Research Articles

Neurorehabilitation and

Segmental Muscle Vibration Improves Neural Repair

24(3) 254­–262
© The Author(s) 2010
Walking in Chronic Stroke Patients With Reprints and permission: http://www.
sagepub.com/journalsPermissions.nav

Foot Drop: A Randomized Controlled Trial DOI: 10.1177/1545968309349940

http://nnr.sagepub.com

Marco Paoloni, MD,1,2 Massimiliano Mangone, MEng,2

Paola Scettri, MD,2 Rita Procaccianti, PhD,1
Antonella Cometa, MD,2,3 and Valter Santilli, MD1,2

Abstract
Background. Studies have described the effects of segmental muscle vibration (SMV) on brain plasticity and corticomotor
excitability. Information on the treatment-induced effects of SMV in stroke patients is, however, still limited. Objectives. To
assess whether the application of SMV to ankle dorsiflexor muscles of chronic stroke patients can improve walking. Methods.
Forty-four patients were randomly assigned to either an experimental group (EG) or a control group (CG) and underwent
12 sessions over 4 weeks of general physical therapy. Patients in the EG also received SMV at 120 Hz over the peroneus
longus and tibialis anterior for 30 minutes at the end of each session. All the participants underwent pretreatment and
posttreatment gait analysis assessments. Time–distance, kinematic, and surface electromyography (EMG) data were used as
outcome measures. Results. A moderate improvement in mean gait speed, normal-side swing velocity, bilateral stride length,
and normal-side toe-off percentage was observed only in the EG. A significant increase in bilateral ankle dorsiflexion angle
at heel contact was associated with increased maximum ankle dorsiflexion and plantarflexion degrees during the swing
phase on the paretic side after treatment in EG. Surface EMG during the swing phase revealed a significant increase in the
activation of the tibialis anterior muscle on the paretic side in the posttreatment assessment in the EG. Conclusions. SMV
added to general physical therapy may improve gait performance in patients with foot drop secondary to chronic stroke.
The authors hypothesize that this may be due to the mechanical vibration stimulation, probably as a consequence of effective
brain reorganization.

Keywords
segmental muscle vibration, gait, instrumental gait analysis, rehabilitation, chronic stroke

