Gait Restoration in Paraplegic in Paraplegic Patients: A ... · Recent advances in science have aided research toward the resto-ration of biped gait in paraplegic patients by means

PAGE 3

Journal of Rehabilitation R&DVol . 20, No . 1, 1983 (BPR 10-38)Pages 3–20

Gait Restoration in ParaplegicPatients : A FeasibilityDemonstration using MultichannelSurface Electrode FES a

ALOJZ KRALJ, D . Sc ., E . E . bProfessor

Faculty of Electrical EngineeringEdvard Kardelj UniversityLjubljana, Yugoslavia

TADEJ BAJD, D . Sc.Assistant Professor

Faculty of Electrical EngineeringEdvard Kardelj UniversityLjubljana, Yugoslavia

RAJKO TURK, M . D.Head of Paraplegia Department

University Rehabilitation InstituteLjubljana, Yugoslavia

JANEZ KRAJNIK, Dip! . Eng.Research Assistant

Institute Josef Stefan, Edvard KardeljUniversity, and University RehabilitationInstitute, Ljubljana, Yugoslavia

HELENA BENKO, P . T.P. T. Instructor

University Rehabilitation InstituteLjubljana, Yugoslavia

'This research work was supported in part byResearch Grant 23-P-59231/F from the Na-tional Institute of Handicapped Research,Department of Education, Washington DC20202, and Slovene Research Council,Ljubljana, Yugoslavia and partly by Univer-sity Rehabilitation Institute, Ljubjlana,Yugoslavia.

b Prof . Kralj, at the time this paper was sub-mitted, was Visiting Professor at Illinois In-stitute of Technology, Dept . of Electrical En-gineering, and Pritzker Institute of MedicalEngineering, Chicago, Illinois.

He may be addressed at Univerza EdvardaKardelja, Ljubljana, Fakulteta za elektroteh-niko, 61001 Ljubljana, Trzaska 25,Yugoslavia .

ABSTRACT

Recent advances in science have aided research toward the resto-ration of biped gait in paraplegic patients by means of functionalelectrical stimulation (FES) . In this paper it is shown how FES-restrengthened muscles of paraplegic patients have been used forsimple FES-assisted standing . Those experiments subsequently ledto biped gait-initializing experiments and to simple forms of bipedgait synthesis . The purpose of this paper is to show the feasibility ofusing FES for standing and for restoring biped gait in many paraple-gic patients—to present the past achievements, focus on problems,and highlight directions for future research.

The results of gait obtained in three complete spinal cord injuredpatients (out of a series of 17) are shown, using four to six channels ofFES . It is also shown how preserved reflex mechanisms of thetransected spinal cord can be incorporated and employed for obtain-ing improved function while at the same time simpliflying the FEShardware . Of the three patients reported on in detail here, twopatients have managed to walk in parallel bars while the third patienthas mastered independent unassisted walking over shorter dis-tances with the aid of a roller walker. The biomechanical and controlproblems of this last patient's gait are presented in detail.

INTRODUCTORY REVIEW

The advances in electronics during the last decade, particularlyminiaturization, improved reliability, and lower prices, have facilitat-ed the research and clinical application of advanced electronics invarious fields, including Functional Electrical Stimulation (FES) . Theachievements in neurophysiology and extensive research in locomo-tion engineering have opened new approaches in rehabilitation . Inspite of these helpful developments, the problem of locomotion inspinal cord injured (SCI) patients still requires better solutions. Inpast decades the main medical problems of these patients wereconsidered first and locomotion in wheelchairs was generally ac-cepted as the only realistic means of mobility . But it is obvious thatenvironmental barrier problems could be more suitably reduced forSCI patients by restoring some degree of non-wheelchair locomo-tion. Doing this by FES might also be less expensive, more function-al, and more attractive for the patient (and society) than the wheel-chair. Therefore, since the year 1972 we have focused our researchon feasibility studies of FES-assisted biped gait in SCI patients . Otherresearchers have followed the same rationale (1, 2, 3, 4) . The earlywork by Kantrowitz (1), Wil lemon et al . (2), and Kralj et al . (3) providedevidence that FES-stimulated muscles in paraplegic patients could

KRALJ et al . : GAIT RESTORATION FEASIBILITY

perform work when restrengthened by means of FES.More recently it has been shown that functions such asstanding and even rising from a sitting position tostanding can be accomplished in complete SCIpatients.

Research was also conducted in exploring thechanges in muscle properties and tissue because ofprolonged periods of FES as used in restrengthening(5, 6, 7, 8, 9) . Peckham et al . (5) and Mortimer et al . (6, 7)focused their work on such physiological changes asmetabolism, contractile properties, muscle fiberchanges, etc. It is important to note that their basicfindings have confirmed the conclusions of earlierauthors . Today it is common knowledge that partlydisused, atrophied, and upper-motor-neuron-lesionedmuscles can be restrengthened by means of FES.These muscles are later used for performing function-al tasks requiring high muscle forces ; for example,providing joint torques for standing-up (3) or kneelocking in gait and in standing (2'8) . The first reporteduses of implanted functional electric stimulators in SCIpatients which provided knee locking via FES of M.Quadriceps muscles were in work performed in 1971,by Willemon et al (2) and in 1979 by Brindley et al (8).These authors made nnu!iichannel implants foren-abling SCI patients to walk in a swing-to or swing-through mode gait assisted by FES (with FES kneelocking substituting for long leg braces) . All the pre-viously described achievements have demonstratedthat FES of paraplegic muscles is feasible . In Ljubljana,Yugoslavia, the researchers have used stimulationthrough surface electrodes to obtain prolonged stand-ing of the patient (10, 11, 12). The work was based ontheir earlier experiences using multichannel FES instroke patients (13, 14, 15) . The research reviewed herehas been focused on the problems of standing-up,maintaining erect pooition, and balance . Most recentemphasis has been given to the problems and require-ments for weight shifting between supporting limbsassisted by FES, and the problems of single-limb anddouble-limb support as they take place during gait . Allthis has led to research on the possibilities of synthe-sizing functionally useful FES-assisted biped gait inparaplegic patients.

