Paraplegia 32 (1994) 54-553 1994 International Medial Soiety of Paraplegia The effets of parallel bars, body weight support and speed on the modulation of the loomotor pattern of spasti pareti gait. A preliminary ommuniation M Visintin MS & H Barbeau PhD Shool of Physial and Oupational Therapy, MGill University, Montreal, Quebe, anada. The effets of walking with and without parallel bars, providing 4% body weight support (BWS) and inreasing speed on the gait pattern of spasti pareti subjets during treadmill loomotion were investigated. In asymmetrially involved subjets, walking without parallel bars led to a more symmetrial gait pattern with dereased ompensation of the less involved side. This was aompanied by hanges in eletromyographi (EMG) and sagittal angular displaement profiles whih favoured a more normal swing phase of the more involved limb. When symmetrially involved subjets walked without parallel bars, inreases in EMG ativity, with prolonged ativation during the stane phase were noted, espeially in the distal musles. Providing 4% BWS failitated gait when walking without parallel bars espeially in the asymmetrially or severely involved subjets who showed marked diffiulty at % BWS. Forty perent BWS led to a derease in lonus assoiated with walking without parallel bars. Higher treadmill speeds inreased lonus in some subjets while in others it only aused a small inrease in EMG amplitude. Impliations for gait training are disussed. Keywords: spasti gait; parallel bars; body weight support; speed; eletromyography; spinal ord injury. Introdution Inomplete spinal ord lesions in man ommonly result in disturbanes of the loomotor pattern. Altered reruitment patterns with premature musle ativation, as well as prolonged eletromyographi (EMG) ativity and delayed musle relaxation have been identified in spasti pareti gait.l The pattern may also be oupled with flattening of the EMG profiles, haraterized by diminished or abolished peaks of EMG ativity. Knutsson2 identified (1) early streth ativation of distal musles, (2) paresis and (3) abnormal oativation of agonist and antagonist lower limb musles in response to loading, as likely to interfere with gait following a spinal ord lesion. oupled with orrespondene: Martha Visintin MS, Jewish Rehabilitation Hospital, Physiotherapy Department, 325 Plae Alton Goldbloom, homedey, Laval, Quebe, anada H7V 1R2. disturbanes in loomotor programming, external fators suh as the use of parallel bars, body weight support (BWS) and walking speed an also influene the loomotor pattern. Suh external fators an be manipulated when retraining gait, aiming at optimizing the loomotor output. Although widely used in the linial setting, the use of parallel bars during gait training appears ontroversial. onrad et al1,3 report that the dereased stability inurred during treadmill loomotion without parallel bars leads to a deterioration in the gait pattern of spasti pareti subjets. In ontrast, onventional gait training for neurologial patients disourages the use of parallel bars as it is believed to lead to an asymmetrial gait with ompensation of the less involved side. 4 The use of BWS, provided by an overhead harness whih supports a perentage of body weight as subjets walk on a treadmill, has been proposed to retrain gait following a
Paraplegia 32 (1994) 54-553 spinal ord lesion. BWS has been shown to failitate gait and eliit a more normal gait pattern with respet to sagittal angular displaement patterns, temporal distane parameters, and EMG ativity of lower limb musles in a group of spasti pareti subjets during treadmill loomotion at their omfortable speed.5 Futhermore, with BWS, subjets were able to walk at higher omfortable treadmill speeds. The harateristi slow walking speeds evident among spasti subjets5,6 is an additional onern when retraining gait. Gait training aims at inreasing the walking speed, attempting to make the gait more effiient. However, few quantitative data exist to eluidate the effets of walking speed on a spasti pareti gait pattern. It is evident that a better understanding of the effets of external fators suh as parallel bars, BWS and speed on spasti pareti gait is required before a omprehensive gait training strategy an be developed. It is the aim of this study to quantitatively and qualitatively desribe the effets of walking with and without parallel bars, with BWS, and at inreased treadmill speeds on the loomotor pattern of spasti pareti subjets. Methods The study was onduted in the human gait laboratory whih has previously been desribed in detail. 