Introduction inputs activated by SMV can alter the excitability of the

Gait disorders affect a large population of stroke survi-
vors.1 Between 20% and 30% of patients who survive the
acute phase are unable to walk, whereas many others have
a moderate to severe walking disability.2 Gait alterations in
hemiplegic patients lead to an increased risk of falls,3,4 loss
in balance, and decreased social participation.5 Restoration
of gait is, therefore, one of the main goals of poststroke
rehabilitation,6 and it is now widely accepted that chronic
hemiplegic patients may also benefit from rehabilitation
and gait training.7-9
Segmental muscle vibration (SMV) is a technique that
applies a vibratory stimulus to a specific muscle using a
mechanical device. When applied to a muscle–tendon unit,
SMV induces the generation of Ia inputs as a consequence
of activation of muscle spindle primary endings.10 The Ia
corticospinal pathway11 by modulation of intracortical inhi­
biting and facilitating inputs to the primary motor cortex.12,13
Indeed, recent transcranial magnetic stimulation (TMS)
studies have shown increased excitability in the primary
motor cortex representation of the vibrated muscle fol­
lowing the application of low-amplitude SMV to the flexor
carpi radialis muscle12 and intrinsic hand muscles13 in heal­
thy subjects. Moreover, SMV applied to healthy subjects
1
Azienda Policlinico Umberto I, Rome, Italy
2
“Sapienza” University, Rome, Italy
3
Neurological Centre of Latium (NCL) Institute, Rome, Italy
Corresponding Author:
Marco Paoloni, MD, Department of Orthopaedic Science, Piazzale Aldo
Moro 3, 00185, Rome, Italy
E-mail: paolonim@tin.it
Paoloni et al 255
effectively compensates for repetitive TMS-induced inhibi-
tion of the prefrontal cortex.14
To our knowledge, the functional outcomes and the
potential therapeutic applications of SMV have not yet been
fully investigated. We hypothesize that SMV may, on the
basis of its presumed action on cortical excitability in addi-
tion to local spinal cord modulation, enhance paretic muscle
recruitment after stroke to diminish foot drop and improve
aspects of the gait cycle. We, therefore, designed a consid-
eration-of-concept randomized controlled trial15 to ascertain
whether the application of low-amplitude/high-frequency
SMV on the ankle dorsiflexor muscles of patients with foot
drop secondary to chronic focal ischemic brain injury imp­
roves walking ability, as measured by instrumental gait
analysis. We also assessed whether SMV is a well-tolerated
modality by investigating any local adverse event that
occurred during the treatment period.
Methods
Participants
Men and women between 35 and 80 years were included in the
study if they presented right or left hemiplegia associated with
foot drop of at least 6 months duration, secondary to magnetic
resonance imaging–documented chronic ischemic stroke.
Patients also had to be able to walk for at least 10 m without
assistive devices. The main exclusion criteria were bilateral
brain lesions, ischemic involvement of the cerebellum or basal
ganglia, aphasia, ankle spasticity (modified Ashworth scale =
3 or 4), psychiatric disease, cognitive impairment (Mini-­
Mental State Exam evaluation <23), and previous history of
neurological diseases, diabetes, or rheumatic and orthopedic
conditions that may interfere with locomotion. Patients were
also excluded if they were invol­ved in other clinical trials or if
they were under antispasticity therapy.
Foot drop was defined as the inability to reach a neutral
position of the ankle (ie, 90° angle between the shank and
the main axis of the foot) during the swing phase of gait.
Procedure
The study protocol was approved by the local ethics
committee. The experimental protocol was explained to
the participants and their informed consent was obtained.
Participants were randomly assigned to either an experi-
mental group (EG) or a control group (CG) by an
independent person who selected a sealed envelope 30 min-
utes before the intervention was due to start. All the
participants underwent an instrumental evaluation before
the training started (pretreatment) and 1 month after the end
of the training per­iod (posttreatment). All gait analysis data
were then analyzed offline by an assessor who was blinded to
the randomization.
All the participants underwent a 50-minute general
physical therapy session, 3 times per week, over a period
of 4 weeks. These sessions involved stretching, muscle
strengthening, balance, and overground walking training.
Participants in the EG alone received a further 30 minutes
of SMV therapy, which was delivered at the end of each
general physical therapy session, with the patient lying
supine.
At the end of each therapy session, patients in the EG
were orally interviewed and clinically evaluated to detect if
muscle pain or skin lesions (ie, irritation, reddening, abra-
sions) occurred.
Segmental Muscle Vibration Therapy
Low-amplitude SMV at a fixed frequency of 120 Hz was
delivered over the target muscles, that is, the tibialis ante-
rior and peroneus longus on the paretic side, by means of a
commercial device (Horus; Akropolis, Rome, Italy). The
Horus is an acoustic wave vibratory device composed of a
control unit and 3 multiple probe transducers with 4 head
transducers each for the transmission of the vibratory stim-
ulus. The acoustic production device, located in the control
unit, generates an acoustic wave that is induced within the
probe cavities and then translated into a vibratory stimulus
within the transducer heads that are fixed to the body sur-
face with elastic bands. The vibration amplitude was set at
10 mm, which is sufficient to drive Ia spindle afferents,16
but avoids muscle fiber injury17 and is subthreshold for the
tonic vibration reflex (TVR)18; the 30-minute stimulation
was delivered in trains of 6 seconds divided by 1-second
pauses.
Instrumental Evaluation
Gait analysis was performed using the ELITE stereopho­
togrammetric system (BTS, Milan, Italy) with 8 infrared
video cameras (TVC; BTS) for the acquisition of kine-
matic variables. Kinematic data were acquired and digitized
with a sampling rate of 100 Hz. Surface myoelectric sig-
