingCPGqualitiesishowevertobeconsidered bilateral.

Followingaspinallesion(B),thedescendingpathwaydirectedtomotorneuronsisinterrupted,whilethepropriospinalnetworkbelowlesionmostlyappearstobe functionallysilentandafferentinputreducedbecauseofdamageandparalysis.Thus,thelocomotorCPGdoesnotreceiveanadequateamountofsubthresholdinput,hence theinabilitytovoluntarilyevokemotorpatterns.

Aninnovativetreatment(C),whichpairsanappropriatesubthresholdmulti-siteelectricalstimulationofthethoracolumbarspinalcordwithaspecificmotortraining,can reactivatepartofthesublesionalelementsofthepropriospinalnetwork,thusrestoringthesubthresholdbackgroundactivitythat,althoughreducedinfrequencyduetothe moreexiguousnetworkextension,canstillreachtheCPG(yellowboxwithredlines).

Althoughthedirectdescendingfibersremaininterrupted,descendinginputcannowreachmotoneuronsthroughthepolysynapticpatternofpropriospinalconnectionsnow transformedfromelectricallynon-responsivetoresponsive.Thedescendingvolitionalcommand,althoughwithreducedamplitudeandincreasedlatency,canbeprocessed bythespinalnetworkssothatwell-coordinatedpatternsofmuscularactivationcanbegenerated.

definethe degree and qualityof thenewly acquiredvoluntary function.Giventheextensivefunctional reorganizationthatcan occurwithin and between thebrain and spinal networks, two pointsseeminevitable.First,giventhemassivelossofconnectivity withasevereSCIitishighlyimprobablethattheconnectionsthat re-emergedinthecasesnotedabovecanbeattributedtothesame connectionsthatwerepresentbeforethelesion.Secondly,thelevel of functionality that developsas a result of the newly formed connectionsislikelytobereflectedinthelevelofsynergismofthe reorganizationofthenetworkswithinandbetweenthebrainand spinalcord.

The fact that regaining voluntary motor control can occur withinafewsessions(Angelietal.,2014;Gerasimenkoetal.,2015), foryearsafterSCI,indicatesapersistenceofthepotentialoftaking advantageof brain-spinal plasticitythat not only hasnot been previouslyrealized,buthasbeensteadfastlydeniedasapossibility. 8.Possibleinterpretationofthevoluntaryrecoveryinducedby eEmc

Basedonourcurrentstateofunderstanding,efficacyofeEmc can be attributed to either the potentiation or reactivation of cellular components that survived the lesion, but which are sufficientlydamagedanddisabledintheirpotentialtoreinstatea functionalsupraspinal-spinalconnectome.Oncesomefunctional connectionhasbeenrestored,eEmcseemstofurtherenhancethe functionallevel byamplifyingandremodelingthis connectome (Fig.3).Thismightbeinducedviasproutingofthefewremaining fibersthataresparedbythedamagebutthatareinaninsufficient numbertorecruitanappropriatequantityofpostlesionaltargets (Calancie et al., 1999; Asensio-Pinilla et al., 2009). A greater numberofsynaptic contactswithsub-lesionaltargetsmight be established, which will consequently increase the number of voluntarilyrecruitednetworks.

Also,eEmcmightbringtheexiguousdescendingaxons,spared bythelesionbutunabletoconductimpulses(Shumanetal.,1997; NashmiandFehlings,2001),toafunctionalstatethatenablesthem toconductactionpotentialstospinalnetworksbelowthelesion. Electricalstimulationofresidualdescendingfibersoveragreater periodoftimemightpromoteremyelination(Ishibashietal.,2006; Wakeetal.,2011)andre-establishconductioninfibersthatwere spared by the initial trauma but apparently became unable to conduct impulses (Kakulas, 1999). To this regard, electrical stimulationofthecorticospinaltractinducedactivity-dependent proliferation and differentiation of oligodendrocyte progenitor cells(Lietal.,2010)andtheformationofmyelininmixedneuronal and oligodendrocyte cultures (Gary et al., 2012). Notably, a protractedelectricalstimulationofthespinalcordusing transcu-taneous electrodes, improved the neurological outcomes of personswithdemyelinatingdisorders(Dooleyetal.,1978).

Moreover,thereduceddescending inputsfollowing alesion, combinedwithsubsequentdisuseduetoparalysis,maydepress synapsesamongresidualdescendingandpropriospinalfibers,as wellasMNs. Synapticdepression due tothe presenceof silent synapses (Malinowand Malenka, 2002)has been identified in someneurological disorders (Shevtsova and Leitch, 2012; Wan etal.,2011;Kerchneretal.,1999),butnotyetinSCIs.

