Brain Research Reviews 40 (2002) 274–291

www.elsevier.com / locate / brainresrev

Review

The effect of locomotor training combined with functional electrical

stimulation in chronic spinal cord injured subjects: walking and reflex
studies
Hugues Barbeau a,b , *, Michel Ladouceur d , Mehdi M. Mirbagheri c , Robert E. Kearney c
a
School of Physical and Occupational Therapy, McGill University, 3630 Promenade Sir-William-Osler, Montreal, QC, Canada H3 G 1 Y5
b
Jewish Rehabilitation Hospital Research Center, Laval, QC, Canada H7 V 1 R2
c
Department of Biomedical Engineering, McGill University, Montreal, QC, Canada H3 A 2 A7
d
Faculty of Applied Health Science, Brock University, St. Catharines, ON, Canada, L2 S 3 A1

Abstract

With the new developments in traumatology medicine, the majority of spinal cord injuries sustained are clinically incomplete and the
proportion is likely to continue to rise. Thus, it is necessary to continue to develop new treatment and rehabilitation strategies and
understand the factors that can enhance recovery of walking following spinal cord injury (SCI). One new development is the use of
functional electrical stimulation (FES) device to assist locomotion. The objective of this review is to present findings from some recent
studies on the effect of long-term locomotor training with FES in subjects with SCI. Promising results are shown in all outcome measures
of walking, such as functional mobility, speed, spatio–temporal parameters, and the physiological cost of walking. Furthermore, the
change in the walking behavior could be associated with plasticity in the CNS organization, as seen by the modification of the stretch
reflex and changes in the corticospinal projection to muscles of the lower leg. In conclusion, recovery of walking is an increasing
possibility for a large number of people with SCI. New modalities of treatment have become available for this population but most still
need to be evaluated for their efficacy. This review has focused on FES assisted walking as a therapeutic modality in subjects with chronic
SCI, but it is envisaged that the care and recovery of SCI in the early phase of recovery could also be improved.
 2002 Elsevier Science B.V. All rights reserved.

Theme: Motor systems and sensorimotor integration

Topic: Spinal cord and brainstem

Keywords: Locomotor training; Electrical stimulation; Spinal cord injury

Contents

1. Introduction ............................................................................................................................................................................................ 275

2. FES assisted walking ............................................................................................................................................................................... 275
3. Changes in the locomotor behaviour with locomotor training using FES assisted walking ............................................................................. 276
4. Changes in functional mobility ................................................................................................................................................................. 277
5. Changes in maximal overground walking speed ......................................................................................................................................... 277
6. Changes in the kinematics of walking with locomotor training using FES assisted walking ........................................................................... 280
7. Changes in the physiological cost of walking............................................................................................................................................. 281
8. Mechanisms involved in the locomotor behaviour changes ......................................................................................................................... 281
8.1. Changes in reflex organization with FES assisted walking .................................................................................................................. 281
8.2. Effect of CPN stimulation on the soleus stretch reflex ........................................................................................................................ 281
8.3. Effect of cutaneomuscular stimulation during walking ....................................................................................................................... 281
8.4. Effect of locomotor training using FES on the soleus stretch reflex ..................................................................................................... 283

*Corresponding author. Tel.: 11-514-398-4519; fax: 11-514-398-8193.

E-mail address: hugues.barbeau@mcgill.ca (H. Barbeau).

0165-0173 / 02 / $ – see front matter  2002 Elsevier Science B.V. All rights reserved.
PII: S0165-0173( 02 )00210-2
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291 275

9. Changes in the corticospinal control of tibialis anterior with FES assisted walking ....................................................................................... 283
10. Discussion............................................................................................................................................................................................. 286
Acknowledgements ...................................................................................................................................................................................... 288
References................................................................................................................................................................................................... 288

1. Introduction adaptations of the circulatory system during exercise

[43,45,89,10,96,116,53].
Spinal cord injury (SCI), by nature, has a very sudden Spinal cord injury is also associated with changes within
impact on an individual, both in a physical sense as well as the central nervous system that include problems of muscle
in an emotional and social sense [83]. The causes of injury activation such as weakness [94,66,35,71,87,41,68,78,
and demographics of SCI have been reported to be 108,27], hyperactive spinal reflexes [70,11,85,47,64], and
constant for the last 20 years [23]. The incidence of SCI loss of sensory function [19,111]. It is believed that all of
varies around the world, but where comprehensive data are these modifications lead to postural problems related to
available, the incidence is usually reported to be between bearing weight, maintaining balance and developing prop-
20 and 50 cases per million per year, approximately half of ulsion.
whom are under 30 years of age [12,22,25,54,102,36, Biomechanical descriptions of walking by SCI subjects
98,99]. In the United States, the annual incidence of SCI is are scarce, but caution should be exercised when compar-
around 10 000 [106]. In Canada, the incidence is 35 SCI ing the results because of the heterogeneity of the samples
cases per million per year and 900 new cases are reported included in those studies. It has been reported that the
per year to the Canadian Paraplegic Association [13]. Of pattern of SCI subjects walking on a treadmill, in com-
all the new cases, a large proportion has some preservation parison to able-bodied participants at matched treadmill
or recovery of sensory and / or motor function caudal to the speed, can be characterized by a shorter cycle duration and
level of lesion. Such an injury is termed an incomplete stride length [15,7]. Furthermore, it was also reported that
lesion. In many areas of the world, incomplete injuries are the total knee excursion was reduced, initial contact was
now more common than clinically complete injuries made with a flexed knee such that the ankle was forced
[33,25,36,98]. As a result of new developments in into more dorsiflexion with little plantarflexion occurring
traumatology medicine, the proportion of incomplete spi- during the stance phase [16,7]. However, in a case study of
nal cord injured (SCI) subjects are likely to continue to an SCI subject walking overground it was shown that cycle
rise. There are several other potential new treatments for duration was increased and that foot contact was made
acute SCI [29,90] that may continue to reduce the severity with increased knee flexion and a plantarflexed foot [55].
of neurological impairment and consequently increase the The changes in the electromyographic patterns have been
potential for walking ability. relatively more thoroughly investigated [21,15,16,55,7,31]
In view of the above findings and trends in the epi- than kinematic and kinetic changes. Briefly, there is
demiology of SCI, it is becoming increasingly important to increased tibialis anterior muscle activation during swing
develop treatment strategies that can enhance the recovery and the triceps surae amplitude was slightly reduced [21]
of motor function, walking in particular, following SCI. In with a broadened and flattened electromyographic profile
conjunction with providing treatment to enhance recovery [15,16,55] and activation early in the stance phase
of walking, rehabilitation specialists must consider the [21,15,16,31,32]. Muscles acting on the knee also show
factors that may influence such recovery. prolonged muscle activation [16,31] that seems to be
related to the activity of muscles acting on the ankle joint
[67].
2. FES assisted walking It can be seen that SCI subjects do have problems in
most of the critical events of the gait cycle. Weakness of
Spinal cord injury, and the inactivity that follows, the knee extensors would certainly modify the gait cycle
triggers numerous modifications of the musculoskeletal, during the loading and midstance phases [15,16]. Weakness
circulatory and central nervous systems that will influence of the ankle plantarflexors would correspond to changes in
the walking behaviour. Examples of the changes in the the terminal stance and pre-swing phases [3]. The initial
musculoskeletal system are the increased stiffness of the and midswing phases would be affected by changes in the
passive components of the ankle joint [74], the increased passive and reflex stiffness of the ankle plantarflexors as
fatigability and modification of the biochemical properties well as weakness of the ankle dorsiflexors [21,16]. Finally,
of motor units [17,118,103,109], and the higher incidence the terminal swing phase would be affected by a decrease
of osteoporosis [6,34,113,56,20]. Furthermore, depending in angular velocity generated by the weakness of the hip
on the level of the injury, there is a modification of the flexors since intersegmental dynamics are dependent on the
276 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
velocity of the movement [92,44,114]. Another factor is
the energy requirement necessary for ambulation. Al-
though wheelchair propulsion approximates the energy
requirement of normal walking [8], the energy cost of
walking by SCI subjects is higher than that for speed-
matched walking by able-bodied participants [105].
Many modalities of intervention have been attempted to
improve walking in SCI subjects. Those modalities range
from pharmacological interventions (for a review, see Ref.
[82]) or surgical modalities [5] to the use of modalities
modifying the external constraints like ambulatory assis-
tive devices [72] or mechanical orthoses (for a review, see
Refs. [79,93]). Another modality is the use of active
orthosis using functional electrical stimulation (FES) as-
sisted walking. There is a long history for the therapeutical
use of electrical stimulation. Greek and Roman physicians
used electrical activity from torpedo fishes before the time
of Jesus Christ [65]. Furthermore, the feasibility of con-
tracting paralysed muscles with electrical stimulation was
already shown in the eighteenth century [30]. More than 30
years ago, functional electrical stimulation (FES) was
developed as an orthotic system to prevent ‘foot drop’ in
hemiplegic subjects or to be used for SCI patients [69,48].
Since then FES has been used to restore a variety of
movements including walking (for a review, see Refs.
[4,119,88]) and was first reported for SCI subjects in 1989
[57]. The goal of FES is to obtain an immediate contrac-
tion of the skeletal muscles that will lead to a functional
movement. As reviewed in the Introduction, the gait of
SCI subjects is impaired to some extent by weaknesses in
the knee and hip extensors as well as ankle dorsiflexors. In
relation to the wide spectrum of motor impairment follow-
ing SCI, there are multitudes of systems for FES assisted
walking. The simplest system uses a stimulation of the
common peroneal nerve through surface electrodes in
conjunction with a footswitch [69]. However, systems
using either four or six channels [4] could also be used. It
is interesting to note that new peroneal nerve stimulator
systems, using implanted stimulators and natural sensing,
are still being developed even 30 years after the initial use
of a common peroneal nerve surface stimulator [42,14,40].
Even though FES assisted walking has been available
for more then three decades, it has not been widely used in
rehabilitation for a number of technical reasons. The
stimulators were bulky, unreliable, prone to breakage and
expensive [104]. Recent studies of simple system for
FES-assisted walking (common peroneal nerve stimulator)
showed an increase of the walking speed (less then 0.1
m*s 21 ) and a reduction of oxygen consumption during gait
[37,104,112].
The combination of simple FES system with locomotor
training shows the usefulness of such devices to improve
walking. Remarkably, walking speed of SCI subjects using
FES assisted walking is sometimes increased even when
the stimulator is temporarily turned off. This retained
increase of walking speed without FES, called the thera-
peutic effect, has been reported anecdotally for hemiplegic
patients [69,107] and in SCI subjects [38]. Changes in their
assistive ambulatory devices as well as changes in the
energy efficiency of the walking behaviour has also been
shown. The relative contributions of locomotor training
and FES-assisted walking to the improvement of walking
speed still need to be determined. The objective of this
paper is to review the results on the effect of locomotor
training using FES assisted walking seen during a longi-
tudinal study in SCI subjects [58,59,75] as well as some of
the mechanisms that could be involved in this effect.
3. Changes in the locomotor behaviour with
locomotor training using FES assisted walking
Unless specified otherwise, the data presented in this
review paper were generated from fourteen incomplete SCI
subjects, with an average age of 33 years (range from 25 to
48.9). These participants as well as subsets of these
participants were samples of convenience. The time be-
tween the injury and the start of the FES assisted walking
program varied from 1.8 to 19.1 years. The neurological
level of the lesion of the participants ranged from the fifth
cervical vertebra (C5) to the first lumbar vertebra (L1).
According to the American Spinal Injury Association
(ASIA) [24], the participants were either in the C (5 / 14) or
D (9 / 14) category. Furthermore, the participants spanned
the range of functional categories [86] from community
walkers (n54), least community walkers (n53) to house-
hold walkers (n56). All of the participants required an
ambulatory assistive device at the onset of the study. The
ambulatory assistive devices used were either a walker
(n56), forearm crutch (n54) or cane (n54).
Prior to the start of the FES assisted walking all the
participants signed a consent form. Following a 4-week
initiation program on the proper use of the FES stimulator,
the participants were asked to use FES assisted walking as
much as possible in their activities of daily living.
FES assisted walking was accomplished by stimulating
the peroneal nerve to help initiate and accomplish the
swing phase and / or quadriceps stimulation provided to
help maintain the extension of the knee during the stance
phase [57]. When up to four channels of stimulation was
required, the participants were given a Quadstim stimulator
(Biomotion, Inc.), which was capable of providing mono-
phasic stimulation with a constant current. The parameters
of stimulation were adjustable for the output current (0–
150 mA), stimulus frequency (10–30 Hz) and stimulus
pulse width (50–500 ms). When the participants required
only one channel of stimulation, they were fitted with
either a Unistim (Biomech Engineering Ltd.) or MikroFES
(Institut Jozer Stefan) device. The Unistim stimulator
provided monophasic stimulation with a constant voltage
output. The output voltage was adjustable from 0 to 100 V,
whereas the stimulus frequency and pulse width were
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291 277