7 Eight spasti pareti subjets, ranging in age from 22 to 42 years (mean = 27.5 years), partiipated in this study. Seven subjets had sustained a traumati inomplete spinal ord lesion to the ervial or thorai spine, one suffered from nonfamilial progressive spasti paraparesis and one subjet had a surgially indued lesion at the level of TlO following resetion of a spinal tumour. Eah subjet was apable of independent overground loomotion with or without the use of external aids. The hroniity of the lesion ranged from 7 months to 21 years. The demographi data, presene of ankle lonus at rest, symmetrial involvement of the lower limbs, omfortable and maximal treadmill speed, and an overground ambulation profile for eah subjet are summarized in Table I. A study of spasti pareti gait 541 Body weight support The subjets walked on a treadmill while % BWS (full weight bearing) and 4% BWS were provided. The BWS apparatus onsisted of a ustom-designed harness whih mehanially supported the patient vertially over the treadmill. The harness onsisted of a pelvi band attahed around the hips and two padded straps whih pass between the legs to attah anteriorly to the pelvi band. The perentage of BWS provided was alibrated using a fore transduer. The fore was normalized to eah subjet's weight (1% ) and the sequene of % BWS provided (% or 4% ) was randomly assigned into two trials given within the same experimental session. Prior to data olletion, eah subjet was habituated at % BWS for 1-5 minutes aording to his walking tolerane. During this pratie session, treadmill speed was slowly inreased from. ms-1 up to eah subjet's omfortable walking speed. It was also determined whether the subjets ould walk at more than one treadmill speed (Table I). All subjets, exept subjet 2, were able to walk at a minimal and at a omfortable treadmill speed. Four subjets (1, 3, 5, 8) were able to walk at a maximal treadmill speed. Data were then olleted for predetermined speeds (minimal, omfortable and maximal speeds) during subtrials at % BWS with parallel bars. Data were also olleted at % and 4% BWS, without parallel bars at minimal or omfortable speeds (subjets: 1, 2, 3, 6). Of the eight subjets, only one (8) required a short leg brae to ontrol foot drop of the left ankle while walking on the treadmill. A 1 minute rest period was given between eah BWS trial to minimize fatigue. Blood pressure and pulse were monitored following eah trial to ontrol for undue stress on the subjets. EMG and Jootswith data EMG ativity was reorded from the gluteus maximus (GM), vastus lateralis (VL), medial hamstrings (MH), tibialis anterior (TA), medial gastronemius (GA), and lateral soleus (SOL) of either the left (subjets 1, 4, 5, 6, 7) or right (subjets 2, 3,
.. :s Table I Details of subjets S s- Subjet Lvel hroniity Ankle Symmtrial omfortable Maximal Ovrground walking." Sex of (years) lonus involvement treadmill treadmill aids used ;::: Age lesion sped (ms-i) speed (ms-i) 1 SP 3. +(R) Yes.4.6 Bilatral foot M drop bras 32 ++(L) 2 6-7 1.5 +++(Bil) (R> L) O.OS.8 (L) long Ig brae; M anadian ruthes 23 3 TlO 1.5 +++(R) (L> R).1.3 (R) foot drop F ++(L) brae; 1 ane 22 4 6-7 4. +++(Bil) Ys.4.4 2 ans M 23 5 7-Tl 1.5 +++(Bil) Yes.15.25 Walkr M 32 6 I-2 21. ++(Bil) Yes.43 Not No aids F tsted 26 7 TS-9 1.5 Yes.15.15 2 anes 2'..., M 2 f S T9-1.6 +++(Bil) (L> R).1.15 (L) foot drop tjol N M brae; walker 42 SP = spasti paraparesis; + = mild; ++ = moderate; +++ = severe. Ul.j:>. N \::l.. ti::i.".., -." ::: 'D 'D u. +- I u. u.
Paraplegia 32 (1994) 54-553 8) lower limb while the subjets walked on the treadmill. When subjets presented with asymmetrial involvement of the lower extremities (i.e. one limb showing minimal spastiity and weakness), data were olleted from the more involved lower limb. When the degree of involvement of the lower limbs was largely symmetrial, data were then olleted from the limb that showed a greater degree of linial spastiity. Bipolar surfae eletrodes (2.5 m enter to enter) were plaed over the belly of eah musle following onventional skin preparation. The EMG signals were preamplified, differentially amplified and bandpassed (1-47 Hz). Footswithes plaed under the heel, fifth metatarsal head and big toe of eah subjet's shoes were used to detet heel strike, foot-flat and toe-off, and to determine the temporal distane parameters. The EMG and footswith signals were then reorded at 3.75 IPS on a 14 hannel FM tape with a frequeny response of 25 Hz. The EMG signals of the six musles along with the footswith signals were then played bak on a polygraph and an artefat-free sequene of 1 or more onseutive yles was hosen to represent eah experimental paradigm for eah subjet. Joint