nals were acquired with a sampling rate of 1000 Hz using
a 16-channel telemetric transmission surface electromyo-
graph (pocket EMG System; BTS). The lower and upper
cutoff frequencies of the Hamming filter were 10 and 400
Hz, respectively, whereas the common mode reaction ratio
was 100 dB. Anthropometric data were collected for each
subject, and retroreflective spherical markers were placed
over prominent bony landmarks.19 Disposable Ag/AgCl
bipolar surface electrodes, greased with electroconduc-
tive gel, were placed on the center of each muscle belly,
according to the European recommendations for surface
electromyography.20 The surface electrodes were placed
bilaterally on the gastrocnemius medialis (GM) and tibialis
anterior (TA) muscles. Offset angles of hip, knee, and ankle
256 Neurorehabilitation and Neural Repair 24(3)
joints on sagittal, frontal, and transversal planes were
obtained for each patient before a walking session started.
For this purpose, subjects were asked to maintain a standing
position for 4 to 5 seconds. The angles obtained were then
subtracted from those obtained during walking, with the
resulting angles then being used for the kinematic analysis.
Subjects were then instructed to walk at a self-selected
speed along a level surface approximately 10 m in length; 5
trials were acquired for each subject.
Time–Distance Data
For the gait analysis evaluation, a stride was considered as
the time between 2 consecutive heel–floor contacts of the
same limb and was subdivided in a stance phase (from the
first heel contact to toe-off) and a swing phase (from toe-off
to the second heel contact). We collected the following data
for both the paretic and normal side: stride length (m),
defined as the distance traveled by a person in a stride; step
width (m), defined as the mediolateral distance between the
feet; step length (m), defined as the distance between the
heel–floor contact of 1 limb and the subsequent heel–floor
contact of the contralateral limb; duration of the stance
phase, defined as the percentage of the whole gait cycle; and
swing velocity (m/s). Cadence (the number of steps over a
minute) and gait speed (m/s) were also calculated. The mean
value of 3 trials was considered for the time–distance data.
Kinematic Data
Three-dimensional marker trajectories during walking
were obtained by using a frame-by-frame tracking system
(Tracklab; BTS). The joint centers of rotation were deter-
mined, and joint angular excursion was calculated21; joint
excursion data were normalized to the stride duration and
reduced to 100 samples over the gait cycle. In the stance
phase, we considered the following parameters for both the
paretic and normal sides: the angle of the hip, knee, and
ankle joints at heel contact and ankle dorsiflexion and plan-
tarflexion range of motion (ROM). In the swing phase, we
considered the following parameters for both the paretic
and normal sides: ankle dorsiflexion and plantarflexion
ROM and the minimum and maximum angles of the ankle
on the paretic side. The mean value of 3 trials was consid-
ered for each kinematic variable.
Electromyographic Data
The surface electromyographic (EMG) signal of the TA and
GM muscles was collected from both the paretic and normal
sides during walking. We considered the EMG signal recorded
during the swing phase to assess muscular behavior related
to foot drop. The raw signal was rectified, low-pass filtered
with an upper cutoff frequency of 5 Hz, and integrated using
a mobile window of 125 ms to obtain the envelope. The
signal was then normalized at the maximum level of the
time-normalized, averaged EMG signal across all the trials.
The maximum EMG level from each muscle was identified
during the swing phase and then used to normalize the muscle
activity to this value.22 A 20% threshold of normalized signal
was used to evaluate the EMG on–off status. The activation
time for each muscle during the swing phase was calcu-
lated and expressed as a percentage of the whole swing phase
duration.
Statistical Analysis
The statistical analysis was performed using the SSP 2.5
statistical package (Smith’s Statistical Package, version
2.75, 2004; Gary Smith, Pomona College, Claremont,
­California). With regard to the primary outcome measures,
that is, maximum ankle dorsiflexion during the paretic side
swing phase and gait speed, an effect size was calculated
by considering the mean result of the EG minus the mean
result of the CG divided by the standard deviation of the
CG for the dependent variable.15 The c2 or 2-sample t test
was applied to compare the pretreatment and posttreatment
data. A 2-way ANOVA with group (experimental vs con-
trol) as the between-subjects factor and time (pretreatment
and posttreatment) as the within-subjects factor was used to
detect any significant differences between the pretreatment
and posttreatment evaluations, between the EG and CG and
within each group. A Tukey post hoc comparison was used
to detect any significant differences between the mean
values when a significant main effect and interaction were
found. The level of significance was set at P < .05 for all
analyses.
Results
A total of 44 patients were randomized (Figure 1) to either
the EG (n = 22) or the CG (n = 22). The demographic and
clinical baseline characteristics of the 2 groups were well
balanced (Table 1). No adverse effects were observed in
either the EG or CG, and all the patients completed the
treatment and performed the posttreatment assessment.
With regard to the time–distance gait parameters, 2-way
ANOVA showed a significant treatment effect (P < .01) for
gait speed, unaffected side’s swing velocity, stride length on
both the paretic and normal sides, and toe-off percentage on
the normal side (Table 2). Significant differences at the post
hoc analysis are shown in Table 2. The kinematic evalua-
tion during the stance phase revealed a significant difference
at the 2-way ANOVA for both the paretic and normal ankle
dorsiflexion angles at heel contact (P < .01); no significant
changes emerged in the other parameters (Table 3). With
regard to the swing phase, significant differences were
observed in the maximum degrees of ankle dorsiflexion and
Paoloni et al 257