At pre-synaptic levels, eEmc might modulate density and propertiesofionicchannelsinthenetwork,suchastransientIAK+ currents(Watanabeet al.,2002)andCa2+activatedK+channels (Nanou et al., 2013). Moreover,it might facilitatevariations in synapticexcitability, byactingonNa+_{channels (Ganguly etal.,}

2000), Ih current (Mellor et al., 2002) and slow-activating K+ channels(Lietal.,2004),or,again,itmightlimittherelease,from postsynapticterminals,of retrograde factorswithaninhibitory significance(endocannabinoids,Sjöströmetal.,2003).

Inaddition,electricalSCSincreasescerebralspinalbloodflow (Kannoetal., 1989;Linderothetal., 1995;Liuetal.,2008)and,thus, its prompt adoption is proposed for rescuing sleeping neurons locatedinthepenumbrazoneofcorticalischemia(Visocchietal., 1994).Asimilareffectmightoccuralsoatspinallevel,butonlyif aroundthechronicspinallesionarelocalizedsomeimpairedbut sparedneuronsthatwouldbenefitfromtheincreasedperfusion. In addition, impulses that travel along residual axons may projecttoelectricallynon-responsiveneuronsbeloworwithinthe lesion.AfteraSCI,severalchangesoccurinintrinsic electrophysi-ological properties ofmotoneurons,suchas thedepthof after-hyperpolarization of action potentials and the amplitude of segmental or descending excitatory post-synaptic potentials (Petruskaet al., 2007).Concurrently, anincreased activationof intrinsicpersistentinwardcurrents(PICs;ElBasiounyetal.,2010) andadownregulationofthepotassium-chlorideco-transporter-2 (KCC2; Boulenguez et al., 2010) can change motoneuronal responses to sensory stimuli. Moreover, an upregulation of glycinergic receptors depresses motoneurons, as demonstrated by facilitation of stepping in chronic paraplegic animals using prolongedfullweight-bearingafteradministratingtheglycinergic inhibitor,strychnine(deLeonetal.,1999).

Furthermore,eEmc might affectbasic and active membrane biophysical propertiesof motoneurons(8) or the expressionof mGluR receptors(KarmarkarandBuonomano,2002).Moreover, eventhedifferentkineticsofpostsynapticNMDAreceptors,linked to a variation in NR2A/NR2B subunits expression, must be considered(Tangetal.,1999).

Finally, awakening silent synapses may require coordinated pre- and postsynaptic modifications,e.g., incorporation of new AMPAreceptorsatpostsynapticsiteandpresynapticmodification inreleasemachinery(VoroninandCherubini,2004).

The spinal cordowns smart abilities, such as learning new motortasks,forgetthemafteralesionordisuseandremember them again after an intense task specific training, that were considered,untilmorerecently,tobeaprerogativesolelyofthe brain(Edgertonet al.,2004, 2001;Jindrichet al.,2009).In the brain, information storage and refinement of neuronal circuits occurs throughthephenomenonof Hebbiansynapticplasticity, which pairs stimulation of pre- and post- synaptic terminals (VoroninandCherubini,2004).Thetimingwithwhichpre-and post-synapticimpulsesfollowoneanothercanmodifysynaptic efficacy in a dependent manner, i.e. spike-timing-dependent plasticity (Caporale and Dan, 2008; Dan and Poo, 2004). This principleisinaccordancewithobservationsinbothcorticaland hippocampalneurons(LevyandSteward,1983),alsoreproduced bymodelingstudies(Songetal.,2000).Moreover,inindividuals withachronicincompleteSCI,spiketiming-dependentplasticity ofsparedcorticospinal-motoneuronalsynapseprovidesa mecha-nism toimprove motor functions of upper limbs (Bunday and Perez,2012).

Basedonspike-timing-dependentfacilitation,paired stimula-tion of multiple inputs, converging onto MNs with different frequenciesand latencies,maybecrucial in recruitingdifferent motoneuronalpools.Thisisinlinewiththeobservationthatdirect stimulationofthespinalcordwithdifferentfrequenciesfacilitates different voluntary motor tasks following SCI (Harkema et al., 2011;Gerasimenkoetal.,2015;Shahetal.,2016).

9.Synergismofneuromodulationandproprioception Afferent inputs projecting to the postural and locomotor networks are the primary contributors in facilitating standing and stepping.Indeed, assisted stepping evokes afferent inputs, whichallow therecoveryoflocomotioninsubjectswithspared residual connections crossing the lesion (Wernig et al., 1995;