Neuroprosthesis for Footdrop

Compared with an Ankle-FootOrthosis: Effects on Postural Control during Walking

Haim Ring, MD, MSc,*† Iuly Treger, MD, PhD,*† Leor Gruendlinger, MS,xand Jeffrey M. Hausdorff, PhD‡x//

From the *Neurologic

Rehabilitation Center, Ra

Medicine, ‡Physical The

University, Tel-Aviv, Isra

Department, Tel-Aviv S

and //Division on Agin

chusetts.

Received May 19, 2008

August 26, 2008.

This work was funded

Journal of Stroke and C

Objectives: We sought to compare the effects of a radio frequency–controlled neuro-

prosthesis on gait stability and symmetry to the effects obtained with a standard an-

kle-foot orthosis (AFO). Methods: A total of 15 patients (mean age: 52.2 6 3.6 years)

with prior chronic hemiparesis resulting from stroke or traumatic brain injury (5.9 6

1.5 year) whose walking was impaired by footdrop and regularly used an AFO par-

ticipated in the study. There was a 4-week adaptation period during which partici-

pants increased their daily use of the neuroprosthesis, while using the AFO for the

rest of the day. Gait was then assessed in a 6-minute walk while wearing force-sen-

sitive insoles, by using the neuroprosthesis and the AFO in a randomized order. An

additional gait assessment was conducted after using the neuroprosthesis for a fur-

ther 4 weeks. Gait speed and stride time (inverse of cadence) were determined, as

were gait asymmetry index and swing time variability. Results: After the 4-week ad-

aptation period, there were no differences between walking with the neuroprosthe-

sis and walking with the AFO (P . .05). After 8 weeks, there was no significant

difference in gait speed, whereas stride time improved from 1.48 6 0.21 seconds

with the AFO to 1.41 6 0.16 seconds with the neuroprosthesis (P , .02). Swing

time variability decreased from 5.3 6 1.6% with the AFO to 4.3 6 1.4% with the neu-

roprosthesis (P 5 .01). A gait asymmetry index improved by 15%, from 0.20 6 0.09

with the AFO to 0.17 6 0.08 with the neuroprosthesis (P , .05). Conclusions: Com-

pared with AFO, the studied neuroprosthesis appears to enhance balance control

during walking and, thus, more effectively manage footdrop. Key Words:

Neuroprosthesis—functional electrical stimulation—ankle-foot orthosis—postural

control—gait.

� 2009 by National Stroke Association

Footdrop is one of the common gait impairments asso-

ciated with hemiplegia; an estimated 20% of all stroke

survivors have a footdrop.1 The conventional approach

to address footdrop is the prescription of an ankle-foot or-

thosis (AFO), but this has significant drawbacks.2 Use of

an AFO may block normal ankle kinematics during gait

and prevent active ankle stability and balance reactions.

al Rehabilitation Department, Loewenstein

nana, Israel, †Departments of Rehabilitation

rapy, Sackler Faculty of Medicine, Tel-Aviv

el, xMovement Disorders Unit, Neurology

ourasky Medical Center, Tel-Aviv, Israel;

g, Harvard Medical School, Boston, Massa-

; revision received August 17, 2008; accepted

in part by Ness Ltd, Ra’anana, Israel.

erebrovascular Diseases, Vol. 18, No. 1 (Janua

Sensory feedback that is needed for integrated motor con-

trol may also be inhibited with an AFO. Furthermore, it

restricts the natural passive range of motion and the flex-

ibility of the ankle and foot, may limit walking ability on

uneven terrains, and may be uncomfortable to use.3 AFOs

can only be worn in shoes, and often the shoe with the

AFO must be larger in size than that of the other foot.

Portions of this work were presented at the 15th International

World Congress of Physical Therapy in Vancouver, British Columbia,

Canada, in June 6, 2007.

Address correspondence to Jeffrey M. Hausdorff, PhD, Movement

Disorders Unit, Neurology Department, Tel-Aviv Sourasky Medical

Center, Weizmann 6 Tel-Aviv, Israel. E-mail: jhausdorff@alum.mit.

edu.