FES : area!istic andbetteroo!utioninparaplegia?--TheertUicia!oynthoaioofgubueingFEShuo ' uttirneo 'looked like a problem whose complexity grows veryrapidly as research improves our awareness of thephysiological, neurological, and bioengineering facts.No doubt this is typical of most sustained work inintrinsically complex areas of study : the investigatorsare sustained by confidence that their own work andthat of others periodically leads to fresh insight andunderstanding—and ultimately to simplification.Along the way the investigators can hope for encour-

aging signs that praoticul c!inioal benefits (or evenmajor gains in life-style) for the majority of paraplegicpatients could emerge from the work in FES-enabledgait.

Signs that achieving uoeful gait via FES is a trulyimportant goal may be seen in various bits of informa-tion . For example, biomechanical studies in the UnitedStates, comparing wheelchair propulsion with bracedcrutch-walking (in swing-to or -through modes of gait)gave evidence that " . . .patients with paraplegia whorequire two KAFO's . . . usually do not continue to walkat a!!âfter!aurningtovva!kinenahebi!itationhonpiiu!(quote from page 1133, Physical Therapy, Vol . 60, No. 9(Ref . 16)) . The high energy demand of today's crutchwalking is the reason . In oon1raot, biped gait via FESnow seems unlikely to impose unacceptably (or evenuncomfortably) high energy demands once a healthypatient has been trained in the use of restrengthenedmuscles for FES-enabled biped gait"

The advantages of using the FES method for re-enabling gait in paraplegic patients are many. Naturalbone support of body weight and existing joints, liga-ments and muscle power are used . No external bracessuch as KAFOs, etc, or force transfers with levers, areneeded for FES orthopedic devices . The muscle pro-vides a self-regenerating energy supply, and electricalstimulation is used only for triggering, therefore FESorthotic devices will have low battery needs whichsuggests good autonomy.

Preserved Reflexes Employed —An additional veryimportant advantage mfFES is that it can take advan-tage of preserved spinal reflexes and neuromuscularmechanisms. When these mechanisms are employed,FES is applied as sensory stimulation to provide trig-gering and facilitation of preserved functions such asthe flexion synergistic response, agonist or antago-nist, undoontra!atmra! extmnsur/f!exorinhibition or ex-citation. Because of the advantages discussed above,the FES method has a theoretical advantage over othermethods (such as the powered "exoskeleton", etc .)which use externul oonirol for locomotion. Such anexternal device typically fails to become (or to beperceived as) an integral part of the patient, and evenwhen the devices function well they are only rarelywell-accepted by patients . In contrast, FES, when de-livered by means of an implanted stimulator, becomesvirtually an integrul part of the patient and has there-fore a good possibility of being well received by thepatient . This reasoning, if confirmed, would clearlysuggest that research toward totally body-implantedFES systems should be emphasized in the future.

'The authors believe that, in the near future, FES-assisted paraplegicgait with crutches will be feasible .

5

Dsadvantages of Using FES —There are also disad-vantages to using FES. These include difficulties withelectrode placement, inadequate selectivity, the needfor costly technology, and insufficient present knowl-edge of neurology, neuroanetonny. FES, and relatedfields and of the information processing required fordecision-making and control . It seems that most ofthese disadvantages could be minimized in the nearfuture by further research.

Up to now, only fragments and notes regarding thework in Ljubljana, Yugoa!evia, with FES-assisted gaitin paraplegic patients have been published . This paperwill explain in some detail the methodology used, andcritically summarize the results obtained so far on 17patients.

GENERAL METHODOLOGY

Patient Suitability

Because of the lack of previous experience, no pa-tient-selection criteria were available initially (in 1972).Spinal-cord-injured patients with thoracic lesions be-tween T5 and T12 were selected . They were at the endof their standard rehabilitation program, facingdischarge d .

In general, the subjects were middle-age and youn-ger patients with good physical health, showing noheart, lung, or bladder problems . Major contraindica-tions were contractures, osteoporosis, joint oaoifica-don ' ulcers, or very severe spasticity

The data obtained from three patients are describedin dotai! in the ^Cfinioo! Obeervatiunu^ aaction . Thepatients are those who proved to perform the best inour series of 17 randomly selected patients. Some ofthe other patients did not complete the program be-cause ofvarious reasons (m .g,being discharged, nna p

riagm' goingtosohou! ' ato .) . Owing to this the patientpopulation which could benefit in the future from thismethod cannot now be precisely specified, but consid-ering only standing enabled by FES, practically all thepatients who suit the criteria in Table 1 can be expect-ed to be able to master the standing in a time of 2–3months.

Testing Muscles for Suitability—Patients were fullyinformed of the research procedures when admittedto the program . Then, after being checked for thegeneral requirements, the patients were tested for

It must be noted that some patients many years postinjury can beadmitted to the program with the same success . It is important thattheir status corresponds to the limitations outlined in the criteriafor selection (Table 1 and Table 2) . The derivation of these suggest-ed criteria ixueo ^,iêd inme ôinica!Ubx :rvauonn^ pomono,this paper.

Journal of Rehabilitation R&D Vol . 20 No . l 1983 (B pn10-38)

TABLE 1.Patient indications for FES

1. Lesion level T-4 through T-12

2. Upper motor neuron lesion

3. Positive results of FeSmananouxonin 9 program

4. No joint contractures or other involvements(e .g ., ossification, osteoporosis)

5. No major skin problems

6. Normal balance sensation

7. Good physical, mental, and emotional condition8. Motivated and cooperative

TABLE 2.Contraindications for FES

1. Osteoporosis

2. Ossification of Joints3. Peripheral lesion of main leg muscles

4. Severe atrophy of muscles so that mstronomeo'in B isimpuxuih|n

5. Pressure sores

6. Obesity7. Very severe spasticity

8. Inadequate sitting balance

muscle response to FES . Only main hip, knee, andankle-joint muscles were tested . Rectangular die.pulses of 0 .3 ms with frequency of 20 Hz were used.The amplitude of stimulation was in the range of 40 Vto 120 V. The stimulator used has been previouslydescribed (14) . Stundardvvater-ooakedfoam-rubber-covered sheet nnetal electrodes were used (17).