constant (23 Hz and 300 ms, respectively). The MikroFES
stimulator also provided monophasic stimulation, with a
constant voltage output that could be adjusted from 10 to
130 V. The frequency of stimulation was fixed at 25 Hz
with a pulse width of 150 ms. Triggering of the common
peroneal nerve stimulation with these devices could be
made either with hand or foot switches that used force-
sensing resistors (Interlink, Inc.) or mechanical contacts.
4. Changes in functional mobility
A functional scale consisted of 13 items scored on a
seven-point scale (for a maximum score of 84 points) was
used to evaluate functional mobility. The psychometric
properties of the scale have been tested and some modi-
fications were done [97]. The mobility score ranged from
45 to 83 points.
FES assisted walking has a functional impact as seen by
the statistically significant increase in functional mobility
within the first year (t-test: 6.702, df: 13, P,0.000). This
increase in functional mobility is dependent on many
factors but is mostly due to changes in the ambulatory
assistive device used during walking and increased walk-
ing speed. Furthermore, some participants (3 / 14) needed a
simpler system of stimulation within the first year of FES
assisted walking.
5. Changes in maximal overground walking speed
To evaluate the maximal overground walking speed
(MOWS), the participants were asked to walk as fast as
possible on a 15-m walkway. The surface of the walkway
was an industrial carpet with a high coefficient of friction.
Participants were instructed to start walking before the
starting line in order to accelerate and reach a steady speed
before the start of the timing of the performance. The
performance in the middle 10 m of the walkway was timed
with a handheld stopwatch and was then transformed into
average MOWS. This outcome measure had good reliabili-
ty for a sample of persons suffering from rheumatism [28].
The participants were asked to perform in different ran-
domized conditions: with the use of the FES orthosis,
without orthosis, and with the different ambulatory assis-
tive devices they could use. The time-course of the

Fig. 1. Longitudinal changes in the combined maximal overground walking speed. (A) Results of subjects with unlimited walking ability in the
community: AC (filled squares, full line), DB (Filled circles, small dotted line), JB (open squares, long dotted line), MR (open circles, full line), MS (open
squares, long dotted line). (C) Results of subjects with household walking ability: DT (filled squares, full line), LS (filled circles, small dotted lines), RL
(open squares, long dotted line). (D) Results of subjects with household walking ability: RP (filled squares, full line), SM (filled circles, small dotted line).
(Adapted from Ref. [58]).
278 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291

changes in MOWS was fitted to either a linear or exponen-
tial association function [58]. When possible, measures of
the MOWS were repeated to establish a baseline measure
of the spontaneous changes that were occurring. No
changes in MOWS were found in participants that were
evaluated at least three times prior to the start of the FES
assisted walking study. These results shows that no
spontaneous recovery was occurring prior to the start of
the FES assisted walking.
The increases in MOWS followed a longitudinal pro-
gression when locomotor training was combined to FES.
Fig. 1 shows the changes in MOWS for the participants.
Changes in MOWS occurred for the three functional
categories of walkers (A, Community walkers; B, Least
community walkers; C and D, Household walkers) with no
striking difference between categories in terms of the
model that best described the longitudinal changes. Within
the first year of FES assisted walking, the MOWS in-
creased by 0.26 m*s 21 (S.D.: 0.24) which was signifi-
cantly different from zero (t54.106, df513, P50.001).
This increase in MOWS ranged from 0.00 to 0.79 m*s 21
as shown in Fig. 1. Furthermore, the increase in MOWS
was correlated to the initial MOWS (r50.57, P50.033).
This increase was normally distributed as tested by the
Kolmogorov–Smirnov distance (KS) [21] (KS: 0.2323,
P.0.10). The increase in MOWS was a percentage of the
initial MOWS (71%, S.D.6 65). In addition, the mag-
nitude of change correlated to the initial walking speed,
showing a greater increase for the participants that were
community walkers at the onset of the study (Fig. 1A). Of
the 14 participants, two had no changes in MOWS in the
combined condition. Of the remaining 12 participants, the
longitudinal changes of four were best described with a
linear model, whereas those of the other eight participants
were best described with an exponential association model.
When the progression followed an exponential associa-
tion model, the average K value was 0.271 (0.238). This
represents an improvement in the walking speed for a
duration of at least 23 weeks.
Fig. 2 shows the MOWS as a function of the functional
category of walker both while using the FES orthosis (FES
training columns) and when the stimulator was turned off
(Off FES columns). It can be seen that the combination of
a FES system with locomotor training was effective for
improving walking in chronic SCI subjects (Fig. 2, first
column). The average gain in speed was 0.26 m / s (see
above) after 1 year of locomotor training with FES. This
gain represented more than double the increase in MOWS

Fig. 2. Changes in maximal overground walking speed for FES combined with locomotor training and when the stimulation was turned off (N514). The
overall effect pre- and post-intervention as well as the effects in subgroups of different initial gait speed are shown (modified from Ref. [58]).
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291 279

that is related to using an FES orthosis (not illustrated, see
Ref. [112], P,0.003). Furthermore, the gain was signifi-
cant across all levels of initial gait speed (Fig. 3, columns
3, 5 and 6). Strikingly, a general gain in walking speed
was seen even when the stimulator was turned off (Fig. 2,
columns 2, 4, 6 and 8).
The combination of a simple FES system with
locomotor training shows the usefulness of such devices to
improve walking [112]. Fig. 3 illustrates the increase of
MOWS in two chronic SCI subjects using FES assisted
walking over a long period of time. A gradual improve-
ment of the maximal walking speed was observed from

Fig. 3. Example (SM from Fig. 1) of the changes in the kinematics with FES assisted walking. Panels (A–C) show the stick figure (every 0.5 s) and
trajectories of retroreflective markers (see Methods) when SM walks without FES at the beginning of FES assisted walking (A), without FES but after 43
weeks of FES assisted walking (B) and with the FES assisted walking at the same evaluation session (C). Panels (D–F) shows the angular excursion of the
hip (D), knee (E) and ankle (F) of the affected lower limb in these trials. Calibration bars for panels (A–C) represent 0.5 m. The hip, knee and ankle
angular excursions (D–F) without FES assisted walking at onset of training, without FES assisted walking after 43 weeks of training are represented by
thin, dotted and thick lines, respectively. (From Ref. [59]).
280 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291