Table 2. Time–Distance Characteristics of Gait in Experimental

Assessment for eligibility (n = 64)
and Control Groups at Pretreatment and Posttreatment
Excluded
Not meeting inclusion criteria Evaluationsa
Enrollment (n = 13)
Refused to participate(n = 7) Experimental Control
Group Group P Valueb
Randomized (n = 44)
Toe-off normal (%)
Pre 67.0 (6.2) 67.2 (9.5) .999
Post 64.1 (5.7) 65.7 (10.1) .932
SMV + General physical therapy General physical therapy P valuec .434 .964
Toe-off paretic (%)
Allocated (n = 22) Allocated (n = 22)
Received full intervention (n = 22) Received full intervention (n = 22) Pre 62.6 (5.8) 65.1 (6.2) .313
Post 59.6 (5.5) 64.1 (8.0) .038
P valuec .040 .732
Cadence (step/min)
Pre 75.1 (12.0) 71.9 (21.5) .980
Post 80.4 (13.5) 74.1 (21.7) .989
Analyzed-(n = 22) Analyzed-(n = 22) P valuec .459 .774
Step length normal (m)
Figure 1. Flow-chart Diagram of the Study. Abbreviations: SMV, Pre 0.34 (0.12) 0.35 (0.09) .999
segmental muscle vibration. Post 0.40 (0.11) 0.37 (0.10) .937
P valuec .196 .960
Step length paretic (m)
Table 1. Demographic and Clinical Characteristics of
Pre 0.36 (0.11) 0.35 (0.07) .929
Patients in Both Groups
Post 0.39 (0.09) 0.35 (0.08) .174
Experimental Control P valuec .089 .974
Group Group P Value Stride length normal (m)
Pre 0.71 (0.20) 0.69 (0.14) .903
Age (years)a 59.5 (13.3) 62.6 (9.5) .06 Post 0.82 (0.18) 0.71 (0.15) .09
Height (cm)a 169.5 (7.9) 170.6 (8.1) .45 P valuec .01 .831
Weight (kg)a 75.8 (12.8) 76.0 (10.8) .22 Stride length paretic (m)
Male (%) 86.4 90.9 .13 Pre 0.70 (0.19) 0.70 (0.13) .970
Female (%) 13.6 9.1 .13 Post 0.79 (0.17) 0.73 (0.17) .238
Time since acute 1.85 (0.59) 1.86 (0.61) .44 P valuec .02 .728
event (years)a Step width normal (m)
Right hemiplegia 50 54.5 .28 Pre 0.18 (0.05) 0.19 (0.03) .948
(% of total) Post 0.17 (0.04) 0.19 (0.04) .549
Left hemiplegia 50 45.5 .24 P valuec .513 .999
(% of total) Step width paretic (m)
Lower extremity 4.04 (0.2) 4.04 (0.2) .50 Pre 0.18 (0.05) 0.20 (0.04) .847
Brunnstrom stagea Post 0.17 (0.04) 0.19 (0.02) .727
a
P valuec .639 .962
Data are expressed as mean (standard deviation). Swing velocity
normal (m/s)
plantarflexion on the paretic side (2-way ANOVA; P < .01; Pre 1.32 (0.34) 1.33 (0.38) .995
Post 1.53 (0.39) 1.35 (0.29) .999
Table 4). Surface EMG during the swing phase revealed a
P valuec .05 .425
significant increase in the percentage of activation of the TA Swing velocity
muscle on the paretic side between the pretreatment and paretic (m/s)
posttreatment assessments (P < .01; Table 5). The effect Pre 1.19 (0.36) 1.22 (0.48) .993
size was .61 for maximum ankle dorsiflexion during the Post 1.35 (0.40) 1.23 (0.42) .885
paretic side swing phase and .33 for gait speed. P valuec .306 .999
Gait speed (m/s)
Pre 0.44 (0.13) 0.44 (0.21) .999
Discussion Post 0.53 (0.13) 0.46 (0.21) .674
P valuec .047 .988
In this pilot consideration-of-concept randomized con- a
Data are expressed as mean (standard deviation). Significant P values are
trolled trial, we measured, to the best of our knowledge for shown in boldface.
the first time, the effect of a combination of SMV and gen- b
Comparison between groups at each evaluation time.
eral physical therapy on walking ability in chronic stroke c
Comparison within each group between pretreatment and posttreatment.
258 Neurorehabilitation and Neural Repair 24(3)

Table 3. Kinematic Characteristics During Stance Phase Table 4. Kinematic Characteristics During Swing Phase
of Experimental and Control Groups at Pretreatment and of Experimental and Control Groups at Pretreatment and
Posttreatment Evaluationsa Posttreatment Evaluationsa