1052-3057/$—see front matter

� 2009 by National Stroke Association

doi:10.1016/j.jstrokecerebrovasdis.2008.08.006

ry-February), 2009: pp 41-47 41

H. RING ET AL.42

As a result, the AFO is often rejected by patients.4 Geboers

et al5 concluded that an AFO does not improve walking

performance as measured by a 10-m walk test, and a re-

cent study also showed that, in the long run, the effective-

ness of the AFO is minimal.6 Despite these limitations,

AFOs have some advantages such as providing firm an-

kle stability for less cognizant patients. It is also very sim-

ple to use and its cost is relatively low. AFOs are probably

the most common treatment for footdrop today.

Externally induced dorsiflexion using functional elec-

trical stimulation (FES) was initially introduced by Liber-

son et al7 in 1961 as an alternative treatment for footdrop.

Since then, several footdrop stimulators have been devel-

oped.3,8-10 Such systems activate the muscles that dorsi-

flex the ankle and evert the subtalar joint during the

swing phase of gait, potentially providing several advan-

tages over the AFO. FES allows both greater passive and

active movement of the ankle, promoting proprioceptive

input that is essential for postural control.2,11 It does not

restrict push-off during the terminal stance, relevant for

several patients who have this ability, which is almost to-

tally restricted by the AFO. It enables foot adaptation to

uneven terrains, whereas the AFO restricts this adapta-

tion because of its firm structure.12 In addition, there is

evidence that stimulation of the common peroneal nerve

may trigger knee and hip flexion and thus facilitate the

flexion pattern needed for foot clearance during

swing.13,14 Other potential benefits of FES are prevention

of disuse atrophy, increased local blood flow, and muscle

re-education.13,15 These advantages and the ability to pro-

gram FES parameters to specific patient requirements10

support the idea that FES may be a preferred choice for

treating footdrop and might possibly yield better balance

control during walking.

Despite these potential benefits, clinical use of FES sys-

tems for correction of footdrop is not yet common.3 Among

the possible reasons for limited use are user-related draw-

backs inherent in previously available FES devices15 and

the lack of studies documenting improved efficacy of these

devices over conventional therapies, such as an AFO.

Although several reports have demonstrated the benefits

of such systems for the correction of footdrop,3,8,9,16 only

two studies directly compared a surface electrode foot-

drop stimulator with an AFO.17,18 Sheffler et al17 reported

promising results, but differences between the two devices

did not reach statistical significance and superiority of the

FES device over the AFO (or vice versa) could not be defin-

itively established. A recent study by Kottink et al18 evalu-

ated the effects of an implantable peroneal nerve

stimulator on walking speed in comparison with the

AFO. The participants, stroke survivors with chronic

hemiplegia, were randomly allocated to the treatment

group or to the control group (who continued their regular

use with their AFO). The implanted FES group improved

walking speed by 23% whereas the improvement in the

control group was only 3%. Although promising, both of

these studies focused on gait speed as their main outcome

measure; the effects on other aspects are not known.

A new FES neuroprosthesis for the treatment of

footdrop (NESS L300, NESS Ltd, Ra’anana, Israel) was re-

cently developed. This system includes features that were

intended to overcome barriers in the application of nonin-

vasive FES technology for lower limb activation. The ef-

fects of this system on mobility were previously

described.19 Consistent with previous FES studies, it

was found that gait speed improved and the physiologic

costs of walking were also reduced after patients walked

with the device for 8 weeks. Here we report on a subgroup

of that first study who were tested with the AFO and the

neuroprosthesis and compared the conventional treat-

ment for footdrop, the AFO, and the neuroprosthesis, spe-

cifically with respect to gait stability and symmetry. These

features of gait have been associated with function and

fall risk in various populations, even after taking into ac-

count gait speed.20-22 Given the system’s ability to adapt

in real time,19 we hypothesized that it would enhance

gait stability and symmetry, compared with the AFO.