angular displaement data Joint angular displaement data were olleted from the same limb as the EMG reordings for eah subjet. The subjets were videotaped as they walked on the treadmill using a shutter video amera. Refletive joint markers were plaed at the shoulder, hip, knee and ankle as well the heel, fifth metatarsal head and the toe region of the lateral border of the shoe. Additional markers were plaed on a vertial and a horizontal bar to be used as absolute oordinates for the video analysis. The trials were reorded on a 3/4 inh videotape at a speed of 6 fields per seond. A remote searh ontroller was used for field by field viewing. The sagittal angular displaements were manually measured from the monitor sreen using a goniometer. One the subjets had reahed a steady state while walking on the treadmill, A study of spasti pareti gait 543 one representative gait yle for eah subjet during eah experimental paradigm was analyzed. The joint angular displaements were measured at every 5% of the gait yle. The trunk and hip angles were alulated with respet to a vertial line, with the neutral position in standing being taken as displaement of the trunk and hip, flexion being positive, and extension negative. Likewise, in alulating the knee and ankle angles, the neutral standing position, with the knee in full extension and the shank axis perpendiular to the foot, was taken as. Knee flexion and ankle dorsiflexion beyond neutral were taken as positive angular displaements, and ankle plantarflexion beyond neutral was taken as negative angular displaement. Results Effets of parallel bars and BWS The effets of removing parallel bar support during treadmill loomotion at omfortable speeds, at % and 4% BWS will be addressed both in those subjets with an asymmetrial and a symmetrial gait pattern. Figure 1 illustrates the EMG patterns of subjet 2 who walked with an asymmetrial gait pattern due to marked impairment of the right lower limb and a near normal left lower limb. He walked on a treadmill with (Figure la) and without parallel bars (Figure Ib) at % BWS and without parallel bars at 4% BWS (Figure l) at his omfortable speed of.8 ms-i. While walking with parallel bars (Figure la), he ompensated for his lak of hip, knee and ankle flexion during the right swing phase by pushing on the parallel bars to ome up onto his left toes while swinging the right limb through fully extended. As illustrated in Figures le-g, there was no flexion at the hip or ankle and minimal flexion at the knee (11 ) during the right swing phase. The EMG profiles (Figure la) revealed ativity in most right lower limb musles during the stane phase. Low level ativity appeared in T A during stane with minimal ativity during the swing phase. During the same experimental session, the subjet was asked to walk without parallel bars (Figure 1 b). As
54 4 Visintin and Barbeau Paraplegia 32 (1994) 54-553 a % with bars b % without bars 4% without bars r"!'ifiiq "' II."J "llj,." +., oj GM...,".,.. +II.. '.1. ". ':11...,t,_ 11' 't4i_- f MH! _ttfl. rt, TA 11,1 -It I,... GA.""'I". :.-1.-l.-l,-----J.-l.-l,-----J - I 1M +"++.,,jot.- H \ \ I,----l.-1,-----J,---l,-l,-l,--J.-1 ff+ '1 I I o H I I... ':'.II... "j,----l,--i.-1.---1,----l,-l.-1,----l J 1).lV 1 se Figure la- aption on p. 54 5. he released the parallel bars, his legs dragged at the end of the treadmill following whih he was able to initiate and omplete only three laborious steps (Figure 1 b ). During these three steps a marked inrease in hip and knee flexion as well as ankle dorsiflexion emerged as the subjet was no longer able to ompensate by using the parallel bars, but required flexion of the right lower extremity in order to advane the limb forward (Figure Id-g). The most evident EMG hange was seen in T A during the swing phase where a burst of ativity appeared resulting in ankle dorsiflexion. Although walking without parallel bars was very diffiult for this subjet, it failitated a more normal swing phase, with flexion at the hip, knee and ankle. When 4% BWS was provided (Figure I) a muh smoother, less strenuous gait resulted, thereby allowing the subjet to take a greater number of steps (a minimum of 1 steps per trial). TA ontinued to show a burst of ativity during the swing phase. The sagittal angular displaement profiles (Fig Id-g) revealed a onsiderably straighter knee at foot-floor ontat (3 ) when ompared to % BWS (56 ). Ankle dorsiflexion was initiated earlier in late stane. Similar findings were observed in subjet 3 (not illustrated) who also walked with an asymmetrial gait pattern by ompensating with her less involved right side in a similar way to subjet 2. When the subjet walked without parallel bars at % BWS, a burst of ativity appeared in T A during the swing phase whih was not present when the parallel bars were used. This orresponded to an inrease in ankle dorsiflexion during the swing phase, similar to that seen in subjet 2. Failitation of gait with 4% BWS was also seen in subjet 3 when BWS resulted in a smoother, less spasti gait. Figure 2a- illustrates the effets of walking with and without parallel bars at % BWS and without parallel bars at 4% BWS in a symmetrial subjet (6) at.43 ms-i. Exept for prolonged VL ativation (Fig 2a), this subjet's lower limb EMG patterns were similar to those seen in