The FES testing was used to verify that the mainmuscles had not suffered lower-motor-neuron lesions(muscles which have suffered such lesions cannot beused for FES),

During these initial stimulation trials, recordingswere made to display maximum force and the fatigu-ing rate of stimulated muscles . At the same time, thestandard testing for reflexes was performed . Stretchreflexes and gross extension and flexion patternedreflex movements were initiated as a proof that thespinal-cord levels below the transection were stillfunotional and had not been destroyed during theacute phase because of ischemia, edema, hematoma,surgery, physical handling, or other reasons . Typically,and in a majority of patients tested, the muscles inner-vated o!oeetothe!eva!ofinjurymre peripheral-dener-vated (do not respond to FES) ; the muscle tissue ismissing because of atrophy and the muscles lost forFES treatment . This can also be determined by visualinspection, because the muscle tissue is lessened tothe extent that bony prominences are clearly seen.Muscles such as M. Quadriceps may show disuseatrophy very rapidly—even in cases with intact periph-


Fi hi

StanuingbvmemnsofFESanddmwheelchair-attamnonamei

era! innervation . In the latter case, even when atrophyhas begun the muscle force can frequently be ne-atrengthenedbynnaanaofFES ' evmntoufairfunuiion-a!!eveiiftheinitio!foromivat!eoetonthoorderof20percent of nornoal funm1ionul strength . (So far, thereare no methods developed enabling the prediction ofthe restrengthening results).

Because of various reasons, levels of injury at T12and below frequently result in peripheral denervationof the muscles involved . Injuries at lumbar levels allresult in periphnral nerve lesions because of damageto the cauda equina.

Muscle Restrengthening Program

After initial testing, the FES muscle restrengtheningprogram was started . Isotonic exercising was per-formed using bilateral movements or reciprocalmovements patterned similarly to gait . FES was initial-ly administered for about 18–15 minutes twice daily.At this point, control was via switches operated by thephysiotherapist or the FES was triggered by a cycler.The patient was then taught how to apply the stimula-tion, or relatives were taught to do it for him.

The FES of spastic muscles can increase, decrease,or have no effect on spasticity. If it does increase thespasticity, FES should be discontinued . Inhibition ofspasticity by means of FES is not fully understood (18,19 ' 20 ' 29) . ThoFES sessions for muscle restrengthen-ng were gradually increased to stimulation times of

150 minutes twice daily. The optimal timing and proce-dure for muscle restrengthening have not yet beenestablished. The gain in muscle force was measuredweekly or biweekly and recordings for fatigue monitor-ing vvermrnade . Records were also made of the pa-tient's weight and length of body segments, for laternmathmrnaticel modelling (12) . When the joint torquesproduced by FES reached a functional level, standing-up and sitting down training was started, allowing thepatient to assist himself freely with his arms . The erectstanding time was gradually increased to the pointwhen fatiguing of knee locking prevents further stand-ing. During this phase of the program, M . quadricepsmuscles were continuously stimulated bilaterally dur-ing standing . Occasionally, FES of hip extensors wasadded for insuring good posture with hyperextendedhips . The patients were taught to keep their balancewith the upper body and to use their arms only forlarger body shifts and to prevent the development ofunstable postural situations . The patients had to mas-ter hand-assisted weight shifting for right and leftsingle stance . (In some patients, FES of hip abductorswas added for weight shifting and prevention of over-lapping of the feet .) For the patient's exercise and laterpossible use in gait, the M . yo!*uaund M . gastrocne-mius were cyclically stimulated, providing lifting ontheir toes .

7

Journal of Rehabilita on R&D Vol . 20 No . / 1983 (BPR 10-38)

Simp'e Gait Training Started

Once the patient felt safe and comfortable whilestanding, simple gait-pattern training could be started,beginning with walking in para!!al bars . As an interme-diate stage to allow walking over longer distances, awalker frame was used (9) . Later, walking by means ofa roller walker could even be performed . One patientwas taught how to adjust the electrodes and the stimu-lator itself for independent roller-walking . Once he hadmastered walking with FES biped gait, measurementassessments were made.

After the FES clinical program was terminated, sixpatients were supplied with their personal two-chan-nel stimulators for home use . These units providedonly standing and muscle exercising . So far, only onepatient has been given a roller-walker with a 4-channelstimuator for home use, enabling him to walk at home.For patients who considered FES standing functional,a special wheelchair-attached collapsible supportingframe was developed, allowing them to use erectstanding at any location indoors or outdoors (Fig . 1)(12, 21).

Patients using FES at home were visited frequentlyand closely monitored . Their most frequent com-plaints were about tedious electrode positioning andthe required redressing ; the latter was particularlyunpleasant in winter. However, surface electrodes willcontinue to be used in the University RehabilitationInstitute to collect more knowledge, refine the nne1h-odo!ogy,ond evaluate the entire procedure.

GAIT SYNTHESIS

The proper timing of FES for each muscle, and thecoordination of the patient's trunk maneuvers, areessential for achieving good walking . Therefore, it isimportant that the patient has control of major events.For a particular function we have to select an appropri-ate stimulation activity-program, subsequently re-ferrmdto here as an FES oequanoei

The act of gait synthesis is the development andestablishment of the FES sequences which producethe necessary movements which will result in theexecution of a particular function in time (e .g . mtand-ing-up, sitting-down, or biped walking) . FES can beapplied in two ways : as efferent-motoric stimulationresulting in a contraction of the selected muscle, or asafferent-sensory stimulation of nerve endings whichprovides triggering of reflex events (e .g . ' the flexionwithdrawal movement which can be incorporated ingait) and hence the stimulation sequence is appropri-ately determined.

Improper FES sequences will provoke movementswhich deviate from the desired function and whichcould even create hazardous situations . Synthesizing

the gait by means of FES consists of selecting anddetermining the order, duration, and synchronizationof events which are necessary to obtain a simple gaitpattern . During the development of the needed FESsequence we have to take into account the limitationsas well as the arsenal of preserved movements whichmay be exerted voluntarily by the patient, and then weenable the remaining needed movements by FES . Thepatient assembles the events by his voluntary move-ments (which may in some cases themselves be usedto trigger FES stimulation of the next movement) andemploys commands under his direct control (as in thecase of a switch which is patient operated) in such amanner as to result in the planned and wanted func-tion . The synthesis of FES-assisted gait requires pro-feaaionel knowledge of biomechanics of gait, offunc-tiona! electrical stimulation, neurophysiology, andmusculoskeletal anatomy and physiology, This proce-dure of determining the FES sequences is a creativeand demanding task.