Methods

Participants

We studied 15 patients with chronic hemiparesis. Par-

ticipants were recruited from two outpatient clinics in re-

habilitation centers. The criteria for patient selection were:

(1) diagnosis of an upper motor neuron lesion; (2) chronic

phase (.6 months postdiagnosis); (3) footdrop (toe drag

during walking); (4) regular use of an AFO as prescribed

by a physiatrist; (5) passive ankle range of motion to neu-

tral; (6) ability to walk at least 10 m independently or with

a cane; and (7) ability to follow multiple-step directions

and score greater than 23 on the Mini Mental State

Exam.23 Patients were excluded if they had a cardiac

pacemaker, skin lesion at the site of the stimulation elec-

trodes, or major depression as defined by Diagnostic and

Statistical Manual of Mental Disorders, Fourth Edition crite-

ria (a potential confounder). Patients were recruited for

a larger study designed to evaluate the effects of a neuro-

prosthesis on gait in patients with footdrop19; the current

study is based on a subset of those patients previously de-

scribed who were tested with their AFO. The criterion for

this subset was use of an AFO for at least 6 months before

the initiation of the study. Eleven male and 4 female pa-

tients participated in the study. Twelve patients were

poststroke and 3 patients were posttraumatic brain injury.

Six patients had right hemiparesis and 9 had left hemipa-

resis. The mean age was 52.2 6 3.6 years. The average

time postbrain injury was 5.9 6 1.5 years.

The Neuroprosthesis

The system includes an electronic orthosis, a control

unit, and a gait sensor that communicate by radio signals.

FES VS. AFO FOR FOOTDROP 43

The orthosis delivers electrical pulses to the common pe-

roneal nerve. These pulses are synchronized by the sensor

to activate the ankle dosiflexors during the swing phase of

gait and thus prevent footdrop. It may also be configured

to stimulate during part of the stance phase to improve

ankle stability while weight bearing. The hybrid orthosis,

designed to enable accurate and repeatable placement, in-

cludes two electrodes (45-mm diameter) and an inte-

grated configurable stimulation unit. The stimulating

electrodes are placed by a clinician before use. One elec-

trode is located over the common peroneal nerve, poste-

rior and distal to the fibular head, and a second

electrode is located over the tibialis anterior muscle to

achieve dorsiflexion with slight eversion. The movement

may be further adjusted by modifying the position of the

electrodes during the fitting process. The patient can then

place the orthosis using one hand. The gait sensor uses

dynamic gait recognition algorithms to detect events dur-

ing walking, and then transmits this information to the

rest of the system. It includes a pressure sensor worn un-

derneath the shoe insole at the heel with a small transmit-

ter that is attached to the shoe rim. When the system is

turned on, the gait sensor identifies the initiation of the

swing phase and triggers the stimulation accordingly.

A miniature control unit allows simple operation and

displays real-time information regarding the system’s

status. A handheld computer personal digital assistant

[PDA] with configuration software and interface is used

by a clinician to set the parameters of the system (e.g., tim-

ing, amplitude, pulse width, pulse frequency) and to ad-

just it to the patient’s gait characteristics (e.g., whether to

add stimulation during the stance time).

The AFO

Participants used their own plastic AFOs that were pre-

scribed by a physician during rehabilitation. Six patients

used a standard plastic off-the-shelf AFO set in neutral

position. Nine patients had a custom AFO with special

adjustments: 4 patients had AFOs with a hinge and

5 patients used an AFO with a dorsiflexion assist moment.

Procedures and Intervention

Patients provided written informed consent, as ap-

proved by our institutional review board. Basic demo-

graphic variables were collected as was significant

medical history. The stimulation (e.g., intensity, pulse fre-

quency) and gait parameters of the neuroprosthesis (e.g.,

extended time - the percentage of the stance time that the

stimulation continues after heel contact) were configured

individually for each patient. There was a 4-week adapta-

tion period during which participants increased their

daily use of the neuroprosthesis while using the AFO

for the rest of the day. The instructions for the participants

were as follows: ‘‘Gradually increase the use of the neuro-

prosthesis to an hour by the end of the first week, to four

hours by the end of the second week; you can use the neu-

roprosthesis up to 6 hours by the end of the fourth week.

During the adaptation period, keep using your AFO for at

least 2 hours a day.’’ After this 4-week adaptation period,

gait was measured under two conditions in a randomized

order: (1) while using the neuroprosthesis; and (2) while

using the AFO. An additional gait assessment was con-

ducted after 4 more weeks of neuroprosthesis use. During

these final 4 weeks, patients were encouraged to use only

the neuroprosthesis. During this period of time, the in-

structions for the participants were as follows: ‘‘Use the

neuroprosthesis all day long while walking.’’