Paraplegia 32 (1994) 54-553 A study of spasti pareti gait 545 d 4 e 4 +---- --- ;\!! I O+r ----+---,l --2 +----,---,-----.-----. --, -3 +----. --,-----.-----. ---. o 1 1 7 f 9 3 o.!!? en Q) Q) o ". -1+--? --. --, o 2 4 6 8 1 2 en O-r.:T--J-----:o"'----+.!!? - «-4 +---- ----,-- ---.--------r4-, % Gait yle % Gait yle 4 6 8 1 H % with bars _. % without bars - 4% without bars Figure 1 The right lower limb EMG ativity of subjet 2 walking on the treadmill, at a speed of.8 ms-i, at (a) % BWS with parallel bars, (b) % BWS without parallel bars, and () 4% BWS without parallel bars. The downward arrows indiate foot-floor ontat, while the upward arrows indiate toe-off, with the solid line depiting stane duration and the spae denoting swing duration, for both right (R) and left (L) lower limbs. In (b), the seond, third and fourth steps of the right lower limb represent the three steps the subjet takes without parallel bars. In (b) and (), note the burst of ativity in TA during the swing phase. In (), note that the subjet is able to take a greater number of steps without parallel bars when 4% BWS is provided. The orresponding sagittal angular displaement patterns of a representative yle for the (d) trunk, (e) hip, (f) knee and (g) ankle are also illustrated. Note the inrease in hip and knee flexion and ankle dorsiflexion during the swing phase when walking without parallel bars and the straighter knee at foot-floor ontat when 4% BWS is provided. normal subjets.8,9 When the subjet walked without parallel bars (Fig 2b), an inrease in EMG ativity of all lower extremity musles was noted espeially at foot-floor ontat. T A showed an inrease in toni ativity during stane. A broadening of both the GA and SOL bursts were noted with early ativation at foot-floor ontat. This was aompanied by a large inrease in EMG amplitude espeially in the GA musle. Exept for a small derease in plantarflexion at the ankle at the end of stane, walking without parallel bars resulted in minimal hanges in lower limb angular displaement profiles in this subjet (Fig 2d-g). Similar findings were noted in subjet 1 (not illustrated). In subjet 6, providing 4% BWS during treadmill loomotion without parallel bars resulted mainly in a derease in EMG amplitudes for all lower extremity musles. This is exemplified in Figure 2 where the EMG profiles losely resembled those at % BWS with parallel bars, illustrated in Figure 2a. Minimal hanges were noted in the sagittal angular displaement patterns when ontrasting % and 4% BWS in this subjet (Fig 2d-g). A similar trend was noted in subjet BP at his minimal treadmill speed of.2 m-i. Providing 4% BWS also failitated gait without parallel bars in two more severely impaired subjets (5, 7) with a symmetrial gait pattern. These subjets were not able to
546 Visintin and Barbeau Paraplegia 32 (1994) 54-553 a GM VL MH TA % with bars '..., fi.f.tl +' 1 +1 'HH GA. ",t,. jill SOL. "'.I'.I'IIIIi.'.II' L R -"" ) OJ o,::.-l d 4 o.' b % without bars '.t. ffitt# - I t!,,t I.'.'.' r e 4 ) OJ o Q. ". "'t- -'I/r" I 4% without bars.4t... Ift t*.'. + "II+.'* '..,.,!.. ill" I"... J 1!-IV rr r 1 se -2 f 7-3 1 9 3 1 OJ o ) ) -1 2 4 6 -"" ) OJ o <l: -4 8 1 2 4 6....... 8 1 +--.. % with bars.--.. % without bars -- 4% without bars % Gait yle % Gait yle Figure 2 The left limb EMG ativity of subjet 6 walking on the treadmill, at a speed of.43 ms-1, at (a) % BWS with parallel bars, (b) % BWS without parallel bars, and () 4% BWS without parallel bars. The downward arrows indiate foot-floor ontat, while the upward arrows indiate toe-off, with the solid line depiting stane duration and the spae denoting swing duration, for both left (L) and right (R) lower limbs. In (b), note the inrease in VL ativity as well as the broadening of ativity in TA, GA and SOL during stane. The orresponding sagittal angular displaement patterns of a representative yle for the (d) trunk, (e) hip, (f) knee and (g) ankle are also illustrated.