0.20 to 0.70 m / s for the first SCI subject (Fig. 3A) and
from 0 to 0.35 m / s in the second SCI subject (Fig. 3B).
Remarkably, in both SCI subjects the walking speed was
increased even when the stimulator was temporarily turned
off, with changes from 0.20 to 0.58 m / s for the first
subject (Fig. 3A) and changes from 0 to 0.20 m / s for the
second subject (Fig. 3B). This retained increase of walking
speed without FES has been termed the therapeutic effect
and is of the same magnitude as reported earlier [58,59].
This therapeutic effect on walking has been reported
anecdotally for hemiplegic subjects [69] and in SCI
subjects [38,60]. Changes in their assistive ambulatory
devices as seen in Fig. 3A with the subject using either
crutches or canes were also observed. From these results, it
was concluded that increases in maximal walking speed
during training with the FES orthosis is mostly due to a
therapeutic effect, which implies that some plasticity
occurred during the training paradigm. The locomotor
recovery observed with this training paradigm appears to
result from the interaction between these plastic changes in
combination with the proper activation of different
peripheral afferents including cutaneous and proprioceptive
inputs (see Section 8 on the mechanisms involved in the
locomotor behaviour changes).

6. Changes in the kinematics of walking with

locomotor training using FES assisted walking
The kinematic patterns in the sagittal plane and temporal
variables were reconstructed from video recordings of the
most affected side. Kinematic outcome measures included
hip, knee, and ankle angular excursions as well as stride
length and duration, and periods of stance and swing [59].
Fig. 4 shows examples of the changes in the spatio–
temporal parameters found with a program of FES-assisted
walking (Fig. 4A–C; one stride in each condition). A
comparison of Fig. 4A and Fig. 4B shows the increased
stride length (0.61 and 1.22 m) and reduce stride time (4.4
and 3.9 s) that occurred with long-term use (43 weeks) of
FES-assisted walking even though the stimulator was
turned off both times. The example also shows that FES-
assisted walking has a minor effect on stride length (1.22–
1.28 m) and stride duration (3.9–3.2 s for this example),
although the average in both conditions is 3.6 s) when the
comparison is made for the same day (Fig. 4, B vs. C).
Fig. 4 also illustrates an example of the changes in the
time-normalized angular joint excursion occurring with the
use of FES-assisted walking as well as changes occurring
with the training program (Fig. 4D–F). One stride is
shown for each condition. The use of FES-assisted walking
increased hip angular excursion by 16.58 (Fig. 4D).
However, the knee was more flexed in the early stance
phase and more extended in the later stance phase (from 0
to 30% and from 50 to 80% of the normalized cycle time,
respectively, Fig. 4E). Although the ankle angular excur-
sion during the stride did not change among the three
Fig. 4. Example of two SCI subjects showing the longitudinal changes of
the maximal overground walking speed occurring with locomotor training
using FES-assisted walking. (A) This panel is an example (SM from Fig.
1) of a qualitative change between the combined (open squares) and
therapeutic (filled squares) conditions as well as an example of the similar
longitudinal changes occurring when the person uses different ambulatory
assistive devices. Squares represent the condition with forearm crutches;
circles represent the condition with two canes. (B) Example (CD from
Fig. 1) of the longitudinal changes in the physiological cost of FES
assisted walking, who had a decrease in PCI combined with an increased
walking speed. Significant differences in the means are represented with
an asterisk. The non-assisted and FES assisted walking conditions are
represented by open and filled symbols, respectively, whereas walking
speed is represented by circles and PCI by squares.
conditions, there was an increased ankle dorsiflexion angle
during the swing phase of gait when FES-assisted walking
was used (Fig. 4F). In addition, an increased ankle
dorsiflexion during the swing phase by 10.98 and decreased
ankle plantar / flexion at foot contact by 5.68 allowed a
better foot placement when using FES assisted walking
(Fig. 4B, C). In contrast, the knee and ankle angular
excursions did not change.
As a group, walking training with FES assisted walking
during the first year increased the comfortable walking
speed by 0.10 m*s 21 (F511.90, dF59, P50.007). This
improvement in walking speed results from an average
increase of 0.12 m in stride length (F54.43, dF59, P5
0.06) and 0.04 Hz in stride frequency (F513.26, dF59,
P50.005). The increase in stride frequency is due to a
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
decreased stance duration of 0.22 s (F59.43, dF59, P5
0.013) with only minor changes in the swing duration
(F52.82, dF59, P50.127). Quite surprisingly, those
changes in the spatio–temporal parameters of walking
were not concomitant with changes in the lower limb
angular excursions. However, when the difference between
using FES assisted walking and no orthosis was analysed,
it was shown that FES assisted walking did not influence
the spatio–temporal parameters of walking but increased
the hip angular excursion and ankle dorsiflexion angle
during the swing phase and the ankle angle at foot contact.
7. Changes in the physiological cost of walking
The physiological cost of walking was measured with
the physiological cost index. The index was established by
dividing the difference between the heart rate during
walking and during sitting with the walking speed mea-
sured in metres per minute. The result of this calculation is
the physiological cost of walking in heartbeats per metre
walked [59].
To evaluate the effect of training with FES assisted
walking on the physiological cost of walking, we studied
nine participants longitudinally from onset of FES assisted
walking. When combined with time, we see five different
types of responses: the physiological cost index remained
constant and the walking speed increased (n53), the
physiological cost index decreased and the walking speed
remained constant (n53), the physiological cost index
decreased and the walking speed increased (n51; Fig. 3B),
the physiological cost index and walking speed increased
(n51), and the physiological cost index increased and the
walking speed remained constant (n51). These results
show a positive effect for eight out of the nine participants
that were evaluated for duration of more then 3 months.
8. Mechanisms involved in the locomotor behaviour
changes
8.1. Changes in reflex organization with FES assisted
walking
Participants with injuries to the central nervous system,
like multiple sclerosis, stroke or incomplete spinal cord
injury, show a significant reduction of the short latency
stretch reflex (and / or H-reflex) modulation during gait
[117,32,100]. In stroke participants, this impaired modula-
tion results from the absence of an increased stretch reflex
threshold in the swing phase in comparison to the stance
phase of gait [81] as seen in able-bodied participants. This
lowered threshold for the stretch reflex in spastic patients
could lead to an increased soleus activity [31]. Lack of foot
281
clearance in the transition from stance to swing is often
observed during walking by spastic participants severely
impairing their walking ability [81,101]. It is often as-
sumed to be related to the weakness of the dorsiflexors, but
could equally well be due to a reflex mediated tone in the
plantar-flexors as they are stretched during the swing phase
of the step cycle. One of the hypotheses to be tested was
whether the stimulation used to make the tibialis anterior
contract would lead to a modification of the soleus stretch
reflex.
8.2. Effect of CPN stimulation on the soleus stretch
reflex
Inhibition of the soleus motoneuron pool with the
stimulation of the common peroneal nerve (CPN), through
reciprocal inhibition, has been shown in humans [18].
However, recent studies [76,80] have shown different
sensitivities to presynaptic inhibition for the H- and the
stretch reflex. In spite of this difference, studies in able-
bodied [2], stroke [109], and multiple sclerosis [63]
participants have shown the feasibility of inhibiting the
soleus stretch reflex by stimulating the CPN. The mecha-
nism of action for this inhibition is still being investigated
but recent data suggests that it could be partially related to
an inhibitory pathway from group II afferents [63].
8.3. Effect of cutaneomuscular stimulation during
walking
Fig. 5 shows the soleus H-reflex modulation patterns in
healthy (Fig. 5A) and SCI (Fig. 5C) subjects. Details of
the stimulation paradigm could be obtained from a previ-
ous report [32]. The test soleus H-reflex was obtained by
stimulating the tibial nerve in the popliteal fossa via a
monopolar surface electrode. The anode was placed imme-
diately above the patella. The test stimulus consisted of a
single 1-ms square wave pulse delivered by the stimulator
(Grass model S88) and isolation units (Grass SIU5 and
CCU1). A recruitment curve was determined initially in
the relaxed standing position, and the intensity that yielded
a minimally detectable M-wave was chosen for subsequent
stimulation. Throughout all standing walking trials, the
effective stimulus strength was controlled by adjusting the
intensity on-line to obtain the same and consistent M-
wave.
The conditioning stimulus, consisting of an 11-ms train
of three 1-ms pulses at 200 Hz (Fig. 5B), was delivered to
the medial plantar region, stimulating the cutaneous and
mechanoreceptors of the sole. The intensity was varied
from sensory threshold (1T), which was barely perceptible,
to a maximally, comfortably tolerable level (maxT), rang-
ing from 2.5 to 3T in all subjects tested. It is important that
the maxT intensity was perceived as innocuous by all
282 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291

Fig. 5. Modulation of the soleus H-reflex (d, mean6S.D.) in a healthy subject and (A) a severely impaired SCI (C) patient during treadmill walking. Note
the absence of H-reflex modulation throughout the entire gait cycle in the SCI patient and the increase in amplitude beyond that observed during quiet
standing (indicated by the dotted horizontal line). With cutaneomuscular stimulation (B), the H-reflex became partially and selectively inhibited in early
stance and swing, thereby revealing a more normal modulation pattern (D), with the M-wave (h, mean6S.D.) remained relatively low and constant. Soleus
activity was flattened and prolonged over the gait cycle (E), with the presence of clonus, whereas TA showed relatively reciprocal activity, except for the
early and midstance period. (Arrow indicates the average time of transition from stance to swing). (From Ref. [32]).