Under each condition, patients walked on level ground

up and down a 50-m hallway at their self-selected, usual

walking speed for 6 minutes while wearing force-sensi-

tive insoles (B&L Footswitches, Tustin, CA) connected

to a data logger (JAS Research Inc, Belmont, MA)24 en-

abling measurement of temporal parameters of gait. The

patients were instructed to walk as far as they could in

6 minutes while turning around each time they reached

the end of the walkway. Average gait speed was deter-

mined by dividing the distance covered in 6 minutes by

360 seconds. Stride time (inverse of cadence) was deter-

mined to assess the walking pace. A gait asymmetry in-

dex and swing time variability (the single support

phase variability of the paretic leg) were calculated as

markers of gait stability and fall risk.20,21,25-27 The asym-

metry index was determined as follows21,28: 100 3

[(swing paretic – swing nonparetic)/(swing paretic 1

swing nonparetic)]. When the asymmetry index 5 0.0,

gait is perfectly symmetric. Symmetry indicates that the

swing time is similar in both limbs. Conversely, high

asymmetry indicates that weight bearing is unevenly dis-

tributed, an imbalance that may lead to an increased risk

of falls.21,22,29 The coefficient of variation (CV) of the

swing time was determined using previously described

methods to quantify balance during walking and the

intrinsic dynamics of steady-state walking. The CV is de-

fined by: SD/mean 3 100.20 The CV assesses the variabil-

ity or dysrhythmicity of gait, a measure previously

associated with fall risk.20,27,30 Swing time variability is

a measure of dynamic balance that is independent of

gait speed.20 To evaluate the participants’ acceptance of

the neuroprosthesis, patients were asked to report on

their preference regarding the AFO and the neuroprosthe-

sis during the last session of the study (week 8). During

the study, the participants were instructed to immediately

report any adverse event.

Statistical Analysis

A repeated measures analysis of variance, using gen-

eral linear models, was performed separately for each of

the different aspects of gait studied to analyze the effects

of the neuroprosthesis use under 3 conditions: AFO,

walking with the neuroprosthesis after the adaptation

H. RING ET AL.44

period, and walking with the neuroprosthesis after 8

weeks. If there were significant differences among the 3

conditions, post hoc analyses compared the AFO condi-

tion with the two neuroprosthesis conditions. Values are

summarized as mean 6 SD. A P value less than .05 was

considered significant.

Results

Gait Measures

Table 1 summarizes the group values for each gait pa-

rameter under the 3 test conditions and the results for

comparisons with the AFO. After the 4-week adaptation

period, there were no differences between walking with

the neuroprosthesis and walking with the AFO (P .

.05). After 8 weeks, the effects of the neuroprosthesis on

gait were significantly greater than those seen with the

AFO in 3 of the 4 outcomes that were measured. Although

there was no significant difference in gait speed with the

neuroprosthesis, there was a significant change in stride

time, gait asymmetry, and swing time variability. The

stride time was shorter (P 5 .02), the gait asymmetry in-

dex (Fig 1) was improved (P , .05), and the single limb

support of the paretic leg (swing time of the nonparetic

leg) also became less variable and more consistent

(P , .01).

Patient Perceptions

The patients’ perceptions of the neuroprosthesis were

very positive. For instance, 13 of the 15 patients reported

that they felt more stable with the neuroprosthesis and 14

patients indicated that their gait looked more normal.

Compared with the AFO, all 15 patients preferred to

use the neuroprosthesis for daily ambulation.

Discussion

The purpose of this investigation was to compare the

effects of a recently developed FES neuroprosthesis on

dynamic postural control with that of a traditional AFO

in a group of brain-injured patients long after the time

frame when spontaneous changes could possibly be

expected. The results support the hypothesis that the

use of the studied neuroprosthesis enhances gait symme-

try and rhythmicity compared with walking using an

Table 1. Effects of AFO and

Measure AFO Ne

6-min Walk gait speed (m/s) 0.58 6 0.06

Average stride time (s) 1.48 6 0.21

Swing time variability, nonparetic leg (%) 5.3 6 1.6

Swing asymmetry 0.20 6 0.09

Abbreviation: AFO, ankle-foot orthosis.