Paraplegia 32 (1994) 54-553 walk without parallel bars at % BWS but were able to do so when 4% BWS was provided. BWS versus parallel bars In those subjets (2, 3) who walked with an asymmetrial, ompensatory gait pattern, 4% BWS appeared to have the most positive effets on the gait parameters when the subjets walked without parallel bars. This is illustrated in Figure 3a-g whih shows the effets of 4% BWS with and without parallel bars. As subjet 2 walked on the treadmill with parallel bars at % BWS he ompensated by supporting himself on the parallel bars, oming up onto his left toes during left stane, while swinging the right leg forwards with a minimal amount of hip, knee and ankle flexion. When 4% BWS was provided, while he walked with parallel bars, minimal hanges were noted in the gait pattern (Figure 3b, d-g) as the subjet ontinued to ompensate. However, when the subjet released the parallel bars A study of spasti pareti gait 547 while walking at 4% BWS, a more normal gait pattern emerged. This was haraterized by an appropriate swing phase with flexion at the hip, knee and ankle (Fig 3-g) and a burst of ativity in T A (Fig 3). Effets of speed Inreasing treadmill speed, at % BWS while walking with parallel bars, resulted in an inrease in lonus of the distal musles for subjets 3, 7 and 8. In subjet 1, lonus was not present at the slowest speed, but was eliited at the higher speeds. An example is illustrated in Figure 4a-g for subjet 8 with inreases in treadmill speed from. 5 ms-1 to.15 ms-i. At the lowest speed of. 5 ms-1 (Fig 4a), oativation of the extensor musles was noted. There was persistent ativity in GM and VL throughout stane. Minimal ativity was evident in MH. A burst of ativity during the swing phase was noted in T A. GA was haraterized by low toni ativity during stane with a GM % with bars..i..!.., -.,'. b 4% with bars.. 4% without bars MH TA t f /.rt " " -, "M t.'., I.... 4.. _ III.j GA... "I I,--J,--J f-----1.----l. " It.----l,--J,--J,--J.---l '.,.t+ "j +.., 1-1". r"!..... 'I"llIj',......----l,--l.---l,-----J,--J J 1 IV 1 se Figure 3a- aption on p. 548.
548 Vis in tin and Barbeau Paraplegia 32 (1994) 54-553 d 4 8 4 ) J) <1l - :::J ) J) <1l Q. I -2 1-3 1 f 7 9 3 ) J) <1l ) ) ) J) <1l ) 32 «- % with bars 4% with bars -4+-------r---'----- - 4% without bars o 2 4 6 8 1 % Gait yle % Gait yle Figure 3 The right lower limb EMG ativity of subjet 2 walking on the treadmill, at a speed of.8 ms-l, at (a) % BWS with parallel bars, (b) 4% BWS with parallel bars, and () 4% BWS without parallel bars. The downward arrows indiate foot-floor ontat, while the upward arrows indiate toe-off, with the solid line depiting stane duration and the spae denoting swing duration, for both right (R) and left (L) lower limbs. In (), note the more phasi EMG ativity in VL and the appearane of a burst of ativity in T A during swing. The orresponding sagittal angular displaement patterns of a representative yle for the (d) trunk, (e) hip, (f) knee and (g) ankle are also illustrated. Note the presene of hip and knee flexion and ankle dorsiflexion at 4% BWS without parallel bars. the onset of low amplitude lonus shortly after initial foot-floor ontat. SOL also showed lonus during that period, followed by a burst of ativity ontinuing until terminal stane. Inreasing the treadmill speed to. 1 ms-1 and. 15 ms-l (Fig 4b, ) resulted in sustained lonw; of all musles during the entire stane phase and, in some instanes, during early swing. This resulted in a laboured gait, espeially at the highest speed, haraterized by inreased diffiulty walking and loni osillations visible at the ankle. Figure 4d revealed a straighter trunk alignment at the higher speeds for subjet 8. Minimal hanges in the hip angular exursion patterns were evident (Fig 4e). At the higher speeds an inrease in knee flexion during mid to terminal stane was noted, with a small derease in maximum swing angle (Fig 4f). A hange in the ankle kinemati profiles was the ourrene of plantarflexion during midstane as the subjet went up on his toes to failitate swinging the left leg through (Fig 4g). An inrease in plantarflexion at push-off was also noted. In subjets 4, 5 and 6 who showed minimal abnormal reflex ativity in the distal musles during loomotion, inreases in speed had a minimal effet on EMG timing and amplitude (not illustrated). The effets were to produe a small inrease in EMG amplitude for lower extremity musles suh as the GM, TA, GA and SOL. Qualitatively, the gait appeared more laboured at the higher speeds.
Paraplegia 32 (1994) 54-553 A study of spasti pareti gait 549 a.5 ms 1 b.1 ms.15 ms 1 GM VL MH TA GA SOL.11 1 se 1,1 I.\ J 1 V.!!! l :::J t=: 11-... -4'/) Q) Ol Q. I -2 1 f 7-3 9 3 1 l Q) Q) :><: -:::::....:.::::o::. +--- -1+------- o 2 4 6 8 1 % Gait yle Q) Ol Q) 32 «.......... -'It'- ''1-- -4+---_r----_r----r_--, o 2 4 6 8 1 % Gait yle..5 ms -1..1 ms 1.15 ms 1 Figure 4 The right lower limb EMG ativity of subjet 8 walking on the treadmill, with parallel bars, at (a) O.5ms-1, (b) O.lOms-l, and () O.15ms-1. The downward arrows indiate foot-floor ontat, while the upward arrows indiate toe-off, with the solid line depiting stane duration and the spae denoting swing duration, for the right (R) lower limb. Note the presene of sustained lonus at the higher treadmill speeds. The orresponding sagittal angular displaement patterns of a representative yle for the (d) trunk, (e) hip, (f) knee and (g) ankle are also illustrated.