Consider the stimulation sequence for standing . InFigure 2 the stimulation sequence is shown forM. quadriceps-assisted standing (A) . The kneee joint islocked . (Standing, assisted also by hip extensorsM. gluteus maximus and M . gluteus medius, is alsoshown (Fig . 2, B) ; this provides better hyperextensionof the hip and hence better posture) . FES for providingcyclical -toe standing" (body lifting) movements dur-ing erect standing requires the activation of M . soleusand M . gastrocnemius muscles bilaterally, as shown inFigure 2 (C).

The gradual increase and decrease of stimulation

FES-AMPLITUDE

TIME

ADDITIONAL ASSISTANCE FORPOSTURE- EXTENSION OF HIP

w .50LEU5and M . GASTROCNEMIUS

TO-Es

~

.

STANDING)

5TAN06UP

ERECT STANDING TIME

* APPLIED BILATERALLY

FIGURE 2The FES for standing-up and erect standing, witn or without toelifting and sitting-down .

. QUADRICEPS

LOCKING OF THE KNEE)

M .GLUTEUS MAXIMUS

Aa

I RISE 'TIME

1-2 s

B

SITTINGDOWN


during standing-up and sitting-down is important be-cause this gives the patient time for executing theproper associated body movements . Without the riseand fall times (gradually increasing/decreasing stimu-lation) the patient may collapse when sitting-down atthe precise moment when the M. quadriceps stimula-tion is automatically switched off . The patient must betaught to be aware of this and to remember that agiven but limited time is available for nornoul perfor-mance of standing-up or sitting-down, in which hisawareness, parfornnance, and assistance play an ino-pmrtantpart.

Gait Pattern Possibilities

It is obvious that there are numerous possible gaitpatterns for a biped gait . The most simple and funda-mental ones, displaying only the main events of abipmdgait .anaterrn*d^sinnp!e^bipodguhputtorno . Togenenatea simple gait we need, first, a double-stance-phase period . This is the period of greatest stability.During this phase the weight transfer between legswill take place . The weight transfer can be provided byhand pushing movement or by stimulation of ankleplantarflexors . Next, a rather short single-stancephase on one side and swing phase to the oppositeside is needed . This single-stance/swing-phase periodis terminated when the swing-leg heel strikes theground . The antigravity support and swing functionscan be provided by FES, while the rest of the functions(body weight transfer, hip abduction during singlestance, stability, etc.) can be provided either by appro-priate correlated voluntary maneuvers of the upperbody and arms by the patient himself or by selectedFES of additional muscles . In the latter case, the FESsequence would become more complicated . Stimulussettings for each muscle include, in principle, time-delay for the onset referenced to an event oroonn'rnund.theFESdurmtionendmrnp!itudu . At the presenttime it is simpler to leave the burden of correlation andproper timing of as many functions as possible to thepatient ; this simplifies the FES hardware . This philos-ophy gave rise to the basic gait mode described in thispaper and the resulting stimulation sequences whichare shown.

FES Sequences

Before describing the FES sequences in detail itmust be explained that, at present, the whole antigrav -ity support function is provided by motoric FES ofmuscles. In contrast, for providing the swing phase,the required flexion of joints can be obtained in twoways : (i) by motoric (efferent) stimulation of the mainf!exoru, or (ii) by the sensory (afferent) stimulation totrigger the flexion withdrawal movement . The firstgiven possibility for providing flexion of joints neededduring the swing phase requires at least two pairs of

electrodes for stimulation ofanWejoint doroal flexorand knee-joint flexor (M . tibialis anterior and M . bicepsfennorio) ' enda forward-leaning posture . The settingnf6 FES parameters, three for each muscle, must bemade. (A setting is composed of the stimulation ann-p!i1ude ' stimulation onset, and duration .) The require-ments would be even more demanding incorporatinghip flexors and using three pairs of electrodes . Theproposed method is not feasible because the hip flex-ors cannot be adequately stimulated with surface elec-trodes . To compensate to a certain extent for the miss-ing hip flexion, reinforced knee flexion may beintroduced, butthelatter oanbaobtained byanaddi-tional FES pair of electrodes for stimulation of uddi-tional knee-joint flexion muscle tissue. On the otherhand, the setting of FES parameters for motoric-effer-ent stimulation is simple, because we can copy themuscles' activity in normal gait.

The second possibility for obtaining flexion of jointsduring swing uses only one pair of electrodes, butmakes advantageous use of the preserved mecha-nisms of the transected apinal cord . Therefore thehardware requirement is reduced, but the adjustmentof parameters influencing the movement is demand-ing . E!actrophyoio!oAical knowledge of how the differ-ent stimulation parameters and electrode locationsinfluence the resulting flexion reflex-withdrawalmovement is essential . From the literature (22, 23, 24,28) it is known that during a flexion reflex movement ofthe lower extremity, hip and knee flexion and ankle-joint dorsiflexion can be obtained . But published datadescribing how to control the degree of obtained flex-ion, and the timing using FES, are rare . It is obviousthat the swing-phase oontrul via flexion reflex ru-sponoeia!eeapreoioe ' andrequiruarnormaoundiudg*-rnant ' butitrequiraa!eoohardvvure . Since the primaryfunction of the swing phase is bringing the leg for'vvard ' the exact swing trajectory is not strictly pre-scribed, but the proper timing of the toe-off and heel-strike are. In all of our simple gait patterns the forwardpropulsion is provided by the patient with his upper-body movements and with forces exerted through thehands. In addition, the equivalent of the required ab-duction forces for the hip are also controlled by thepatient's hands and the side-leaning movements of histrunk. FES can efficiently provide the hip abductionforces if needed, as our trials showed . The activation ofM . tensor fasciae !xtaavvith M . gluteus medius andminimus requires the setting of two additional pairs ofelectrodes and the adjustment of 6 new stimulationparameters.

The abduction and hip extension torques can also beobtained just by using the M . quadriceps channel ; inthat case the upper M . quadriceps electrode is movedand placed proximal to the crest of the pelvis andplaced slightly lateral-dorsal . This placement can be

9

Journal of Rehabilitation R&D Vol . 20 No . 1 1983 (BPR 10-38)

used partly for co-contraction of M . gluteus mediuswith the tensor faciae !a!a*.