P values in parentheses show results of comparison with AFO (at 4 we

AFO. Previous work suggested that FES can improve cer-

tain aspects of gait in chronic hemiparetic patients with

footdrop (e.g., an improved energy consumption).3,8,9,16

In this study, we extend those previous findings by dem-

onstrating that the beneficial effects on gait are apparently

superior to the benefits achieved with an AFO. During the

initial adaptation period, the neuroprosthesis effect on

gait was similar to that obtained after chronic use with

an AFO, but after 8 weeks, the positive impact of the neu-

roprosthesis was greater than the AFO. The neuropros-

thesis improved the walking rhythmicity, and the gait

timing became less variable and more consistent, com-

pared with that seen with the AFO. Although gait speed

tended to improve, the changes were not statistically sig-

nificant. These results suggest that the use of the studied

neuroprosthesis is likely to enhance postural control dur-

ing gait, better than that achieved with an AFO.

This study also sheds light on a very important clinical

question. In contrast to the common clinical belief that

FES has an effect only in the swing phase and, therefore,

could not be an alternative for patients who present sta-

bility difficulties during the stance phase, a positive effect

of the FES during the stance (e.g., more consistent single

limb support, enhanced swing symmetry that relies on

a more stable stance) was observed. This could be ex-

plained by the ability of the neuroprosthesis to extend

the stimulation past the heel strike, providing an eccentric

contraction of the dorsiflexors during the loading re-

sponse, which assists the heel-rocker mechanism and

gives a better perception of the terrain resulting from di-

rect contact with the shoe. In addition, the foot can be

maintained in slight eversion during the initial stance,

keeping the movement of the center of pressure through

the midline and not along the lateral border, as would

often be the case with hemiplegic gait. Further biome-

chanical studies should confirm these possible

explanations.

Few studies have evaluated the effects of AFOs on pos-

tural stability and balance in hemiplegic patients. Wang et

al6 examined the effects of the AFO on balance perfor-

mance in patients with hemiparesis of short (,6 months)

and long (.12 months) duration. The measurements in

that study included balance evaluations by the Balance

Master and the Berg Balance Scale and gait speed and

cadence measurement during a 10-m walk. The authors

neuroprosthesis on gait

uroprosthesis adaptation Neuroprosthesis postadaptation

0.61 6 0.06 (.953) 0.67 6 0.06 (.142)

1.47 6 0.18 (.971) 1.41 6 0.16 (.022)

5.1 6 2.0 (.436) 4.3 6 1.4 (.009)

0.19 6 0.09 (.406) 0.17 6 0.08 (.048)

eks).

0.16

0.18

0.2

0.22

0.24

AFO Neuroprosthesisadaptation

Neuroprosthesispost adaptation

Sw

in

g A

sym

metry (in

dex)

Figure 1. Gait asymmetry index in 3 different conditions (AFO, neuro-

prosthesis adaptation [i.e., after 4 weeks of use], and neuroprosthesis posta-

daptation [i.e., after 8 weeks]). Error bars reflect SE. Continued use of

neuroprosthesis apparently improved interlimb symmetry, above and beyond

that seen with AFO.