Experimental Control Experimental Control

Group Group P Valueb Group Group P Valueb

Normal ankle Normal

angle HC   ankle ROM
Pre -2.99 (6.55) -1.90 (5.34) .185 Pre 11.35 (4.79) 11.09 (4.07) .999
Post 3.93 (5.28) 0.07 (4.43) .098 Post 11.38 (5.44) 10.37 (3.82) .946
P valuec .0001 .634 P valuec .999 .983
Paretic ankle Paretic
angle HC   ankle ROM
Pre -7.86 (5.31) -4.30 (7.08) .281 Pre 8.01 (3.75) 10.54 (6.91) .500
Post -1.15 (7.10) -2.76 (6.86) .874 Post 9.40 (4.05) 9.45 (4.00) .999
P valuec .001 .888 P valuec .593 .931
Normal knee Paretic
angle HC   ankle PF
Pre 13.97 (6.97) 10.66 (8.56) .421 Pre -9.27 (5.09) -7.06 (6.83) .843
Post 14.78 (6.01) 11.00 (6.52) .369 Post -2.93 (7.98) -5.61 (5.42) .704
P valuec .972 .999 P valuec .001 .948
Paretic knee Paretic
angle HC   ankle DF
Pre 8.66 (6.26) 8.83 (9.86) .999 Pre -1.26 (6.40) 3.48 (5.93) .341
Post 8.67 (6.38) 8.61 (6.24) .999 Post 6.47 (8.35) 3.84 (4.30) .747
P valuec .999 .999 P valuec .0002 .999
Normal hip
angle HC Abbreviations: ROM, range of motion; PF, maximum plantarflexion angle;
DF, maximum dorsiflexion angle.
Pre 26.19 (6.37) 27.31 (7.48) .961 a
Data are expressed as mean (standard deviation). Significant P values are
Post 25.82 (9.26) 27.34 (6.08) .926 shown in boldface.
P valuec .997 .965 b
Comparison between groups at each evaluation time.
Paretic hip c
Comparison within each group between pretreatment and posttreatment.
angle HC
Pre 18.62 (7.04) 20.59 (14.77) .913
Post 18.56 (6.95) 19.93 (11.14) .974 Table 5. Duration of Activation of Tibialis Anterior and
P valuec .999 .997 Gastrocnemius Medialis Muscles on Both Sides in the
Ankle ROM Experimental Group, Expressed as a Percentage of
normal the Whole Swing Phase Durationa
Pre 17.86 (6.27) 20.22 (14.57) .751
Post 18.32 (5.85) 17.43 (3.19) .983 Percentage of Activation
P valuec .997 .710
Ankle ROM Paretic side tibialis anterior
plegic Pre 68.33 (33.11)
Pre 17.12 (5.22) 18.06 (6.25) .922 Post 83.57 (19.43)
Post 16.27 (4.49) 16.97 (5.37) .968 P valueb .0038
P valuec .924 .912 Normal side tibialis anterior
Pre 82.67 (19.53)
Abbreviations: HC, heel contact; ROM, range of motion. Post 75.12 (34.11)
a
Data are expressed as mean (standard deviation). Significant P values are P valueb .002
shown in boldface. Paretic side gastrocnemius medialis
b
Comparison between groups at each evaluation time.
c Pre 49.26 (39.80)
Comparison within each group between pretreatment and posttreatment.
Post 46.25 (43.45)
P valueb .328
Normal side gastrocnemius medialis
patients. A moderate improvement in gait speed and sagittal Pre 36.11 (39.83)
ankle kinematics during the swing phase of the gait cycle, Post 20.42 (30.11)
both of which persisted 1 month after treatment ended, was P valueb .07
achieved in the EG but not in the CG. a
Data are expressed as mean (standard deviation). Significant P values are
In particular, the gait analysis evaluation revealed in shown in boldface.
EG but not in CG a significant increase in gait speed and b
Comparison within experimental group between pretreatment and
swing velocity on the normal side; swing velocity also posttreatment.
Paoloni et al 259
increased on the paretic side, though without reaching
significance.
As gait speed is considered to be an important marker of
deficit severity and functional ability after stroke,23-25 speed
gains resulting from the application of physical treat-
ments are the object of considerable interest in clinical
gait studies.1 Patients in the EG, though not those in the
CG, displayed a significant increase in walking speed
(+0.09 m/s), comparable with that achieved by means of
some current exercise therapy regimens, that ranged between
0.04 and 0.20 m/s.1 Despite this improvement, our patients
could not be classified in a different walking speed func-
tional level after treatment. Indeed, the results of both the
pretreatment (0.44 m/s) and posttreatment (0.53 m/s) exam-
inations fell within the restricted community ambulation
group, as commonly reported in the literature.1 Our finding
of an effect size for gait speed of 0.33, which is consid-
ered less than moderate,15 suggests that our protocol was
underpowered in regard to this specific outcome or better
outcomes may be achievable with greater intensity and
duration of the combination intervention.
Several factors contribute to slow walking speeds in indi-
viduals with chronic hemiplegia. Weakness of the paretic
plantar flexor muscles during the push-off phase of the gait
cycle is considered to be of primary importance.26,27 Recent
findings, however, highlight the importance of dorsiflexor
muscles strength on the paretic side in determining gait
velocity.28,29 Indeed, dorsiflexor weakness causes inade-
quate foot clearance during the swing phase of the gait
cycle, thereby increasing the swing time and reducing gait
speed.29 It is noteworthy that patients in our sample showed
increased swing velocity in both legs after treatment, which
implies that the treatment was most effective in the swing
phase of the gait cycle. We also observed a significant
increase in the mean activation time of the TA on the paretic
side during the swing phase of the gait cycle, though this
increase was not associated with significant changes in the
activation pattern of the paretic leg musculature. Indeed,
almost all the patients displayed some degree of coactiva-
tion of the plantarflexion and dorsiflexor muscles throughout
the swing phase in both the pretreatment and posttreatment
evaluations (data not shown), as might be expected because
coactivity between agonist/antagonist pairs is used as a coor-
dinative strategy to adapt to primary impairments in muscle
force output in stroke patients.30 Although correlations have
been found between improvements in synergistic motor con-