Thus forward propulsion in the push-off phase ofgait can be provided by FES of appropriate muscles,but again, additional hardware and settings are need-ed—it can be seen that additional functions requirehardware and the need to set more parameters . Thelatter, in turn, present additiunal requirements for in-formation processing and control.

The reasoning just described provides the back-ground for understanding our selection of the initialexperiments and the use of the simplest possible gaitmodes . This approach requires the least hardware andcontrol processing . Following the experience gainedin multichannel stimulation we actually incorporatedthe switch ON/OFF control, providing only timing forthe starting and duration (termination) of swing andstance phase for each leg. In this case the burden forproper timing is taken by the PT and later the patienthimself (once he controls the switches) and hence nomxternul aignal processing is required.

The FES sequence which, so far, has been found tobe most practical and has been mostly used, is given inFigure 3 . The figure shows traces 1-1' for M . Quadri-

ceps activation and traces 2-2' for the flexion reflexactivation (together, 4ohanna!u) . The required stimu-lation amplitudes are pre-set and maintained constant(readjustments are made only because of fatiguing)and set at an 80 percent level for maximum M . quadri-ceps stimulation and fairly above the threshold for justtriggering the flexion reflex . Also in Figure 3, the pro p-artirninOforthehandovvitohesiagiven ' andthastimu-lation (traces 3-3') for obtaining hip extension andabduction shown.

FIGURE 3The six-channel stimulation sequence using left (LS) and right (RS)side hand switches controlled by the P.T.

FE6-AMPLITUDE M .GLUTEUS MEDIUS

andM .TEN5ORFASCIAE LATAE

FLEXORRESPONSETRIGGERING

TIME 3 OFF

OFF

LEFT' M .QUADRICEPS

OFF

QUADR!CEP5

OFF

RIGHT

FLEXOR RESPONSF TRIGGER N6

L

OFF

RIGHT STANCE PHASE'

3 M . 5LUTEUS MEDI

LEFT HAND

ON-SWINGOFF-STANCE

SWITCH

US and Mfi TENSORFASCIAE LATAE

OFF

ON-SWINGRIGHT HAND OFF -STANCE

'SWITCH

DOUBLE

LEFT5TANC5

SINGLE,STANCEI

,-__..n..,

OFF

RIGHTSINnGI.E

STANCE

OFF


Applications in Three Patients

One of the first stimulation sequences for simplegait used on patient No . l e consisted predominantly ofmotoric stimulation . The gait cadence was determinedby the physiotherapist walking behind the patient ; thephysiotherapist pressed the switches and gave voicesignals to the patient for taking each step . The for-ward-leaning posture of the patient influenced thestep length . The forward swing of the leg was assistedby gravity (Figure 4) . Omitting channel 3–3' in Figure 3

simplified the hardware ; gait was still feasible but withmore assistance from the patient . Later, flexion reflex

' The three patients <mo .l ' No.2, and No .3) described here are theû,st-p° rforminm^,ubjoots from the 17-patient clinical series ofthis research program, as noted earlier. They are also discussedunder "General mmxvuomnyânu "Clinical 0unervauons^ ' em*whee in this paper .

triggering was improved by changing the electrodelocation, so that the gait improved because of longerstep length and better swing phase—the requirementfor forward-leaning posture was then eliminated.

The same FES sequence was used for the secondpatient (No .2) shown in Figure 5 . The site for flexionreflex stimulation was selected distally one centimeteranterior from the ouputfibu!ee at the site of N . peron-eus. This location requires higher stimulation ampli-tudes. Later, the site of active electrode placement inthe popliteal fossa over the peroneal nerve was used,providing larger flexion joint angles with lower stimu-lation amplitudes. The addition of stimulation chan-nels 3–3' (Fig .3) resulted in jerky movements of thetrunk because of strong hip-extension torques ; thepatient was able to walk in a soldier-like erect positionas shown in Figure 5.

FIGURE 4Patient No . 1 while walking (a) in double-stance phase and (b) in left swing phase . Excessive side lean can be seen .


FIGURE 5Patient No . 2 during double-stance (a) with applied electrodes for4- channel and J-channel stim0tion, auuuring svg phase.

In accordance with the findings for the third patient(No.3), a 4-channel FES sequence was adopted (M.quadriceps and flexion-reflex stimulation bilaterally).This patient also tried to control the stimulation eventshimself, using switches placed into the handles of aroller-walker. This arrangement enabled better coordi-nation of preserved body movements with the move-ments produced by the stimulation. In such gait theroller-walker was used for providing balance and par-tial It also afforded better mobility and hencelonger runs . This third patient is shown walking inFigure G . (In Figure 6 the knee strap fastens twoelectrodes .)

One would expect that the extension-patternedspasms would have interfered with the stimulation forobtaining flexion, but this did not occur for us in ourpatients—the flexion-reflex stimulation overrides theextension pattern . All of these three patients reportedthat FES daur*aoedthoir^spoatioity"

Most of the other SCI patients in the FES programalso reported decreased opaotioityaoaresult ofstimu-lation. In some patients, FES similar to that used for

the muscle restrengthening is used prior to walking forabout 30 minutes to reduce spasticity. in some pa-tients, FESe«eroining of hip extensors - Id abductorswas introduced for reducing addL sit'and/or hip-flexion spasms, particularly in patients 'th shortenedhip flexors because of prolonged sitting . During thestanding-up the hip flexors are stretched,which fre-quently triggers hip-flexion and abdominal spasms.With adequate hip-flexor stretching, this problem canbe minimized and in a period of 2– 4weeks, overcome .

PAGE 12

FIGURE 6No. 3 patient walking outside . Good stability but excessive pelvis shifting can be seen (a,

c).

CLINICAL OBSERVATIONS

Patient Selection : A Backward Look

As noted earlier, because of lack of previous experi-ence no patient-selection criteria were available . Ini-tially, traumatic lesion patients were selected, havingno rnedical prob!erno, and with a stable spine . Theywere all well along in their rehabilitation . Most of theselected patients were neurologically characterized aspatients with complete lesions.