55 Visintin and Barbeau Disussion Effets of parallel bars and body weight support Ambulating without parallel bars has been deemed by some investigators to aentuate gait disturbanes in spasti syndromes.1, 3,lO In this study, removing parallel bars in symmetrially involved spasti pareti subjets resulted in an inrease and prolongation of EMG ativity during stane. onrad et al3 attributed omparable EMG hanges, in a group of spasti pareti subjets, to nonspeifi protetive gait mehanisms present during instanes of emotional stress or anxiety when the subjet's stability is threatened. However, in the present experiment, subjets were seurely supported in the harness even when no BWS was provided (under full weight bearing onditions) while they walked without parallel bars. All subjets reported that the harness provided a feeling of safety whih eliminated any fear of falling. Regardless of this, a deterioration in the gait pattern haraterized by an inrease and broadening of the EMG ativity was noted in subjets with symmetrial gait involvement. These results suggest that parallel bars provide lateral stability whih ompensates for the dereased postural reations observed in spasti subjetslll while walking on a moving surfae. The hanges in EMG patterns espeially noted in the distal musles, losely resemble those observed in an immature gait when postural reations are not yet fully developed. 11-13 Providing 4% BWS for suh patients resulted in more normal EMG profiles and it an be suggested that BWS ompensated for the derease in postural stability. This is supported by a previous study whih reports that providing 4% BWS, while walking on a treadmill, results in a derease in perentage total double support time and an inrease in single limb sup ort time in spinal ord injured subjets. F An interesting observation was the more symmetrial gait pattern whih emerged in asymmetrially involved subjets when they walked without holding on to parallel bars at % BWS. Although the gait was laborious and nonfuntional (limited to three Paraplegia 32 (1994) 54-553 steps), removing the parallel bars produed a more normal swing phase by eliiting EMG ativity in T A aompanied by ankle dorsiflexion. The hip, knee, and ankle flexion present when walking without parallel bars had not emerged under onditions where the subjet was allowed to ompensate by using parallel bars. The more normal gait pattern only emerged with an inrease in loomotor demand and no possibility for ompensation. In order to failitate gait when asymmetrial subjets walked without parallel bars, 4% BWS was provided. The subjets were able to take a larger number of steps with greater ease, making gait training without parallel bars more feasible. Severely impaired subjets with symmetrial involvement, who were not able to walk without parallel bars at % BWS, were able to do so when 4% BWS was provided. Retraining gait with BWS, while inreasing the demand on the loomotor system by removing parallel bars, is worth onsidering in the early stages of gait training. The approah appears imperative in failitating a more symmetrial gait pattern while disouraging gait asymmetries from developing. Although both parallel bars and BWS an be onsidered as a form of support during gait training, there appears to be definite differenes between the two. Parallel bars appear to yield a more ompensatory gait pattern, probably by failitating weight transferene to the less involved side in asymmetrially involved subjets. BWS enourages a more symmetrial gait pattern by supporting a perentage of body weight entrally, and disouraging ompensation with the less involved extremity. The differenes between these two types of support for both symmetrial and asymmetrial patients need to be further eluidated. The effets of inreasing treadmill speed Variations in EMG ativity, angular displaement profiles, and temporal distane parameters as a funtion of walking speed in healthy subjets have been extensively reported in the literature14-17 This is in ontrast to a lak of objetive information whih exists quantifying the effets of in-
Paraplegia 32 (1994) 54-553 reased walking speed on gait parameters in spinal ord injured patients. Dietzl8 alluded to musle hypertonia as one of the main reasons for the spasti pareti subjets' inability to walk at faster speeds. In this study, one of the effets of inreasing walking speed was the appearane or inrease in lonus in four subjets. Burke & Lanel9 have reported that the amplitude of streth reflexes are diretly proportional to the veloity of streth. In three other subjets, inreasing walking speed resulted in only a small inrease in musle EMG amplitude. Spasti pareti subjets are known to walk at speeds onsiderably lower than that of normal subjets.s,6 In the present group of subjets, the omfortable treadmill speeds ranged from.8 ms-1 to.4 ms-i, and the highest maximal treadmill speed was.6 ms-i for subjet 1. At suh low walking speeds it beomes imperative to distinguish gait harateristis aused by the pathology and those whih are a result of slow walking speeds.2o.2l Until reently,22 pathologial gait profiles were ompared to standardized normative data derived from healthy subjets walking at omfortable speeds. 