trol and walking ability31 or walking speed,23 recent findings
suggest that functional gait improvement is more closely
related to compensatory, adaptive mechanisms of the nonpa-
retic leg and trunk in motor control than to the recovery of the
appropriate muscle activation sequence.32 We suggest, there-
fore, that the increase in the mean activation time of the ankle
dorsiflexor muscles during the swing phase may be the
primary determinant of gait improvement in our sample
of patients; this hypothesis is supported by observations in
other conditions of functional impairment, such as func-
tional ankle instability, where the reduced duration of
muscular activation of the dorsiflexor and evertor muscles
has proved to be an important cause of motor impairment.33
Moreover, our results revealed significant changes in
sagittal ankle joint kinematics, which further supports a
treatment effect of SMV. At the posttreatment evaluation,
the patients in the EG displayed a significant increase in
both the paretic and normal ankle dorsiflexion angles at the
heel contact phase of stance and a reduced ankle plantarflex-
ion angle as well as greater maximum ankle dorsiflexion
angle during the swing phase of the paretic side. These
kinematic changes allowed our subjects to improve both
the foot clearance during the swing phase and the foot
position during initial contact in the stance phase at the
posttreatment evaluation. Increased ankle dorsiflexor muscle
activity may explain our kinematic findings, as the normal
activation pattern of the TA muscle during gait cycle sug-
gests.21 To ensure the maximum reproducibility of the
kinematic measurement in 2 sessions, offset angles during
standing were obtained for each patient before a walking
session started and then subtracted from those obtained
during walking, the resulting angles then being used for the
kinematic analysis. Moreover, marker positioning was made
according to a validated and standardized model19 based
on well-defined bony landmarks. This is, to our knowl-
edge, one of the best ways of minimizing systematic errors
during kinematic measurements made with optoelectronic
stereophotogrammetric systems.34 With regard to the spa-
tial parameters, the stride length of both the paretic and
normal legs significantly increased after treatment in the
EG, whereas step length increased, though not significantly,
on both the normal and paretic sides. Interestingly, the gain
in stride length was similar on both sides, which thus ensured
that the gait pattern remained symmetrical. We speculate
that this occurred because the patients in our sample gener-
ated a similar force in both legs when walking,35 even after
treatment ended.
The exact mechanism through which SMV acts on the
sensorimotor system is still unclear. Trains of high-frequency,
low-amplitude mechanical vibrations, applied to a muscle–
tendon unit, generate Ia afferent fiber discharges because of
the activation of muscle spindles.10 The modulation of Ia
inputs alters the excitability of the corticospinal pathway11
as well as the activation of cortical motor regions.36 During
low-amplitude biceps muscle tendon vibration, most of
the functionally linked M1 (primary motor cortex) cells
respond mainly with excitatory firing.36 The reason for this
phenomenon may be found in the corticocortical excitatory
connections between somatosensory areas elicited by the Ia
input and motor areas, such as M1. The increased neuronal
260 Neurorehabilitation and Neural Repair 24(3)
excitability evoked by SMV has been demonstrated in recent
TMS studies,12,13 in which a decrease in motor threshold, a
facilitation of motor-evoked potentials in response to single-
pulse TMS, and a reduction in intracortical inhibition (ICI)
associated with an increase in intracortical facilitation were
observed in the vibrated muscle. Moreover, the fact that the
excitability remained unchanged in other cortical motor
representations suggests that this increased neuronal exci­
tability appears to be specific to the vibrated muscle’s
movement representation.12,13 The interaction between pro-
prioceptive input evoked by peripheral muscle vibration and
mechanisms of increased intracortical excitability most prob-
ably takes place upstream of the corticospinal neurons and
reflects reduced ICI, arguably due to a reduced excitability
of GABAergic interneurons.12 Because the architecture of
inhibitory circuits is crucial to dynamic processes and map
organization within the cortex,37 this reduced inhibition
might induce temporary associations between adjacent cell
groups, thereby enabling reorganization. Group Ia afferent
discharges also appear to contribute to the triggering of
locomotor phase transitions, as suggested by findings indi-
cating that vibration of the quadriceps during the stance
phase of walking in human subjects induces an earlier onset
of TA activity.38 It should be noted, moreover, that muscle
vibration has effects not only in the contralateral but also in
the ipsilateral hemisphere, thereby modulating the relation-
ship between the 2 hemispheres,39 which may explain, at
least in part, the contralateral effects on the kinematic and
spatial–temporal characteristics observed in our sample.
Moreover, muscle vibration appears to be effective in
increasing homonymous muscle activation during gait in
spinal cord injury patients,40 thereby leading us to consider
a possible effect of this modality on local spinal cord modu-
lation. Particularly, it has been hypothesized that the local
increase in muscle activity associated with vibration during
walking could occur through stretch–reflex pathways,
mediated by muscle Ia afferents activation.40
The increase in joint proprioception may be another