subjects because they have to undergo many walking and
standing trials with the paired (conditioning test) and
unpaired (test) stimulation.
As illustrated in Fig. 5C, severely impaired SCI subjects
showed little or no phasic modulation of the soleus H-
reflex during walking as compared to the normal subject
(Fig. 5A). These subjects generally had severe clinical
signs of spasticity and marked gait deficits. As compared
to the mean control amplitude in quiet standing (see dotted
horizontal line near top of Fig. 5), the H-reflex during
walking was generally, though not significantly, elevated
beyond the mean level during standing. With conditioning
cutaneomuscular stimulation, the H-reflex was reduced
significantly in the early stance and swing phases, thereby
producing a pattern of H-reflex that seemed to be modu-
lated physically through the step cycle (Fig. 5D). This
selective inhibition as a result of cutaneomuscular stimula-
tion was seen in three out of the four severely impaired
subjects [32]. The soleus activity in these subjects general-
ly were flattened and prolonged in the step cycle, with the
presence of clonic activity (Fig. 5E). The tibialis anterior
activity, when present, was either a small (see Fig. 5F) and
narrow burst in midswing concurrent with foot drag or a
pattern of coactivation with soleus, especially through
early and mid stance (not illustrated).
All SCI subjects reported an ease for the limb to swing
through and a decreased effort to walk secondary to the
reduction in limb stiffness, when the conditioning stimula-
tion is appropriately timed to occur just before or during
swing. This was especially the case for the moderately or
severely impaired subjects who had difficulty dorsiflexing
the foot, often resulting in toe drag during swing.
The H- and the stretch reflex have been shown to have
different sensitivities to presynaptic inhibition [76]. The
effects of CPN stimulation on the soleus stretch reflex
during walking has been investigated in stroke participants
[110]. Stretch reflexes where elicited during walking on a
treadmill by an electromechanical device capable of
stretching the human left ankle. A detailed description on
the design of the device is reported elsewhere [1]. Briefly,
the device consisted of a functional joint attached to the
participant’s ankle joint and an actuator system, involving
a powerful AC-motor / gear system. The functional joint
was mounted in alignment with the ankle joint and was
connected to the actuator system, which was placed next to
the treadmill, by means of two flexible bowden wires.
Position feedback was used to regulate the actuator and
enable it to follow the movement of the ankle without
influencing the pattern of gait. Whenever it was desired
during the gait cycle, it was possible to evoke a stretch of
the ankle extensor muscles by rotating the ankle joint with
a displacement of up to 108. Besides the conditioned and
test stretches two types of control recordings where
collected: strides with the conditioning stimulus only and
unaffected strides. The stretches were randomly mixed
with the control recordings and applied infrequently. A
recording (with stretch or control) was taken every two to
four strides. To avoid fatigue, rest periods were added
according to the participant’s needs. The conditioning
stimulus was applied to the deep peroneal nerve through a
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
monopolar surface electrode (32 mm in diameter) which
was placed distally and somewhat laterally in relation to
the patella, close to where the deep peroneal nerve
branches off the common peroneal nerve. To make sure
that the common peroneal nerve was not stimulated the
tendons of the peroneus longus and brevis muscles were
palpated while applying test stimuli. An anode was placed
halfway on the shin (436 cm oval). The conditioning
stimulus consisted of a train of five 1-ms pulses applied at
a frequency of 200 Hz for a total duration of 21 ms. The
stimulus intensity was set at four times the motor threshold
of the tibialis anterior muscle (MT), except for participant
2. For this participant a stimulus intensity of 3 MT was
used because 4 MT was uncomfortable. A stimulus
intensity of 4 MT was chosen to elicit the highest
inhibition possible as seen in a previous study effectuated
in sitting participants [109].
The result of this study [110] shows that the soleus
stretch reflex can be inhibited by a conditioning stimulus to
the CPN. Panels (A–F) of Fig. 6 shows an example of the
recordings made when the stretch reflex was (Fig. 6D–F)
or was not (Fig. 6A–C) preceded by a conditioning
stimulus. It can be seen, in that example, that the stretch
applied in the early part of the swing phase triggered both
a large burst of EMG activity in the soleus muscle as well
as some clonic activity (Fig. 6B). However, no stretch
reflex EMG activity could be detected when the stretches
applied in the early part of the swing phase were preceded
by the conditioning stimulus (Fig. 6E). Fig. 6G shows the
mean results of the inhibition of the stretch reflex as a
function of the conditioning-test delay of the stroke
participants. It can be seen that the greatest inhibition was
achieved with a delay of about 100 ms and that the stretch
reflex could be inhibited by 75%. Hence, it was demon-
strated that the conditioning stimulus was able to restore
artificially the modulation of the stretch reflex during the
walking cycle.
8.4. Effect of locomotor training using FES on the
soleus stretch reflex
Another question that needs to be answered is whether
long-term use of FES assisted walking has an effect on the
segmental reflex organization. The contributions of the
intrinsic and reflex components to dynamic ankle stiffness
were assessed at different ankle angles by a parallel-
cascade system identification method before and after at
least 18 months of FES assisted walking. Briefly, the
participant laid supine with their foot attached to a pedal of
a stiff, position controlled, electro-hydraulic actuator.
Pseudorandom binary sequences of ankle position were
applied about different ankle angles in the range of motion
of the participants. The amplitude of the perturbations
were adjusted to produce the maximum reflex torque while
the participants maintained a contraction equivalent to 10%
of the maximal voluntary contraction. Transducers mea-
sured joint position and torque and the electromyographic
283
activity of the tibialis anterior and lateral gastrocnemius
were also recorded. A more complete description of the
parallel-cascade model can be found elsewhere [49].
The effect of training with FES assisted walking on the
intrinsic and reflex dynamic stiffness in spinal cord injured
participant was also evaluated. Fig. 7 shows segments of
experimental trials from a typical SCI subject before (left
column) and after (right column) long-term use of FES-
assisted walking. As illustrated, the EMG (panels E and F)
responses were much smaller post-FES than pre-FES. In
both cases the Pseudoramdom Binary Sequence Trials
(PRBS) inputs had the same peak-to-peak amplitude
(0.035 rad) and zero-crossing rate (125 ms). Gastroc-
nemius (GS) EMG was unidirectionally rate sensitive; with
a short-latency burst of EMG that followed each move-
ment in the dorsiflexing (positive) direction whereas
equivalent movements in the opposite direction evoked no
response. However, the EMG amplitude post-FES (panel
D) was much smaller than pre-FES (panel C). The torque
responses were also similar in nature, comprising of two
distinct components: (i) a symmetric variation that corre-
lated with the ankle position and its derivatives with no
significant delay, likely representing the contribution of
intrinsic mechanics, and (ii) a transient, twitch-like in-
crease in torque that was associated with dorsiflexing
displacements, due to reflex mechanisms. These two
components are seen more clearly in the bottom row where
segments of the torque record are shown on an expanded
scale with the reflex components shaded. As with EMG,
the reflex torque magnitude was much smaller post-FES
(panels F and H) than pre-FES (panels E and G), by a
factor of about 2 in this particular SCI subject. It can be
seen that the intrinsic stiffness, which decreases in SCI as
compared to healthy subjects, was increased after the
period of training with FES assisted walking. Whereas the
passive, active reflex and passive reflex components, which
increase in SCI subjects, were decreased during the same
period. These effects were consistent for all positions.
To characterize the amplitude of these changes, we
computed the percentage change of stiffness gain before
and after 18 months of FES assisted walking over the
range of motion (Fig. 8) for each subject. On average, the
group of four incomplete SCI subjects increased their
intrinsic stiffness gain by 97% (S.E.: 44%), decreased their
passive stiffness gain by 39% (S.E.: 8%), and decreased
their reflex stiffness gain by 89% (S.E.: 5%). The pattern
was similar in all chronic SCI subjects when trained with
the FES [75].
9. Changes in the corticospinal control of tibialis
anterior with FES assisted walking
Even though changes in the segmental reflex organiza-
tion could be seen with the improvement of the walking
behaviour, changes in muscle strength could also be
284 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291

Fig. 6. Example of the inhibition of the soleus stretch reflex in the early swing phase of gait in hemiplegic participants from a conditioning stimulus
applied to the deep peroneal nerve. Average responses (n55) to an unconditioned (A–C) and a conditioned stretch (D–F) from one participant. Thick lines
show the recordings with a stretch and the thin dotted lines the control recordings. The control recordings of panels (D–F) show the conditioning stimulus
with no ankle stretches. The stretch reflex inhibition, as calculated from the ratio of the amplitude of the conditioned over the test stretch reflex, is
presented for four different conditioning-test intervals (panel G). The dotted horizontal line represents no changes between the conditioned and test stretch
reflex. In panel (G) the error bars represent the standard deviation. Significant changes are marked with an asterisk. (Modified from Ref. [110]).

responsible for this improvement. Changes in muscle
strength have been shown previously with FES assisted
walking and were thought to be due to peripheral mecha-
nisms such as hypertrophy of the muscle fibers and
changes in muscle fibres type. However, recent studies
have shown that increased afferent input could be corre-
lated with an increased corticospinal control of the
homonymus muscle [39,91]. A recent study [52], investi-
gated the effect of the CPN stimulation on the size of
transcranial magnetic motor envoked potentials (MEP) in
the tibialis anterior muscle. It was shown that after 30 min
of a regimen of repetitive electrical stimulation that
mimicked the stimulation parameters seen during FES
assisted walking, there was a significant increase in the
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291 285

Fig. 7. Segments experimented trials of pre- and post-FES. (A, B) position (p); (C, D) reflex EMG (EMG R ) (E, F) torque (TQ); (G, H) segments of the
torque record between dashed lines are shown on the expanded scale with the reflex components shaded (From Ref. [110]).
286 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291

Fig. 8. Example of changes in ankle dynamics with FES and locomotor training. (A) Passive stiffness gain; (B) intrinsic stiffness gain and (C) reflex
stiffness gain as functions of ankle position for one SCI subject. Pre-FES, Dotted lines, Post-FES, Solid-lines. (Modified from Ref. [75]).