FES VS. AFO FOR FOOTDROP 45

reported that for patients with hemiparesis for at least 12

months, AFOs did not have a significant impact on bal-

ance and gait. Two explanations were proposed for the

limited benefits of an AFO, reflecting effects on both the

ascending and descending pathways. The proprioceptive

input is decreased as a result of ankle supports with re-

strictive properties, and the physical restrictions on ankle

joint movement prevent the re-establishment of a normal

ankle strategy.31 Indeed, in the current study, enhanced

balance control was achieved while walking with an

FES neuroprosthesis, which does not limit ankle move-

ment or decrease proprioceptive input. Another interest-

ing potential explanation may be related to the method

of activation. In addition to recruiting the dosiflexors,

stimulation of the common peroneal nerve facilitates

knee and hip flexion.13,14 The neuroprosthesis used in

the current study is based on gait recognition algorithms

that are designed to optimize control of the foot move-

ment in the appropriate point of the gait cycle. Rather

than simply detecting the gait events (e.g., heel-off and

heel contact), the algorithms calculate a moving average

of swing/stance time and loading to continuously adapt

to various parameters. For instance, when a patient

moves from a firm terrain (e.g., hard surfaced floor) to

a soft terrain (e.g., lawn, sand), the peak loads will

change. The system will immediately react to the changed

environment and will adapt accordingly. The precise tim-

ing of the foot movement, and the hip and knee flexion fa-

cilitation, may allow walking to become more automatic

and enable transfer of cognitive resources away from

gait. In contrast, this facilitation is less likely to take place

while using an AFO. Further investigations using dual

task paradigms that have been used to identify the auto-

maticity of gait components27 may be helpful for testing

these ideas.

Nevertheless, the results did not demonstrate a mark-

edly increased gait speed when using the neuroprosthesis

compared with the AFO. The direct action of the neuro-

prosthesis may improve the safety of gait and patient’s con-

fidence irrespective of the effect on velocity. Furthermore,

the patients’ average walking speed with the AFO was

0.6 m/s, which offers little potential for improvement in

such a short time and without any special gait training

intended to improve gait velocity. These findings are con-

sistent with the findings of Granat et al,9 who also showed

an improvement in some gait parameters related to

balance but not in gait speed when using a peroneal

stimulator.

A strength of the current study is the use of the 6-min-

ute walk test. This is a functional test that covers several

aspects of gait performance and mobility.32 It differs

from the standard 10-m, self-paced test of gait because

it requires sustained walking activity over a relatively ex-

tended period of time.33 The demonstrated advantages of

the FES neuroprosthesis during this functional task high-

light the potential relevance of the results to everyday sit-

uations. The improvement at 8 weeks also suggests that

continued use may lead to further normalization of gait;

however, additional studies are needed to evaluate this

intriguing possibility.

Given a choice between the FES neuroprosthesis and

the AFO, all patients preferred the neuroprosthesis for

daily ambulation. Several factors may have contributed

to this preference for the neuroprosthesis. The effects on

gait demonstrated in this study likely played a role, but

other potential benefits to the patients such as better ap-

pearance, ability to walk with similar-sized shoes, the

feeling of more active walking, and lighter weight may

have also been involved. The high acceptance rate of the

neuroprosthesis, its preference over an AFO, and its pos-

itive effect on gait suggest that it may be a preferred

choice for use by patients with chronic hemiplegia and

may contribute to wider use of this technology in the neu-

rorehabilitation field. However, the specific benefits of

this device can be more fully addressed only in a study

that compares its effects with other FES systems, perhaps

while studying electromyography (EMG) and other bio-

mechanical properties. The readiness of the hemiplegic

population and clinicians to routinely use this neuropros-

thesis in extended use also remains to be seen.

This study has several limitations including the rela-

tively small sample size and the protocol duration (i.e.,

8 weeks). Further investigations should be undertaken

to confirm the study results with larger populations and

longer durations of use.

In addition, kinetics and kinematics studies may be

useful for more completely characterizing the observed

changes. In contrast to the current study, which focused

on patients with chronic hemiplegia, it may also be help-

ful to compare the use of the neuroprosthesis with that

obtained with an AFO in patients in the acute and sub-

acute stages of stroke. Another potential limitation of

the current study is that the protocol did not include

a measurement with the AFO before the adaptation pe-

riod with the neuroprosthesis. It is possible that the gait

H. RING ET AL.46

at 4 weeks with the AFO was affected by the training with

the neuroprosthesis during that 4-week period. Although

this possibility may have lead to an underestimation of

the effects of the neuroprosthesis and cannot be com-

pletely ruled out, we suggest that it is not a likely expla-

nation for the observed advantages of the

neuroprosthesis. This idea is supported by the fact that

the AFO walk was the reference and the fact that all pa-

tients had chronic conditions, minimizing the possibility

of spontaneous improvement in AFO walking. It is also

important to keep in mind that the neuroprosthetic effects

may have improved muscle function and timing, even

without any device. We used the measurement as a refer-

ence at week 4 to reflect the state of the participants who

had been using the AFO chronically for many years.