It has proved desirable to minimize the relativelylong muscle restrengthening time . This time appearsto be shortest in patients with slight spasticity andwho start approximately 6—12 months after injury.These patients in most cases did not have substantialdisuse atrophy and no contractures or other problems(e .g ., no pressure sores, obesity, etc .),

In most cases patients having mild spasms turnedout to be the best suited because of the well-preservedmuscles . (Patients with severe spasticity also had well-preserved muscles, but in most cases they also hadcontraindications such as "tight" joints or even con-tractures .) Patients having pressure sores, or who areobese, or not cooperative, or who have contractures,etc ., are least suitable! In Table 1 the major indications

Professor Kral) has pointed out that, while patients with pressuresores and other problems were not used in the research reportedhere, such individuals are not necessarily excluded from the use ofFES once their problems are cured, if they then conform with theselection criteria as outlined in Tables 1 and 2-Editor .

for patient selection are given, and in Table 2 the majorcontraindications are listed : these lists reflect the data,experience, and opinions acquired and developedover the course of the work described here.

From a neurological point of view, the most suitablecandidates are patients having some preserved inner -vation to the long upper-back muscles, with respira-tory reserve and some trunk control . The data so farhave indicated that patients with lesions between T5and T12 are the most suitable candidates.

If during the initial evaluation of the patient the left-side right-side muscle forces are found to be sub-stantially different, those patients are not suitable forFES.

Frequently patients have joint stiffness with associ-ated spastic muscles, such as hip-joint flexion contrac-ture and plantarflexion contracture associated withspastic M . gastrocnemius and M . soleus. In most casesthe spasticity can be reduced with FES exercising (18,25)i The stimulation of agonist muscles can inhibitspasticity by reciprocal inhibition of the antagonists,or the stimulation of spastic muscle can result in low-ering of spasticity by several possible mechanismsdiscussed elsewhere (25).

The muscle-restrengthening FES procedure is com-bined with FES for inhibition of spasticity if necessary.Based on the severity and the synergy of spasticity, anappropriate FES pattern is selected . We frequently usethe stimulation pattern of unilateral stimulation of M.quadriceps (if needed, other extensors are added),combined with opposite-side flexion-reflex withdraw-

13

at movement stimulation.

Restrengthening : Procedures, Effects

The duty cycle for a single-muscle restrengthening(e.g ., M . quadriceps) is selected with a ratio such as 1 :2(2 seconds on and 4 seconds off), or perhaps 1 :3.Shorter duty cycles (longer stimulation times) result inrapid neuromuscular function transmitter depletion,and cessation of movements . We believe the aim ofmuscle restrengthening is better served if the muscleshave the opportunity to become fatigued because ofwork; that is why longer duty cycles (4 s on and 8–12 soff) seem to be more appropriate.

It should be stressed that at this time the optimalrestrengthening procedure which would provide thebest force increase and longest fatiguing time is notknown . Nor do we know how to select the stimulationfor best spasticity inhibition, or how in general to fitthe FES program to the patient for optimal results.Obviously the initial joint angles, duration and timingof FES loading of muscles, etc ., are important, but atpresent no scientifically proved criteria are availablefor their selection. However, in nearly 80 percent of the17 treated patients of the present study, the musclestrength increased and fatiguing decreased . In addi-tion, we observed improved skin condition which weattribute to improved blood supply. Similiar resultshave been noted and reported by other researchers (2,3, 6, 8, 9) . Additional research is needed to providemore knowledge and for establishing the methodolo-gy for prediction of results of the FES strengtheningprogram.

The results for spasticity reduction associated withFES are very unpredictable . At this point, measurableimprovements can be observed in only 50 percent to60 percent of the patients.

Three Patients : A Record of Improvement

The data from the three patients mentioned earlier(from our series of 17) are described in detail here.

Patient No . 1, age 31, had a complete traumaticlesion at T5, caused by automobile accident . The pa-tient was admitted to the program 12 months postin-jury after the completion of standard rehabilitation.

Patient No . 2, age 25, with a traumatic completelesion at T10, was admitted to the FES program aftercompletion of rehabilitation, several monthspostinjury.

Patient No . 3, 23 years old, suffered a car accidentresulting in complete lesion at T5 and was admitted tothe FES program two years after the injury.

Up to now only these three patients have been close-ly examined with regard to the feasibility of gait resto-ration by means of FES . The results of this gait restora-tion are here described chronologically—with eachpatient we gained some knowledge and were able to


FIGURE 7The vertical ground-reaction force vector location . Points 1, 2, 3 and6 for quiet standing ; the points 7, 8, 9 and 4, 5 have resulted becauseof weight-shifting caused by FES.

improve the application method.The first patient of these three underwent extensive

testing and training, starting in August, 1978, beforeattempting the first step. Much effort was made inregard to safety and to ensure that no injury was likely;this was particularly important because of the possi-bility of collapsing, which could cause more injury.After prolonged standing (of at least 20–30 minutes)was mastered and the patient was familiar with theerect standing posture, teaching for single-limb stancebegan.

In Figure 7, the points 1, 2, 3, and 6 represent thepositions of reaction for quiet standing . Points 7, 8, 9,and 4 and 5, represent the shifting of ground-reactionvector caused only by FES of hip abductors . Clearly,transfer of ground-reaction vector beneath a single legis hard to produce by use of FES alone . Therefore, inmost cases, the patient was taught to assist the bodyweight transfer to one limb by means of arm control.FES of hip abductors was in most cases thenunnecessary.

At this stage the patient started to walk in parallelbars . Good physiotherapy assistance was secured.The physiotherapist, behind the patient, verbally in-structed the patient to transfer the body weight withcommands like "left" and "forward right", etc ., at thesame time pressing the switches for terminating thestance-phase stimulation in the leg which instantlyreceived stimulation for the swing phase as described

PAGE 16


FIGURE 1 1The average ground-reaction forces record . It is seen that the sup-porting force is mainly located at the arch region of the foot, and theforce supporting vector displacement is narrow and does not use, a sin normal gait, the whole sole length.

FIGURE 10The consecutive records of goniograms for 2 left steps and 3 right steps of No . 3 patient.

with the knees extended and locked . The goniogramsof this gait mode have no similarity with plots ofnormal gait ; that fact is more easily recognized if weanalyze the ground-reaction forces record shown inFigure 11 . The average record is shown for the first runof (NK=20) 20 steps . The average stance time lasts 82percent and the swing time is only 18 percent of thestride cycle, compared to normals where the corre-sponding numbers are 60 percent versus 40 percent.The double-stance time lasts 27%–32% of a stride . InFigure 11 we recognize that the horizontal axes timeinterval 0—27 corresponds to the double-stance time,the horizontal time interval 27—47 to the swing phase,and the duration 0–84 to one step vertical force record(starting by the lower part and ending by 84-lowertrace) . The time scale is given for the records in Figures

11, 12, and 13 in percentages of the stride time (whichis 11 .2 s).