possible mechanism of action through which SMV acts in
determining observed gait changes. Ankle joint position
sense does not appear to be directly related to gait perfor-
mance, but it may contribute significantly to variations in
gait velocity and stride length,28 probably as a compensatory
strategy in those patients who are unaware of where their
foot is during gait. In this regard, it is noteworthy that novel
rehabilitative approaches that combine simultaneous motor
and sensory stimulations appear to substantially improve
muscular strength and joint position sense in chronic stroke
patients,41 thereby confirming the strong impact of somato-
sensory stimulation on motor recovery.42
With regard to the physical parameters of SMV in our
work, the vibration frequency was set at 120 Hz, that is,
within a range (75-120 Hz) that seems to be particularly
effective on the central nervous network underlying motor
control,43 whereas vibration amplitude was set at 10 mm,
which was sufficient to drive Ia spindle afferents16 while
remaining subthreshold for the TVR18 and avoiding muscle
fiber injury.17 The long stimulation duration (ie, 30 min/day
for 12 days) was chosen because long-lasting effects on
cortical excitability have recently been obtained by means
of sustained vibratory stimulation on healthy subjects,44
suggesting that the effects of SMV might be related to the
duration of proprioceptive stimulation. We also chose a
device that allowed us to activate more than 1 muscle simul-
taneously (ie, TA and peroneus longus) as well as to firmly
fix the probes to the body by means of elastic bands so as
to avoid loss of contact with the muscle surface during
treatment.
Limitations
The main limitation of our study is the lack of a blinding
procedure for patients owing to the impossibility of simu-
lating the SMV. We did, however, try to limit the possible
bias by blinding evaluators and using what is considered to
be a highly objective evaluation, that is, instrumental gait
analysis. The results may also have been affected by the
additional time of intervention of 30 minutes in the EG. In
addition, we did not investigate whether the same results
can be achieved with a shorter period of stimulation or
better results with a longer duration of interventions. Future
experimental designs are warranted to determine the most
effective stimulation period. A longer course and increased
daily treatment dose would have increased the costs of this
study, which was not supported by external funding, so we
had to limit our treatment period.
The present study did not include an evaluation of
functional outcome measures (eg, functional ambulation
categories and motricity index) that would have allowed
us to interpret the results for their clinical meaning. We
decided not to include these assessments because they may
have become biased by lack of blinding of the subjects. Last,
as we only scheduled a 1-month posttreatment assessment,
no information could be obtained regarding the long-lasting
effects of our SMV intervention.
Conclusion
Our results show that SMV added to general physical ther-
apy may improve gait performance in patients with foot drop
secondary to chronic stroke. Future studies are warranted
to determine the optimal dosage of SMV, and in particular,
the minimum dose necessary to reach a therapeutic res­
ponse plateau. A phase 3 randomized controlled trial with an
appropriate sample size calculation, based in part on the
effect size determined in this study, can also be considered.
Paoloni et al 261
Declaration of Conflicting Interests
The authors declared no potential conflicts of interests with
respect to the authorship and/or publication of this article.
Funding
The authors received no financial support for the research and/or
authorship of this article.
References
  1. Dickstein R. Rehabilitation of gait speed after stroke: a criti-
cal review of intervention approaches. Neurorehabil Neural
Repair. 2008;22:649-660.
  2. van de Port IG, Kwakkel G, van Wijk I, Lindeman E. Suscep-
tibility to deterioration of mobility long-term after stroke: a
prospective cohort study. Stroke. 2006;37:167-171.
  3. O’Loughlin JL, Robitaille Y, Boivin JF, Suissa S. Incidence
of and risk factors for falls and injurious falls among the
community-dwelling elderly. Am J Epidemiol. 1993;137:
342-354.
  4. Jorgensen L, Engstad T, Jacobsen BK. Higher incidence of
falls in long-term stroke survivors than in population con-
trols: depressive symptoms predict falls after stroke. Stroke.
2002;33:542-547.
  5. Hamzat TK, Kobiri A. Effects of walking with a cane on bal-
ance and social participation among community-dwelling
post-stroke survivors. Eur J Phys Rehabil Med. 2008;44:
121-126.
  6. Duncan PW, Sullivan K, Behrman A, et al. Protocol for the
Locomotor Experience Applied Post-stroke (LEAPS) trial: a
randomized controlled trial. BMC Neurol. 2007;7:39.
  7. Kwakkel G, Wagenaar RC, Twisk JW, Lankhorst GJ,
Koetsier JC. Intensity of leg and arm training after primary
middle-cerebral-artery stroke: a randomised trial. Lancet.
1999;354:191-196.
  8. Byl NN, Pitsch EA, Abrams GM. Functional outcomes can
vary by dose: learning-based sensorimotor training for pati­
ents stable poststroke. Neurorehabil Neural Repair. 2008;22:
494-504.
  9. Dias D, Laíns J, Pereira A, et al. Can we improve gait skills
in chronic hemiplegics? A randomised control trial with gait
trainer. Eura Medicophys. 2007;43:499-504.
10. Roll JP, Vedel JP, Ribot E. Alteration of proprioceptive mes-
sages induced by tendon vibration in man: a microneuro-
graphic study. Exp Brain Res. 1989;76:213-222.
11. Steyvers M, Levin O, Van Baelen M, Swinnen SP. Corticospi-
nal excitability changes following prolonged muscle tendon