1. 2 3 Shiavi et al16 investigated the EMG profiles of healthy subjets walking at very low speeds (lowest range:.34 ms-i) in order to provide a template for omparison with pathologial gait while ontrolling the speed effet. They reported that MH musle was biphasi at the very low walking speeds, with most subjets displaying a burst of ativity at the stane-swing transition whih was abolished as the speed inreased. This type of MH EMG profile has been observed in this as well as other studies1,5 and may be a result of low walking speed and not neessarily of abnormal motor programming. Shiavi et al16 also reported that gait parameters beame more variable at the very low walking speeds and that musles responded to individual movement requirements. Subjets desribed their gait as beoming less automati. The low walking speeds may partially explain the inreased variability in gait parameters observed among spasti pareti subjets when ompared to normal subjets.24 Friso et al14 studied lower limb angular displaements A study of spasti pareti gait 551 patterns in normal subjets at different walking speeds and reported the frequent absene of a yield at the knee during initial stane at lower speeds. This is a finding frequently observed in spasti pareti gait,s and the subjets' low walking speeds may be partly responsible for this. In general, subjets in this study exhibited a low EMG amplitude for most musles while walking at their omfortable treadmill speeds. Knutsson24 desribed one of the abnormal ativation patterns as a marked derease in EMG ativity during gait in ertain musles, although the subjets were able to generate fore in these musles during other tasks. The dereases in EMG ativity noted among spasti pareti subjets may partly be resulting from their low walking speeds as well as entral paresis. Impliations for gait training The omplexity of the disturbed loomotor pattern, oupled with postural instability and abnormal reflex ativity, following a spinal ord lesion, make it diffiult to establish a universal training strategy to reeduate gait. A lear understanding of the gait defiits and their auses is required in individual ases in order to propose a omprehensive gait training approah.1o,24,2s Likewise, an understanding of the influene of external parameters suh as parallel bars, BWS and speed are needed in order to inorporate them into a training strategy. Although removal of parallel bars has been thought to produe a deterioration in the gait pattern, it has been demonstrated here that in patients with an asymmetrial gait pattern, removing parallel bars will derease the opportunity for ompensation thus allowing for the expression of a more normal gait pattern. If without parallel bars the loomotor task is a diffiult one, BWS an be inorporated in the training regimen to failitate gait. For those subjets who are unable to walk without parallel bars at % BWS, providing 4% BWS allows gait training without parallel bars and minimizes the development of a ompensatory gait pattern. In symmetrially involved patients parallel bars an be removed and postural stability inreased with BWS while retrain-
552 Visintin and Barbeau ing gait. Suh patients appear to be andidates for gait training with BWS, ahieving the goal of inreasing speed, inreasing postural stability, while eliiting more normal EMG and angular displaement profiles of the lower extremities. Inreasing walking speed did not always result in a deterioration in the gait pattern of spasti pareti subjets. Those presenting with minimal abnormal reflex ativity during gait showed just a small inrease in EMG amplitude although they were not able to walk at speeds muh beyond that of their omfortable level (subjet 5:.15 ms-i-.25 ms-i; subjet 1:. 4 ms-i-.6 ms-i). Suh subjets would probably benefit from gait training with BWS at speeds higher than their omfortable walking speed.5 In subjets where inreased speed led to an inrease in abnormal reflex ativity, drug therapy would be indiated Paraplegia 32 (1994) 54-553 prior to gait training.6,26.27 When retraining gait and other funtional tasks, it beomes important to manipulate the environment and provide external onditions whih favour the response being sought and enhane funtional reovery.28 Providing the appropriate environmental onditions is ritial during gait training in order to maximize the loomotor potential while preventing ompensatory gait deviations. It beomes lear from the above results that following a spinal ord injury a omprehensive and interative approah is required in order to maximize the potential for loomotor reovery. An understanding of the underlying abnormal motor programs and reflex ativity, oupled with an understanding of the influene of external parameters suh as parallel bars, BWS and speed during loomotion, are essential to ahieve this goal. Referenes 1 onrad B, Beneke R, Meink HM (1985) Gait disturbanes in paraspasti patients. In: Delwaide PJ, Young RR, editors. linial Neurophysiology in Spastiity. Vol 1. Restorative Neurology. Elsevier Siene Publishers B.V. (Biomedial Division), Netherlands: 155-174. 