MEP (Fig. 9A). The results of eight able-bodied particip-
ants showed that these changes in the TA MEP were seen
for more than 20 min following the cessation of the CPN
stimulation (Fig. 9A) and were not related to an increase in
motoneuron excitability as evaluated by the monosynaptic
stretch reflex (Fig. 9B). However, the long latency com-
ponent of the stretch reflex, which is thought to be acting
through a transcortical pathway, was increased following
the repetitive electrical stimulation of the CPN nerve. This
result shows that plastic changes within the CNS could
also explain the increase in muscle strength.
10. Discussion
Functional recovery of walking is a goal for many SCI
subjects and for the intervention of many health profes-
sionals involved in SCI rehabilitation. Evidence presented
in this review on the effect of FES assisted walking give a
rational for the use of such an orthosis in the rehabilitation
of walking in the SCI population. Remarkably, the abso-
lute therapeutic effect of using FES assisted walking was
greatest for faster SCI subject walkers even though the
orthotic effect was, at best, minimal.
The aims of these studies were to characterize the
magnitude and time-course of the changes occurring with
the use of the FES assisted walking for SCI subjects. The
main results are that SCI subjects using a FES orthosis has
an improved mobility score due to an increased maximal
overground walking speed as well as a change in the type
of ambulatory assistive device used. This increase in
walking is a result of an increased cycle length and cycle
frequency, with the latter being related to a decreased
stance time. These changes are also reflected in the
increased efficiency of walking as measured by the de-
crease in the physiological cost index.
Previous studies of simple systems of FES assisted
walking for SCI subjects [103,37,38] have shown a
minimal effect on walking speed. Our study showed an
average threefold increase of the walking speed, which can
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
Fig. 9. Changes in the MEP of the TA with repetitive electrical
stimulation of the CPN. Panel A presents a representative example of the
changes in MEP following repetitive electrical stimulation of the CPN
(filled circles, TA; open circles, soleus). Panel B shows the group results
for MEP of the TA (filled bars), short (open bars) and long (hatched bars)
latency stretch reflex of the TA at four different times. The four different
times are before repetitive electrical stimulation, less than 10 min
following the end of repetitive electrical stimulation, more than 10 min
following the end of repetitive electrical stimulation and more than 60
min following the end of repetitive electrical stimulation. Statistical
significance is shown with stars (**, P,0.01 and *, P,0.05). (Modified
from Ref. [52]).
be explained by two factors: the choice of outcome
variable and the length of the study. As reviewed above,
the MOWS of our participants increased on average by
0.26 m*s 21 whereas the improvement in the walking speed
seen during the kinematic evaluation was about 0.10
m*s 21 . In previous studies, the time since the start of FES
assisted walking was either not taken into account [103] or
was controlled with only a relatively short duration
[37,38]. However, the models fitted to the experimental
data from the present study reveal that time since the start
of FES assisted walking is an important a factor in the
explanation of the increase in MOWS.
The result, from Fig. 2, shows that most of the increase
in maximal overground walking speed stems from the
therapeutic effect of FES assisted walking, and provides
evidence of plasticity in walking behaviour in the chronic
287
stage of SCI. The therapeutic effect of FES assisted
walking may arise from many factors such as plasticity of
both the peripheral and central nervous system. Activated
muscles, by electrical or voluntary means, incur changes in
the properties of their fibre [95]. No study, so far, has
reported the effect of FES assisted walking using common
peroneal nerve stimulation on changes to the muscle fibre
properties of either tibialis anterior or triceps surae.
However, studies of electrical stimulation of the peroneal
nerve have shown changes in the muscle properties that
were stimulus dependent [50,51]. Our participants reported
that lower leg muscles hypertrophied while using FES
assisted walking (most probably from an increased cross-
sectional area related to the increased muscular activity).
Plasticity occurs in many sites within the nervous
system. Changes in synaptic activity have been shown
from the connections in the monosynaptic stretch reflex
[115] to cortical sensorimotor maps [26]. It is relevant and
important to this study in that electrical stimulation of the
common peroneal nerve has been shown to reduce the
amplitude of the H-reflex in the soleus muscle [105].
Furthermore, some participants reported a decreased inci-
dence of spasms as well as better bowel and bladder
control. These changes in the activity-dependent feedback
from different receptors, coupled with changes in the
muscles could trigger a reorganization of the planning and
generation of the walking behaviour.
These modifications can be seen behaviourally by an
increased strength of the lower limbs as well as changes in
the output of the motor patterns. An increase in ankle
dorsiflexion strength has been reported for hemiparetics
that received common peroneal nerve stimulation while
walking or sitting [73]. Three types of responses to a
programme of electrical stimulation of the quadriceps was
shown in SCI subjects: some participants increased both
their voluntary and electrically activated strength; others
only their electrically activated strength and some showed
no changes in either condition [57]. These three types of
changes in the strength of the lower limb are similar to the
changes reported above for the MOWS in our participants.
Change in walking behaviour with the use of FES assisted
walking has been reported in the population of persons
who have experienced a cerebral vascular accident. A case
study of a person who had experienced a cerebral vascular
accident and was using FES assisted walking showed,
basically, an improvement in the control of weight accept-
ance on the paretic side, changes in the kinematic pattern
of walking with a return to normal knee flexion at toe-off,
peak knee flexion and knee extension at heel strike [9].
An increase in walking speed that can be fitted by an
exponential association function has also been seen in SCI
subjects with a complete motor function loss who used
FES assisted walking [46], as well as by participants in a
computer-assisted gait training program for hemiparetic
subjects [77]. Because the changes were similar in all
conditions, it is hypothesized that the changes occurring
288 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
with the use of FES assisted walking represent systemic
changes that could also be used in other sensorimotor
activities.
The results of this study have many functional implica-
tions that can be relevant to the whole spectrum of SCI
subjects because of the heterogeneity of our sample,
although greater improvements in the maximal overground
walking speed were seen in participants that were walking
faster at the onset of the study. The reported time-courses
of the changes in MOWS allow us to postulate that clinical
trials on the effect of an FES orthosis for SCI subjects
should use a longer period than 12 weeks of training as
well as evaluate the therapeutic effect of the orthosis.
Because of the discrepancies between the results of this
study and previous studies [104,37,38] of the effect of FES
assisted walking in SCI subjects, it is proposed that the
evaluation of the changes in the walking behaviour occur-
ring with an intervention should take into account at least
two factors such as the control and capacity of the walking
behaviour [60]. These factors can be evaluated using
walking tasks that demand a modification of the walking
pattern from the usual walking pattern at a comfortable
speed on a flat, overground surface to maximal speed.
Furthermore, this study shows a potential for the use of
FES orthosis in SCI rehabilitation by providing a training
tool that will help in restoring the walking behaviour.
Future directions of the research should be an inves-
tigation of the factors associated with the therapeutic
effect. On the behavioural aspect, factors associated with