Figure 9 shows a relatively narrow field of standarddeviation and the same is true of the vertical force plotin the ground-reaction record given in Figure 12 for thesecond run. In Figure 12 the average ground-reactionvertical force is displayed (solid line) for 26 steps ; thestandard deviation range is displayed by the dottedline.

In Figures 11, 12, and 13 the locations of the force

FIGURE 12The ground reaction forces displayed together with the standarddeviation for 26 steps . Also the spatial distribution of the center ofpressure is shown .

FUN] -

40

200

,„„,aft.

0

ID0

120

29-10-81/6

40 80

. .. ... . . . . .. .

TED11TEDI2TEDI3TED I4

ID NK 100N SD

20 10 39 64

TI

50

17

AVERAGE

STANDARD DEVIATION

-ae

1 D NV. 100' SD TI

25 10 42 901 04

.e 2 e

mu3

•<,

0,

TEDISTEDI6TEDI7TEDI8

PAGE 18


FEN]

4O .

200

140 '

120

29–10–81/6 29– 10–81/6

TEDI1

TEDISTEDI2

TEDI6TEDI3

TEDI?TEDI4

TEDIS

ID NK 100% SD TI

20 10 .39 .G4 .5010 .42 .90 .88

0z0 ' 60'40 '

60 '

FIGURE 13The comparison of ground-reaction forces plots for run 1 and run 2.Good repeatibility and symmetry can be seen (solid line = first run,dashed line = second run).

pick-up transducers (all together eight on each force-shoe) are shown together with the shoe contour . Theforce shoes are made by Kljajic, et al . (26, 27) at theJosef Stefan Institute in Ljubljana and have proved tobe very convenient for statistical kinetic gait studies.

In Figure 12 also, the ground-reaction vertical forcevector trajectory is plotted along the shoe sole . Theheight of the vertical lines (Figure 13) pictured on theforce-vector trajectory plotted across the shoe solerepresents the actual force amplitudes and their loca-tions in time (intervals of TI=0 .50 in Figure 13) . InFigure 12, the rectangles plotted onto force displace-ment trajectory display the standard deviation rangefor both X and Y displacements .

The average records for vertical force and verticalforce-vector displacements versus distance and timefor both runs are plotted in Figure 13 . Again it can beseen that the gait was symmetrical, and showed goodrepeatability . The ground-force supporting vector dis-placement is very narrow and relatively stationary . Fornearly 90 percent of the stride time, the vector re-mained in the area of the foot arch.

Patient No . 3, even after walking for a relatively longtime, (60 minutes or more), did not complain of beingtired, but at this state of our work we have not yet triedto confirm the assumption that the FES-generatedbiped gait does have the important characteristic oflow power consumption which is associated with thenormal biped gait mode .

19


DISCUSSION

This preliminary research work should be under-stood as a feasibility demonstration . The work de-scribed has shown that perhaps as many as 80 percentof paraplegic patients who conform with the selectioncriteria (Table 1 and Table 2) are able to stand forprolonged periods by means of surface FES, andstanding balance sufficient for performing single-limbsupport required during biped gait has been routinelyachieved . The patients were able, in spite of theirrelatively high lesions (even at T5) to control the trunkalignment by means of trunk-leaning movements withthe help of their remaining voluntarily controlled longbody muscles (M. erectus spinae, M . rectus abdo-minis, M . latissimus dorsi, etc .) and the hands.

The multichannel surface FES described did enableparaplegic patients to exert simple biped gait patternsif they were assisted by parallel bars or a walker. It wasdemonstrated that the patients were able to walk forreasonably long periods and over moderate distances.One patient ambulated even if unattended.

Additional research is needed to refine and improvethis method to make it functional and safe for daily useby paraplegic patients. If this work is pursued it seemsthat functional standing enabled by FES need not bevery far from being implemented in daily life.

Implanted Electrodes : the Probable Method—Theexperience gained so far has shown that patients 'acceptance will probably usually be better if FES isdelivered by means of an implanted stimulator . Andalthough as investigators we hope that much moreknowledge regarding the restoration of paraplegic pa-tients' gait enabled by FES can be obtained by usingsurface electrodes, we nevertheless believe the ortho-tic FES via implanted selfcontained units will prove tobe the practical and suitable method for the daily useby patients.

Preserved Reflexes and Feedback

The paraplegic patient represents a unique andcomplicated system not yet fully understood . The re-sults reviewed here have shown clearly that researchis warranted toward better understanding and controlof the biomechanical events associated with gait res-toration by means of FES . Together with this, moreinformation must be gained regarding the functionaluse of preserved reflex organization of the transectedspinal cord . Research for implanted multichannel FESdevices, and nerve electrodes, must be stressed.

In regard to biomechanical problems, we intend toconduct research in the biomechanics of standing-up,weight-shifting between supporting legs, and themuscle coordination associated with different eventsin gait . Research aimed toward the possibilities of

increasing gait speed, biomechanical understanding,improved control, stability, and other events in a multi-channel FES-assisted gait is needed . Further attentionshould be paid particularly to the biomechanical prob-lems involved in controlling gait direction changes byFES.

All the described activities might be performed bet-ter and more simply if biofeedback could be added.And for all the mentioned topics, the control problemsshould be studied with the goal of developing patient-activated control . At the end of our discussion, withoutarguing, it can be imagined that FES-assisted paraple-gic gait is a new alternative mode for locomotion in therehabilitation of some of these patients, and a meansof overcoming barriers posed to them

REFERENCES

1. Kantrowitz A (1963) Electronic Physio-logic Aids . A Report of the Maimon-ides Hospital of Brooklyn, NY.

2. Willemon WK, Mooney V, McNeal D,Reswick J (1971) Surgical ImplantedPeripheral Neuroelectric Stimulation.Internal report, Rancho Los AmigosHospital, Los Angeles, CA.