2 Knutsson E (198) Musle ativation patterns of gait in spasti hemiparesis, paraparesis and erebral palsy. In: Fugel-Meyer A, editor. Stroke with Hemiplegia. Sand J Rehabil (Suppl 7): 47-52. 3 onrad B, Beneke R, arne hi J et al (1983) Pathophysiologial aspets of human loomotion. In Desmedt JE, editor. Motor ontrol Mehanisms in Health and Disease. Raven Press, New York: 717-726. 4 Bobath B (1978) Adult Hemiplegia: Evaluation and Treatment. 2nd edn. Heinnemann Medial Books, London, UK. 5 Visintin M, Barbeau H (1989) The effets of body weight support on the loomotor pattern of spasti pareti patients. an J Neural Si 16: 315-325. 6 Barbeau H, Fung J, Stewart J, Visintin M (1988) Impairment of spasti parapareti gait: impliations for new rehabilitation strategies. Pro Fifth Biennial onf an So Biomeh: 12-16. 7 Barbeau H, Wainberg M, Finh L (1987) Desription and appliation of a system for rehabilitation. Med Bio Eng omput 25: 341-344. 8 Fung J, Barbeau H (1987) Quantifiation of the eletromyographi ativity in normal human gait. Pro IEEE onf Biomed Tehnol, Monteh: 41-44. 9 Hirshberg GG, Nathanson M (1952) Eletromyographi reording of musular ativity in normal and spasti gaits. Arh Phys Med Rehabil 33: 217-226. 1 Beneke R, onrad B (1986) Disturbanes of posture and gait in spasti syndromes. In: Bles W, Brandt TH, editors. Disorders of Posture and Gait. Elsevier Siene Publishers B.V. (Biomedial Division), Netherlands: 231-241. 11 Berger W (1986) Development of gait in hildren. In: Bles W, Brandt TH, editors. Disorders of Posture and Gait. Elsevier Siene Publishers B.V. (Biomedial Division): 315-324. 12 Forssberg H (1985) Ontogeny of human loomotor ontrol I. Infant stepping, supported loomotion and transition to independent loomotion. Exp Brain Res 57: 48-493. 13 Sutherland DH, Olshen R, ooper LB, Woo SLY (198) The development of mature gait. J Bone Joint Surg Am 62: 336-353. 14 Frigo, Eng D, Tesio L (1986) Speed-dependant variations of lower-limb joint angles during walking. Am J Phys Med 65: 51-62. 15 Kirtley, Whittle MW, Jefferson RJ (1985) Influene of walking speed on gait parameters. J Biomed Eng 7: 282-288. 16 Shiavi R, Bugle HJ, Limbird T (1987) Eletromyographi gait assessment, part 1: Adult EMG profiles and walking speed. J Rehabil Res Dev 24: 13-23.
Paraplegia 32 (1994) 54-553 A study of spasti pareti gait 553 17 Yang JF, Winter DA (1985) Surfae EMG profiles during different walking adenes in humans. EEG lin Neurophysiol 6: 485-491. 18 Dietz V (1986) Impaired reflex ontrol of posture and gait in spasti paresis. In: Bles W, Brandt TH, editors. Disorders of Posture and Gait. Elsevier Siene Publishers B.V. (Biomedial Division), Netherlands: 243-252. 19 Burke D, Lane JW (1973) Studies of the reflex effets of primary and seondary spindle endings in spastiity. In: Desmedt JE, editor. New Developments in Eletromyography and linial Neurophysiology. Vol 3. Karger, Basal: 475-495. 2 Andriahi TP, Ogle JA, Galante JO (1977) Walking speed as a basis for normal and abnormal gait measurements. J Biomeh 1: 261-264. 21 Longhurst S (198) Variability of EMG during slow walking. Pro Speial onf an So Biomehan: 1-11. 22 Shiavi R, Bugle HJ, Limbird T (1987) Eletromyographi gait assessment, part 2: Preliminary assessment of hemipareti synergy patterns. J Rehabil Res Dev 24: 24-3. 23 Knutsson E, Rihards (1979) Different types of disturbed motor ontrol in gait of hemipareti patients. Brain 12: 45-439. 24 Knutsson E (1985) Studies of gait ontrol in patients with spasti paresis. In: Delwaide PJ, Young RR, editors. linial Neurophysiology in Spastiity. Vol 1. Restorative Neurology. Elsevier Siene Publishers B.V. (Biomedial Division), Netherlands: 175-183. 25 Grimm RJ (1983) Program disorders of movement. In: Desmedt JE, editor. Motor ontrol Mehanisms in Health and Disease. Raven Press, New York: 1-11. 26 Barbeau H, Rihards L, Bedard PJ (1982) Ation of yproheptadine in spasti parapareti patients. J Neural Neurosurg Psyhiatry 45: 923-926. 27 Wainberg M, Barbeau H, Gauthier S (1986) Quantitative assessment of the effets of yproheptadine on spasti pareti gait: A prelimianry report. J Neural 233: 311-314. 28 Bah-y-Rita P (1983) Rehabilitation versus passive reovery of motor ontrol following entral nervous system lesions. In: Desmedt JE, editor. Motor ontrol Mehanisms in Health and Disease. Raven Press, New York: 185-192.