the limitation in the maximal walking speed should be
investigated. In a case report, it was shown that training
with FES assisted walking was associated with an in-
creased maximal frequency of stepping [84]. Changes in
the nervous and musculoskeletal system caused by long-
term FES assisted walking should also be studied. Pre-
liminary evidence shows a modification of the ankle
plantarflexor stiffness dynamics associated with long-term
use of FES assisted walking [75]. In that study [75], it was
shown that in some SCI subjects using FES assisted
walking, there was a decrease in the reflex and passive
stiffness gains as well as an increase in the intrinsic
stiffness gain. This long-term modification of the stretch
reflex gain could be the product of a repetitive con-
ditioning from the action of a peroneal nerve group II
inhibitory pathway projecting to the ankle plantarflexor
motorneuron pool [61,62]. Another question that needs to
be answered is the apparent discrepancy between the
modifications in the spatio–temporal parameters and
changes in the kinematics of the lower limb. This question
arises from the fact that changes in the kinematics of
walking are present earlier than any changes in the walking
speed. This evidence is the basis for the hypothesis that
changes in the angular pattern create a permissive situation
that needs to be incorporated into the walking pattern with
training. Another argument for this hypothesis can be
found in the trajectory of the foot reported in Fig. 4C. This
figure represents the trajectories of the different retro-
reflective markers during one typical walking cycle while
using FES assisted walking. When the participant uses FES
assisted walking, the foot is brought back before it contacts
the ground. This could be explained by a lack of control of
the intersegmental dynamics contributing to the late part of
swing phase [120]. A second explanation is that the
dynamic projection of the centre of gravity may not follow
the improvement in the forward advancement of the foot
resulting in a backward force that would cause disequilib-
rium. Further investigations in the control of the foot
trajectory and projection of the centre of mass during
swing could be made to resolve this question.
In conclusion, locomotor recovery has become possible
for an increasing number of incomplete people with SCI.
New modalities of treatment are available for this popula-
tion but most still need to be evaluated for their efficacy.
This presentation reviewed evidence that FES assisted
walking could be used as a therapeutic modality even in a
chronic stage of SCI. It is then possible to envision that, by
using more efficient technology and readily available
therapeutic modalities, the care and recovery of SCI
subjects could be improved.
Acknowledgements
We would like to extend many thanks to Irene
Shanefield and Vira Rose for their very important contribu-
tions for the bibliography and secretarial support, to
Gevorg Chilingaryan for his assistance in statistical analy-
sis and to Dr. Joyce Fung for critical editorial work that is
greatly appreciated. We would also like to thank the Fonds
de la recherche en Sante´ du Quebec (FRSQ) and the JRH
foundation for their continuing support.
References
[1] J.B. Andersen, T. Sinkjaer, An actuator system for investigating
electrophysiological and biomechanical features around the human
ankle joint during gait, IEEE Trans. Rehab. Eng. 3 (1995) 299–306.
[2] J.A. Apkarian, S. Naumann, Stretch reflex inhibition using electrical
stimulation in normal subjects and subjects with spasticity, J.
Biomed. Eng. 13 (1991) 67–73.
[3] T. Bajd, A. Kralj, T. Karnik, R. Savrin, H. Benko, P. Obreza,
Influence of electrically stimulated ankle plantar flexors on the
swinging leg, Artif. Organs 21 (1997) 176–179.
[4] Y. Bajd, A. Kralj, H. Turk, H. Benko, J. Sega, Use of functional
electrical stimulation in the rehabilitation of patients with incom-
plete spinal cord injuries, J. Biomed. Eng. 11 (1989) 96–102.
[5] G. Barolat, Surgical management of spasticity and spasms in spinal
cord injury: an overview, J. Am. Paraplegia Soc. 11 (1988) 9–13.
[6] G.M. Bedbrook, G.J. Sedgley, The management of spinal injuries
past and present, Int. Rehabil. Med. 2 (1980) 45–61.
[7] R. Benecke, B. Conrad, Disturbance of posture and gait in spastic
syndromes, in: W. Bles, T. Brandt, (Eds.), Disorders of Posture and
Gait, Elsevier Science, Amsterdam, NL, 1986, pp. 231–241.
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
[8] R. Blessey, Energy cost of normal walking, Orthop. Clin. North Am.
9 (1978) 356–358.
[9] U. Bogataj, N. Gros, M. Kljajic, R. Acimovic-Janezic, Enhanced
rehabilitation of gait after stroke: a case report of a therapeutic
approach using multichannel functional electrical stimulation, IEEE
Trans. Rehabil. Eng. 5 (1997) 221–232.
[10] D.C. Bunten, A.L. Warner, S.R. Brunnemann, J.L. Segal, Heart rate
variability is altered following spinal cord injury, Clin. Auton. Res.
8 (1998) 329–334.
[11] B. Calancie, J.G. Broton, K.J. Klose, M.J. Difini, D.R. Ayyar,
Evidence that alterations in presynaptic inhibition contribute to
segmental hypo- and hyperexcitability after spinal cord injury in
man, Electroencephalogr. Clin. Neurophysiol. 89 (1993) 177–186.
[12] A. Campos da Paz, P.S. Beraldo, M.C. Almeida, E.G. Neves, C.M.
Alves, P. Khan, Traumatic injury to the spinal cord. Prevalence in
Brazilian hospitals, Paraplegia 30 (1992) 636–640.
[13] Canadian Paraplegic Association. Spinal Cord Injury. Available:
http: / / www.canparaplegic.org / spinal.html, 1st August, 1999.
[14] C. Childs, M.K. Haugland, E.M. Nielsen, Mapping of energy field
for a two channel implantable nerve stimulator. Proceedings of the
4th International Congress of the International Neuromodulation
Society and Third Annual Congress of the International Functional
Electrical Stimulation Society, Sept. 16–20th, 1998, Lucerne, Swit-
zerland, pp. 261.
[15] B. Conrad, R. Benecke, J. Carnehl, J. Hohne, ¨ H.M. Meinck,
Pathophysiological aspects of human locomotion, Adv. Neurol. 39
(1983) 717–726.
[16] B. Conrad, R. Benecke, H.M. Meink, Gait disturbances in paraspas-
tic patients, in: P.J. Delwaide, R.R. Young (Eds.), Clinical Neuro-
physiology in Spasticity, Elsevier, Amsterdam, 1985, pp. 155–174.
[17] T.C. Cope, S.C. Bodine, V.R. M Fournier, Edgerton, Soleus motor
units in chronic spinal transected cats: physiological and mor-
phological alterations, J. Neurophysiol. 55 (1986) 1202–1220.
[18] C. Crone, H. Hultborn, B. Jespersen, J. Nielsen, Reciprocal Ia
inhibition between ankle flexors and extensors in man, J. Physiol.
389 (1987) 163–185.
[19] K.S. Crozier, V. Graziani, J.F. Ditunno Jr., G.L. Herbison, Spinal
cord injury: prognosis for ambulation based on sensory examination
in patients who are initially motor complete, Arch. Phys. Med.
Rehabil. 72 (1991) 119–121.
[20] G. Demirel, H. Yilmaz, N. Paker, S. Onel, Osteoporosis after spinal
cord injury, Spinal Cord 36 (1998) 822–825.
[21] V. Dietz, J. Quintern, W. Berger, Electrophysiological studies of gait
in spasticity and rigidity. Evidence that altered mechanical prop-
erties of muscle contribute to hypertonia, Brain 101 (1981) 431–
449.
[22] F. Dincer, A. Oflazer, M. Beyazova, R. Celiker, O. Basgoze, K.
Altioklar, Traumatic spinal cord injuries in Turkey, Paraplegia 30
(1992) 641–646.
[23] J.F. Jr Ditunno, C.S. Formal, Chronic spinal cord injury, N. Engl. J.
Med. 330 (1994) 550–556.
[24] J.F. Ditunno Jr., W. Young, W.H. Donovan, G. Creasey, The
international standards booklet for neurological and functional
classification of spinal cord injury, Am. Spin. Injury Ass. Paraplegia
32 (1994) 70–80.
[25] G.S. Dixon, J.N. Danesh, T.H. Caradoc-Davie, Epidemiology of
spinal cord injury in New Zealand, Neuroepidemiology 12 (1993)
88–95.
[26] J.P. Donoghue, Plasticity of adult sensorimotor representations,
Curr. Opin. Neurobiol. 5 (1995) 749–754.
[27] M. Drolet, L. Noreau, J. Vachon, H. Moffet, Muscle strength
changes as measured by dynamometry following functional rehabili-
tation in individuals with spinal cord injury, Arch. Phys. Med.
Rehabil. 80 (1999) 791–800.
[28] M.G. Eileen, E.M. Gerecz, Y.B. Kassam, H.M. Buchanan, W.W.
Buchanan, P.S. Tugwell, 50-foot walking time: a critical assessment
of an outcome measure in clinical therapeutic trials of antirheumatic
drugs, Br. J. Rheumatol. 127 (1988) 372–374.
289
[29] A.I. Faden, Pharmacological treatment of central nervous system
trauma, Pharmacol. Toxicol. 78 (1996) 12–17.
[30] B. Franklin, On the effects of electricity in paralytic cases. Phil.