3. Kralj A, Grobelnik S (1973) Functionalelectrical stimulation—a new hopefor paraplegic patients? Bull ProsthRes BPR 10-20, Spring 1973, pp . 75–102.

4. Molskaja NE . (1979) Basic principlesof electrical stimulation in spinal cordinjured patients (in Russian) . Proc.2nd All-Soviet Conference onElectrostirnulation of Organs and Tis-sue . Eds : Kolesnikov, GF. Kiev, USSR,p 85.

5. Peckham H, Mortimer T, Marsolais E(1976) Alteration in the force andfatigability of skeletal muscle in qua-driplegic humans following exerciseinduced by chronic electrical stimula-tion . Clin Orthop 114 :326–334.

6. Mortimer T, Peckham H (1973) Intra-muscular Electrical Stimulation . Neu-ral Organization (and its Relevance toProsthetics) . Fields WS, Leavitt LA,Eds . New York, Intercontinental Medi-cal Book Corp . pp 131–146.

7. Mortimer T, Kaufman D, RoessmannO (1980) Intramuscular electrical stim-ulation : tissue damage . Ann BiomedEng . 8(3) :235– 244.

8. Brindley S, Polkely C, Rushton D(1978–79) Electrical splinting of theknee in paraplegia . Paraplegia 16(4):428–435, February 1979.

9. Kralj A, Bajd T, Turk R (1980) The influ-ence of electrical stimulation on mus-cle strength and fatigue in paraplegia.Proceedings International Confer-ence on Rehabilitation Engineering,Toronto, pp 223–226.

10. Kralj A, Bajd T, Turk R, Benko TH(1979) Paraplegic patients standingby functional electrical stimulation .

PAGE 20


XII Int . Conf . on Medical and Biologi-cal Engineering, Jerusalem, Israel, p.59 .3.

11. Kralj A, Bajd T, Turk R (1980) Electricalstimulation providing functional useof paraplegic patient muscles . MedProgr Technol 7(1) :3-9.

12. Bajd T, Kralj A, Turk R (1982) Stand-ing-up of a healthy subject and a para-plegic patient . J . Biomech 15(1) :1-10.

13. Kralj A, Trnkoczy A, Acimovic R (1971)Improvement of locomotion in hemi-plegic patients with multichannelelectrical stimulation . Human Loco-motor Engineering, Institution of Me-chanical Engineers, London, 45.

14. Strojnik P, Kralj A, Ursic I, (1979) Pro-grammed six-channel electrical stim-ulator for complex stimulation of legmuscles during walking, IEEE Trans.Biomed Eng BME 26(2) :112-116.

15. Stanic U, Acimovic R, Gros N,Trnkoczy A, Bajd T, Kljajic M (1978)Multichannel electrical stimulationfor correction of hemiplegic gait.Scand J Rehabil Med 10(2) :75-92.

16. Cerny K, Waters R, Hislop H, Perry J(1980) Walking and wheelchair ener-getics in persons wih paraplegia.Phys Ther 60(91 :1133-1139.

17. Vodovnik L, Bajd T, Trnkoczy A, KraljA, Gracanin F, Strojnik P (1981) Func-tional electrical stimulation for con-trol of locomotor systems, CRC Criti-cal Reviews in Bioengineering, pp63-131.

18. Bowman B, Bajd T (1981) Influence ofelectrical stimulation on skeletal mus-cle spasticity . In : Advances in ExternalControl of Human Extremities . Popo-vic D, Ed. Belgrade, Yugoslov Com-mittee for Electronics and Automa-tion, pp 567-576.

19. Rebersek S, Vodonik V, Gros N, Ste-fanovska A, Acimovic R (1981)Electrotherapy of spastic muscles inhemiplegia, as above, pp 217-228.

20. Bajd T, Bowman B (1982) Testing andmodelling of spasticity, J Biomed Eng4(2) :90-96.

21. Kralj A (1981) Electrical stimulation oflower extremities in spinal cord in-jury. Proceedings of the NATO Ad-vanced Study Institute in Spinal CordInjury Rehabilitation Engineering,Stoke Mandeville Hospital, Ayles-bury, England . (in press) .

22. Lee K, Johnston R (1976) Electricallyinduced flexion reflex in gait trainingof hemiplegic patients : induction ofthe reflex . Arch Phys Med Rehabil57 :311-319.

23. Kralj A, Bajd T, Kvesic Z, Turk R (1981)Electrical stimulation of incompleteparaplegic patients . Proceedings, 4thAnnual Conference on RehabilitationEngineering, sponsored by Rehabili-tation Engineering Soc . of N . America(RESNA), Washington DC, pp 226-228.

24. Bajd T, Hufford P, Lunsford B, NelsonH (1980) Spinal Injury . Rancho LosAmigos Rehab . Engineering Center,Downey, Calif ., Annual Reports ofProgress, pp 19-20.

25. Vodovnik L, Van der Meulen JP (1970)Modelling of spasticity and its com-pensation by electrical stimulation.Proc . Int . Symp. External Control ofHuman Extremities, Yugoslavia Com-mittee for Electronics and Automa-tion, Belgrade, pp 503--517.

26. Krajnik J, Kljajic M, Malezic M, StanicU, Stopar M, Acimovic R, Gros N(1981) Ground reaction measuringshoes as an aid in the clinical assess-ment of pathological gait . Proc. IntSymp . External Control of Human Ex-tremities, Yugoslavia Committee forElectronics and Automation, Bel-grade, pp 380--389.

27. Kljajic M, Krajnik J, Trnkoczy A (1979)Determination of ground reaction andits distribution on the foot by measur-ing shoes . Digest of the 12th Int . Conf.on MBE, (International Fed . for Medi-cine and Biology) Jerusalem.

28. Bajd T, Kralj A, Turk R, Benko H, SegaJ (1982) A four-channel electrical sti-mulator enabling the walking of para-plegic patients, submitted for print inJ Phys Ther.

29. Jaeger RJ & Kralj A (1982) Functionalelectrical stimulation changes jointcompliance . 5th Annual Conf. on Re-habilitation Engineering, Houston,Texas, Aug . 1982 . Rehabilitation Engi-neering Soc. of North America(RESNA)—Abstract .

Documents

Gait Restoration in Paraplegic in Paraplegic Patients: A ... · Recent advances in science have aided research toward the resto-ration of biped gait in paraplegic patients by means