vibration. Neuroreport. 2003;14:1901-1905.
12. Rosenkranz K, Pesenti A, Paulus W, Tergau F. Focal reduc-
tion of intracortical inhibition in the motor cortex by selective
proprioceptive stimulation. Exp Brain Res. 2003;149:9-16.
13. Rosenkranz K, Rothwell JC. Differential effect of muscle
vibration on intracortical inhibitory circuits in humans.
J Physiol. 2003;551:649-660.
14. Rollnik JD, Siggelkow S, Schubert M, Schneider U, Dengler R.
Muscle vibration and prefrontal repetitive transcranial magnetic
stimulation. Muscle Nerve. 2001;24:112-115.
15. Dobkin BH. Progressive staging of pilot studies to improve
phase III trials for motor interventions. Neurorehabil Neural
Repair. 2009;23:197-206.
16. Brown MC, Engberg I, Matthews PBC. The relative sensi-
tivity to vibration of muscle receptors of the cat. J Physiol.
1967;192:773-800.
17. Necking LE, Lundström R, Dahlin LB, Lundborg G, Thornell LE,
Fridén J. Tissue displacement is a causative factor in vibration-
induced muscle injury. J Hand Surg. 1996;21:753-757.
18. Hagbarth KE, Eklund G. The effects of muscle vibration in
spasticity, rigidity, and cerebellar disorders. J Neurol Neuro-
surg Psychiatry. 1968;31:207-213.
19. Davis RB, Ounpuu S, Tyburski D, Gage JR. A gait analy-
sis data collection and reduction technique. Hum Mov Sci.
1991;10:575-587.
20. Hermens HJ, Freriks B, Disselhorst-Klug C, Rau G. Devel-
opment of recommendations for SEMG sensors and sen-
sor placement procedures. J Electromyogr Kinesiol. 2000;
10:361-374.
21. Vaughan CL, Davis BL, O’Connor JC. Dynamics of Human Gait.
2nd ed. Cape Town, South Africa: Kiboho Publishers; 1999.
22. Ricamato AL, Hidler JM. Quantification of the dynamic prop-
erties of EMG patterns during gait. J Electromyogr Kinesiol.
2005;15:384-392.
23. Schmid A, Duncan PW, Studenski S, et al. Improvements
in speed based gait classifications are meaningful. Stroke.
2007;38:2096-2100.
24. Goldie PA, Matyas TA, Evans OM. Deficit and change in gait
velocity during rehabilitation after stroke. Arch Phys Med
Rehabil. 1996;77:1074-1082.
25. Patterson SL, Forrester LW, Rodgers MM, et al. Determinants
of walking function after stroke: differences by deficit sever-
ity. Arch Phys Med Rehabil. 2007;88:115-119.
26. Nadeau S, Gravel D, Arsenault AB, Bourbonnais D. Plan-
tarflexor weakness as a limiting factor of gait speed in stroke
subjects and the compensating role of hip flexors. Clin Bio-
mech. 1999;14:125-135.
27. Milot MH, Nadeau S, Gravel D. Muscular utilization of the
plantarflexors, hip flexors and extensors in persons with
hemiparesis walking at self-selected and maximal speeds.
J Electromyogr Kinesiol. 2007;17:184-193.
28. Lin SI. Motor function and joint position sense in relation to
gait performance in chronic stroke patients. Arch Phys Med
Rehabil. 2005;86:197-203.
29. Lin PY, Yang YR, Cheng SJ, Wang RY. The relation between
ankle impairments and gait velocity and symmetry in people
with stroke. Arch Phys Med Rehabil. 2006;87:562-568.
30. Lamontagne A, Richards CL, Malouin F. Coactivation dur-
ing gait as an adaptive behavior after stroke. J Electromyogr
Kinesiol. 2000;10:407-415.
262 Neurorehabilitation and Neural Repair 24(3)
31. Kollen B, van de Port I, Lindeman E, Twisk J, Kwakkel G.
Predicting improvement in gait after stroke: a longitudinal
prospective study. Stroke. 2005;36:2676-2680.
32. Buurke JH, Nene AV, Kwakkel G, Erren-Wolters V, Ijzerman
MJ, Hermens HJ. Recovery of gait after stroke: what changes?
Neurorehabil Neural Repair. 2008;22:676-683.
33. Santilli V, Frascarelli MA, Paoloni M, et al. Peroneus longus
muscle activation pattern during gait cycle in athletes affected
by functional ankle instability: a surface electromyographic
study. Am J Sports Med. 2005;33:1183-1187.
34. Chiari L, Della Croce U, Leardini A, Cappozzo A. Human
movement analysis using stereophotogrammetry. Part 2: instru-
mental errors. Gait Posture. 2005;21:197-211.
35. Balasubramanian CK, Bowden MG, Neptune RR, Kautz SA.
Relationship between step length asymmetry and walking
performance in subjects with chronic hemiparesis. Arch Phys
Med Rehabil. 2007;88:43-49.
36. Fourment A, Chennevelle JM, Belhaj-Saïf A, Maton B.
Responses of motor cortical cells to short trains of vibration.
Exp Brain Res. 1996;111:208-214.
37. Yen C, Wang R, Liao K, Huang C, Yang Y. Gait training
induced change in corticomotor excitability in patients
with chronic stroke. Neurorehabil Neural Repair. 2008;22:
22-31.
38. Verschueren SM, Swinnen SP, Desloovere K, Duysens J.
Vibration induced changes in EMG during human locomo-
tion. J Neurophysiol. 2003;89:1299-1307.
39. Swayne O, Rothwell J, Rosenkranz K. Transcallosal sensorim-
otor integration: effects of sensory input on cortical projections
to the contralateral hand. Clin Neurophysiol. 2006;117:855-863.
40. Cotey D, Hornby TG, Gordon KE, Schmit BD. Increases in
muscle activity produced by vibration of the thigh muscles
during locomotion in chronic human spinal cord injury.
Exp Brain Res. 2009;196:361-374.
41. Cordo P, Lutsep H, Cordo L, Wright G, Cacciatore T, Skoss R.
Assisted movement with enhanced sensation (AMES): cou-
pling motor and sensory to remediate motor deficits in chronic
stroke patients. Neurorehabil Neural Repair. 2009;23:67-77.
42. Conforto AB, Kaelin-Lang A, Cohen LG. Increase in hand
muscle strength of stroke patients after somatosensory stimu-
lation. Ann Neurol. 2002;51:122-125.
43. Steyvers M, Levin O, Verschueren SM, Swinnen SP. Frequency-
dependent effects of muscle tendon vibration on corticospinal
excitability: a TMS study. Exp Brain Res. 2003;151:9-14.
44. Marconi B, Filippi GM, Koch G, et al. Long-term effects on
motor cortical excitability induced by repeated muscle vibra-
tion during contraction in healthy subjects. J Neurol Sci.
2008;275:51-59.