Tran. 50 (1757) 481–483.
[31] J. Fung, H. Barbeau, A dynamic EMG profile index to quantify
muscular activation disorder in spastic paretic gait, Electroenceph.
Clin. Neurophysiol. 73 (1989) 491–495.
[32] J. Fung, H. Barbeau, Effects of conditioning stimulation on the
soleus H-reflex in normal and spastic paretic subjects during
walking and standing, J. Neurophysiol. 72 (1994) 2090–2104.
[33] J. Garcia-Reneses, R. Herruzo-Cabrera, M. Martinez-Moreno, Epi-
demiological study of spinal cord injury in Spain 1984–1985,
Paraplegia 28 (1991) 180–190.
[34] D.E. Garland, C.A. Stewart, R.H. Adkins, S.S. Hu, C. Rosen, F.J.
Liotta, D.A. Weinstein, Osteoporosis after spinal cord injury, J.
Orthop. Res. 10 (1992) 371–378.
[35] K.A. Gerhart, E. Bergstrom, S.W. Charlifue, R.R. Menter, G.G.
Whiteneck, Long-term spinal cord injury: functional changes over
time, Arch. Phys. Med. Rehabil. 74 (1993) 1030–1034.
[36] B.K. Go, M.J. DeVivo, J.S. Richards, The epidemiology of spinal
cord injury, in: S.L. Stover, J.A. DeLisa, G.G. Whiteneck (Eds.),
Spinal Cord Injury: Clinical Outcomes From the Model Systems,
Aspen, Gaithersburg, MD, 1995, pp. 21–55.
[37] M. Granat, J.F. Keating, A.C. Smith, M. Delargy, B.J. Andrews, The
use of functional electrical stimulation to assist gait in patients with
incomplete spinal cord injury, Disabil. Rehabil. 14 (1992) 93–97.
[38] M.H. Granat, A.C.B. Ferguson, B.J. Andrews, M. Delargy, The role
of functional electrical stimulation in the rehabilitation of patients
with incomplete spinal cord injury: observed benefits during gait
studies, Paraplegia 31 (1993) 207–215.
[39] S. Hamdy, J.C. Rothwell, Q. Aziz, K.D. Singh, D.G. Thompson,
Long-term reorganization of human motor cortex driven by short-
term sensory stimulation, Nature Neurosci. 1 (1998) 64–68.
[40] M. Hansen, M.K. Haugland, Adaptive fuzzy logic restriction rules
for error correction and safe stimulation patterns during functional
electrical stimulation, J. Med. Eng. Technol. 25 (2001) 156–162.
[41] J.F. Haughton, L.W. Little, R.K. Powers, L.R. Robinson, B. Golds-
tein, M / RMS: an EMG method for quantifying upper motoneuron
and functional weakness, Muscle Nerve 17 (1994) 936–942.
[42] J.A. Hoffer, R.B. Stein, M.K. Haugland, T. Sinkjaer, W.K. Durfee,
A.B. Schwartz, G.E. Loeb, C. Kantor, Neural signals for command
control and feedback in functional neuromuscular stimulation: a
review, J. Rehabil. Res. Dev. 33 (1996) 145–157.
[43] S.P. Hooker, J.D. Greenwood, D.T. Hatae D, R.P. Husson, T.L.
Matthiesen, A.R. Waters, Oxygen uptake and heart rate relationship
in persons with spinal cord injury, Med. Sci. Sports Exerc. 25
(1993) 1115–1119.
[44] M.G. Hoy, R.F. Zernicke, The role of intersegmental dynamics
during rapid limb oscillations, J. Biomechanics 19 (1986) 867–877.
[45] M. Irizawa, M. Yamasaki, S. Muraki, T. Komura, K. Seki, K.
Kikuchi, Relationship between heart rate and oxygen uptake during
submaximal arm cranking in paraplegics and quadriplegics, Ann.
Physiol. Anthropol. 13 (1994) 275–280.
[46] E. Isakov, J. Mizrahi, T. Najenson, Biomechanical and physiological
evaluation of FES-activated paraplegic patients, J. Rehabil. Res.
Dev. 23 (1993) 9–19.
[47] M. Jamshidi, A.W. Smith, Clinical measurement of spasticity using
the pendulum test: comparison of electrogoniometric and videotape
analyses, Arch. Phys. Med. Rehabil. 77 (1996) 1129–1132.
[48] A. Kantrowitz, Electronic Physiologic Aids, Maimonides Hospital,
Brooklyn, NY, 1963.
[49] R.E. Kearney, R.B. Stein, L. Parameswaran, Identification of
intrinsic and reflex contributions to human ankle stiffness dynamics,
IEEE Trans. Biomed. Eng. 44 (1997) 493–504.
[50] D. Kernell, O. Eerbeek, B.A. Verhey, Y. Donselaar, Effects of
physiological amounts of high- and low-rate chronic stimulation on
fast-twitch muscle of the cat hindlimb. I. Speed- and force-related
properties, J. Neurophys. 58 (1987) 598–613.
290 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
[51] D. Kernell, Y. Donselaar, O. Eerbeek, Effects of physiological
amounts of high- and low-rate chronic stimulation on fast-twitch
muscle of the cat hindlimb. II. Endurance-related properties, J. [72]
Neurophys. 58 (1987) 614–627.
[52] S. Khaslavskaia, M. Ladouceur, T. Sinkjaer, Increase in tibialis
anterior motor cortex excitability following repetitive electrical [73]
stimulation of the common peroneal nerve, Exper. Brain Res.
(2001), accepted.
[53] M. Kjaer, F. Pott, T. Mohr, P. Linkis, P. Tornøe, N.H. Secher, Heart
rate during exercise with leg vascular occlusion in spinal cord- [74]
injured humans, J. Appl. Physiol. 86 (1999) 806–811.
[54] S. Knutsdottir, Spinal cord injuries in Iceland 1973–1989. A follow
up study, Paraplegia 31 (1993) 68–72. [75]
[55] E. Knuttson, Studies of gait control in patients with spastic paresis,
in: P.J. Delwaide, R.R. Young (Eds.), Clinical Neurophysiology in
Spasticity, Elsevier Science, Amsterdam, NL, 1985, pp. 175–183.
[76]
[56] P. Kocina, Body composition of spinal cord injured adults, Sports
Med. 23 (1997) 48–60.
[57] A.R. Kralj, T. Bajd, Functional Electrical Stimulation: Standing and
[77]
Walking After Spinal Cord Injury, CRC, Boca Raton, FL, 1989.
[58] M. Ladouceur, H. Barbeau, Functional electrical stimulation-assisted
walking for persons with incomplete spinal injuries: longitudinal [78]
changes in maximal overground walking speed, Scand. J. Rehab.
Med. 32 (2000) 28–36.
[59] M. Ladouceur, H. Barbeau, Functional electrical stimulation-assisted [79]
walking for persons with incomplete spinal injuries: changes in the
kinematics and physiological cost of overground walking, Scand. J. [80]
Rehab. Med. 32 (2000) 72–79.
´
[60] M. Ladouceur, A. Pepin, K.E. Norman, H. Barbeau, Recovery of
walking after spinal cord injury, Adv. Neurol. 72 (1997) 249–255.
[61] M. Ladouceur, T. Sinkjaer, Stimulation of the common peroneal
nerve as a method for the inhibition of the stretch reflex in spastic [81]
ankle extensors muscles of participants with multiple sclerosis,
Aggiornamenti in Medicina Riabilitativa 1 (1999) 121.
[62] M. Ladouceur, P.H. Veltink, T. Sinkjær, The modulation of the
soleus stretch reflex by electrical conditioning stimuli to the [82]
peroneal and plantaris nerves. On-line proceedings of the 5th
Internet World Congress on Biomedical Sciences 98 at McMaster
University, Canada. (1998). Available: http: / / www.mcmaster.ca /
inabis98 / neuroscience / ladouceur0813 / index.html [83]
[63] M. Ladouceur, P.H. Veltink, T. Sinkjær, Modulation of the soleus
stretch reflex by electrical conditioning to the peroneal and plantaris
nerves, Abstract Soc. Neurosci. Abstr. 25 (1999) 121. [84]
[64] A. Lamontagne, F. Malouin, C.L. Richards, F. Dumas, Evaluation of
reflex- and nonreflex-induced muscle resistance to stretch in adults
with spinal cord injury using hand-held and isokinetic [85]
dynamometry, Phys. Ther. 78 (1998) 964–975.
´
[65] J.C. Le Flohic, J.P. Boucher, Notions de Stimulation Electrique
´
Therapeutique: Neuro-musculaire & Antalgique, UQAM, Montreal, ´ [86]
PQ, 1990.
[66] A.J. Lenman, F.M. Tulley, G. Vrbova, M.R. Dimitrijevic, J.A.
Towle, Muscle fatigue in some neurological disorders, Muscle [87]
Nerve 12 (1989) 938–942.
[67] A. Leroux, J. Fung, H. Barbeau, Adaptation of the walking pattern
to uphill walking in normal and spinal-cord injured subjects, Exp. [88]
Brain Res. 126 (1999) 359–368.
[68] A.D. Levi, C.H. Tator, R.P. Bunge, Clinical syndromes associated [89]
with disproportionate weakness of the upper versus the lower
extremities after cervical spinal cord injury, Neurosurgery 38 (1996)
79–83.
[69] W.T. Liberson, H.J. Holmquest, D. Scott, A. Dow, Functional [90]
electrotherapy: stimulation of the peroneal nerve synchronized with
the swing phase of the gait in hemiplegic patients, Arch. Phys. Med.
Rehabil. 42 (1961) 101–105. [91]
[70] J.W. Little, E.M. Halar, H-reflex changes following spinal cord
injury, Arch. Phys. Med. Rehabil. 66 (1985) 9–22.
[71] J.W. Little, R.K. Powers, P. Michelson, D. Moore, L.R. Robinson, B.
Goldstein, Electrodiagnosis of upper limb weakness in acute quad-
riplegia, Am. J. Phys. Med. Rehabil. 73 (1994) 15–22.
E.H. Melis, R. Torres-Moreno, H. Barbeau, E.D. Lemaire E,
Analysis of assisted-gait characteristics in persons with incomplete
spinal cord injury, Spinal Cord 37 (1999) 430–439.
R. Merletti, F. Zelaschi, D. Latella, M. Galli, S. Angeli, M. Bellucci
Sessa, A control study of muscle force recovery in hemiparetic
patients during treatment with functional electrical stimulation,
Scand. J. Rehab. Med. 10 (1978) 147–154.
M. Mirbagheri, R.E. Kearney, M. Ladouceur, H. Barbeau, Abnormal
stretch reflex mechanics in spastic spinal cord injured subject, Exp.
Brain Res. 141 (2001) 446–459.
M. Mirbagheri, M. Ladouceur, H. Barbeau, R.E. Kearney, The effect
of long-term FES-assisted walking on intrinsic and reflex dynamic
stiffness in spinal cord injured subjects. IEEE transaction in neural
system and rehabilitation engineering (2001) submitted.
H. Morita, N. Petersen, L.O. Christensen, T. Sinkjaer, J. Nielsen,
Sensitivity of H-reflexes and stretch reflexes to presynaptic inhibi-
tion in humans, J. Neurophysiol. 80 (1998) 610–620.
R. Nakamura, K. Suzuki, Y. Yamada, T. Handa, Computer-assisted
gait training (CGAT) of hemiparetic stroke patients: whose recovery
is most predictable?, Tohoku J. Exp. Med. 166 (1992) 345–353.
B.M. Needham-Shropshire, K.J. Klose, M.E. Tucker, C.K. Thomas,
Manual muscle test score and force comparisons after cervical spinal
cord injury, J. Spinal Cord Med. 20 (1997) 324–330.
A.V. Nene, H.J. Hermens, G. Zilvold, Paraplegic locomotion: a
review, Spinal Cord. 34 (1996) 507–524.
J.B. Nielsen, M. Enriquez-Dentron, T. Sinkjaer, H. Morita, L.O.D.
Christensen, N. Petersen, Presynaptic inhibition of excitatory post-
synapic potentials evoked by muscle stretch and electrical nerve
stimulation in cat lumbar motoneurones, Abstract Soc. Neurosci.
Abstr. 25 (1999) 123.
J.F. Nielsen, J.B. Andersen, H. Barbeau, T. Sinkjaer, Input–output
properties of the soleus stretch reflex in spastic stroke patients and
healthy subjects during walking, Neurorehabilitation 10 (1998) 151–
166.
K.E. Norman, The effects of clonidine, cyproheptadine and baclofen
on locomotor pattern in subjects with incomplete spinal cord injury.
Doctoral thesis, School of Physical and Occupational Therapy,
McGill University, 1996.
Paralysed Veterans of America. An Introduction To Spinal Cord
Injury. Available: http: / / www.pva.org / sci / 98scigd / 98scigd1.htm,
1st August 1999.
´
A. Pepin, M. Ladouceur, H. Barbeau, Factors limiting the maximal
walking speed in subjects with a spinal cord injury with an
incomplete motor function loss. J. Neurol. Rehabil., Submitted 2002.
K. Perell, A. Scremin, O. Scremin, C. Kunkel, Quantifying muscle
tone in spinal cord injury patients using isokinetic dynamometric
techniques, Paraplegia 34 (1996) 46–53.
J. Perry, M. Garret, J.K. Gronley, S.J. Mulroy, Classification of
walking handicap in the stroke population, Stroke 26 (1995) 982–
989.
C.M. Powers, C.J. Newsam, J.K. Gronley, C.A. Fontaine, J. Perry,
Isometric shoulder torque in subjects with spinal cord injury, Arch.
Phys. Med. Rehabil. 75 (1994) 761–765.
J. Quintern, Application of functional electrical stimulation in
paraplegic patients, Neurorehabilitation 10 (1998) 205–250.
J. Raymond, G.M. Davis, A. Fahey, M. Climstein, J.R. Sutton,
Oxygen uptake and heart rate responses during arm vs. combined
arm / electrically stimulated leg exercise in people with paraplegia,
Spinal Cord 35 (1997) 680–685.
D.H. Rhoney, M.S. Luer, M. Hughes, J. Hatton, New pharmacologic
approaches to acute spinal cord injury, Pharmacotherapy 16 (1996)
382–392.
M.C. Ridding M, B. Brouwer, T.S. Miles, J.B. Pitcher, P.D.
Thompson, Changes in muscle responses to stimulation of the motor
cortex induced by peripheral nerve stimulation in human subjects,
Exp. Brain Res. 131 (2000) 135–143.
 H. Barbeau et al. / Brain Research Reviews 40 (2002) 274–291
[92] G.E. Robertson, D.A. Winter, Mechanical energy generation, absorp-
tion and transfer amongst segments during walking, J. Biomech. 13
(1980) 845–854.
[93] G. Rubin, E. Cohen, Prostheses and orthoses for the foot and ankle,
Clin. Podiatr. Med. Surg. 5 (1988) 695–719.
[94] W.Z. Rymer, R.K. Powers, Muscular weakness in incomplete spinal
cord injury, Compr. Ther. 13 (1987) 3–7.
[95] S. Salmons, J. Henriksson, The adaptative response of skeletal
muscle to increased use, Musc. Nerv. 4 (1981) 94–105.
[96] A. Schmid, M. Huonker, J.M. Barturen, F. Stahl, A. Schmidt-
¨ D. Konig,
Trucksass, ¨ D. Grathwohl, M. Lehmann, J. Keul, Catechol-
amines, heart rate, and oxygen uptake during exercise in persons
with spinal cord injury, J. Appl. Physiol. 85 (1998) 635–641.
[97] L. Seaby, Mobility: a functional domain for clinical assessment and
program evaluation in physiotherapy. Thesis for the School of
Physical and Occupational Therapy, McGill University, Montreal, ´ Naaman, H. Barbeau, J. Bugaresti, E. Aimone, Multi-center
Canada, 1987.
[98] H. Shingu, M. Ohama, T. Ikata, S. Katoh, T. Akatsu, A nationwide
epidemiological survey of spinal cord injuries in Japan from January
1990 to December 1992, Paraplegia 33 (1995) 183–188.
[99] B. Silberstein, S. Rabinovich, Epidemiology of spinal cord injuries
in Novosibirsk, Russia, Paraplegia 33 (1995) 322–325.
[100] T. Sinkjaer, J.B. Andersen, J.F. Nielsen, Impaired stretch reflex and
joint torque modulation during spastic gait in multiple sclerosis
patients, J. Neurol. 243 (1996) 566–574.
[101] T. Sinkjaer, Muscle, reflex and central components in the control of
the ankle joint in healthy and spastic man, Acta. Neurol. Scand. 96
(1997) Suppl. 10.
[102] A. Soopramanien, Epidemiology of spinal injuries in Romania,
Paraplegia 32 (1994) 715–722.
[103] R.B. Stein R, T. Gordon, J. Jefferson, A. Sharfenberger, J.F. Yang,
´
J.T. de Zepetnek, M. Belanger, Optimal stimulation of paralysed
muscle after human spinal cord injury, J. Appl. Physiol. 72 (1992)
1393–1400.
´
[104] R.B. Stein, M. Belanger, G. Wheeler, M. Wieler, D. Popovic, A.
Prochazka, L.A. Davis, Electrical systems for improving locomo-
tion after incomplete spinal cord injury: an assessment, Arch. Phys.
Med. Rehabil. 74 (1993) 954–959.
´
[105] R.B. Stein, J.F. Yang, M. Belanger, K.G. Pearson, Modification of
reflexes in normal and abnormal movements, Prog. Brain. Res. 97
(1993) 189–196.
[106] S.L. Stover, P.R. Fine, Spinal cord injury: the facts and figures,
Birmingham, AL, University of Alabama at Birmingham, 1986.
[107] P.N. Taylor, J.H. Burridge, A.L. Dunkerley, D.E. Wood, J.A.
Norton, C. Singleton, I.D. Swain, Clinical use of the Odstock
dropped foot stimulator: its effect on the speed and effort of
walking, Arch. Phys. Med. Rehabil. 80 (1999) 1577–1583.
291
[108] C.K. Thomas, E.Y. Zaidner, B. Calancie, J.G. Broton, B.R.
Bigland-Ritchie, Muscle weakness, paralysis, and atrophy after
human cervical spinal cord injury, Exp. Neurol. 148 (1997) 414–
423.
[109] P.H. Veltink, M. Ladouceur, T. Sinkjær, Inhibition of the Triceps
Surae stretch reflex in spastic stroke patients by stimulation of the
deep peroneal nerve, Arch. Phys. Med. Rehabil. 81 (2000) 1016–
1024.
[110] M.M. Voormolen, M. Ladouceur, P.H. Veltink, T Sinkaer, Soleus
stretch reflex inhibition in the early swing phase of gait using Deep
Personal Nerve Simulation, Neuromodulation 3 (2000) 99–109.
[111] R.L. Waters, R.H. Adkins, J.S. Yakura, I. Sie, Motor and sensory
recovery following incomplete tetraplegia, Arch. Phys. Med.
Rehabil. 75 (1994) 306–311.
[112] M. Wieler, R.B. Stein, M. Ladouceur, M. Whittaker, A.W. Smith, S.
evaluation of electrical stimulation systems for walking, Arch.
Phys. Med. Rehabil. 80 (1999) 495–500.
[113] E. Wilmet, A.A. Ismail, A. Heilpom, D. Weiraeds, P. Bergmann,
Longitudinal study of the bone mineral content and soft tissue
composition after spinal cord section, Paraplegia 33 (1995) 674–
677.
[114] D. Wisleder, R.F. Zernicke, J.L. Smith, Speed-related changes in
hindlimb intersegmental dynamics during the swing phase of cat
locomotion, Exp. Brain. Res. 79 (1990) 651–660.
[115] J.R. Wolpaw, C.L. Lee, Memory traces in primate spinal cord
produced by operant conditioning of H-reflex, J. Neurophysiol. 60
(1989) 563–572.
[116] M. Yamamoto, F. Tajima, H. Okawa, T. Mizushima, Y. Umezu, H.
Ogata, Static exercise-induced increase in blood pressure in
individuals with cervical spinal cord injury, Arch. Phys. Med.
Rehabil. 80 (1999) 288–293.
[117] J.F. Yang, J. Fung, M. Edamura, R. Blunt, R.B. Stein, H. Barbeau,
H-reflex modulation during walking in spastic paretic subjects,
Can. J. Neurol. Sci. 18 (1991) 443–452.
[118] J.F. Yang, R.B. Stein, J. Jhamandas, T. Gordon, Motor unit
numbers and contractile properties after spinal cord injury, Ann.
Neurol. 28 (1990) 496–502.
[119] G.M. Yarkony, E.J. Roth, G. Cybulski, R.J. Jaeger, Neuromuscular
stimulation in spinal cord injury. I. Restoration of functional
movement of the extremities, Arch. Phys. Med. Rehabil. 73 (1992)
78–86.
[120] R.F. Zernick, K. Schneider, J.A. Buford, Intersegmental dynamics
during gait: implications for control, Adv. Psychol. 78 (1991)
187–204.