Nonetheless, future studies may also wish to compare

neuroprosthesis use at 8 weeks with that of AFO and no

prosthesis at 8 weeks to more fully understand the ob-

served changes in gait over time.

Conclusions

These findings suggest that, compared with an AFO,

the studied neuroprosthesis apparently yields better bal-

ance control and symmetry during walking and thus

may more effectively manage footdrop caused by stroke

or traumatic brain injury.

Acknowledgments: We thank the patients for their time

and participation.

References

1. Wade D, Wood W, Heller A, et al. Walking after stroke:Measurement and recovery over the first 3 months. ScandJ Rehabil Med 1987;19:25-30.

2. Kottink AI, Oostendorp LJ, Buurke JH, et al. The orthoticeffect of functional electrical stimulation on the improve-ment of walking in stroke patients with a dropped foot: Asystematic review. Artif Organs 2004;28:577-586.

3. Stein RB, Chong S, Everaert DG, et al. A multicenter trialof a footdrop stimulator controlled by a tilt sensor. Neuro-rehabil Neural Repair 2006;20:371-379.

4. Taylor PN, Burridge JH, Dunkerley AL, et al. Patient’sperceptions of the Odstock dropped foot stimulator(ODFS). Clin Rehabil 1999;13:439-446.

5. Geboers JF, Wetzelaer WL, Seelen HA, et al. Ankle-footorthosis has limited effect on walking test parametersamong patients with peripheral ankle dorsiflexor paresis.J Rehabil Med 2002;34:80-85.

6. Wang R, Yen L, Lee CC, Lin PY, Wang MF, Yang YR. Ef-fects of an ankle foot orthosis on balance performancein patients with hemiparesis of different durations. ClinRehabil 2005;19:37-44.

7. Liberson WT, Holmquest HJ, Scott D, et al. Functionalelectrotherapy, stimulation of the peroneal nerve syn-chronized with the swing phase of the gait of hemiplegicpatients. Arch Phys Med 1961;42:101-105.

8. Taylor PN, Burridge JH, Dunkerley AL, Wood DE,Norton JA, Singleton C, Swain ID. Clinical use of theOdstock dropped foot stimulator: Its effect on the speed

and effort of walking. Arch Phys Med Rehabil 1999;80:1577-1583.

9. Granat MH, Maxwell DJ, Ferguson AC, et al. Peronealstimulator; evaluation for the correction of spastic foot-drop in hemiplegia. Arch Phys Med Rehabil 1996;77:19-24.

10. Lyons GM, Sinkjaer T, Burridge JH, et al. A review ofportable FES based neural orthosis for the correctionof footdrop. IEEE Trans Neural Syst Rehabil Eng 2002;10:260-279.

11. Wright A, Mann GE, Swain ID. Training effects of electri-cal stimulation and the conventional ankle foot orthosisin the correction of dropped foot following stroke. Pre-sented at: First Annual Conference of FESnet; September2–3, 2002; Glasgow, Scotland.

12. Fatone S, Hansen AH. Effect of ankle-foot orthosis onroll-over shape in adults with hemiplegia. J Rehabil ResDev 2007;44:11-20.

13. Alfieri V. Electrical stimulation for modulation of spastic-ity in hemiplegic and spinal cord injury subjects. Neuro-modulation 2001;4:85-92.

14. Burridge J, Taylor P, Hagen S, et al. The effect of commonperoneal nerve stimulation on quadriceps spasticity inhemiplegia. Physiotherapy 1997;83:82-88.

15. Peckham PH, Knutson JS. Functional electrical stimula-tion for neuromuscular applications. Annu Rev BiomedEng 2005;7:327-360.

16. Shimada Y, Matsunaga T, Misawa A, et al. Clinicalapplication of peroneal nerve stimulator system usingpercutaneous intramuscular electrodes for correction offootdrop in hemiplegic patients. Neuromodulation2006;9:320-327.

17. Sheffler LR, Hennessey MT, Naples GG, et al. Peronealnerve stimulation versus an ankle foot orthosis for correc-tion of footdrop in stroke: Impact on functional ambula-tion. Neurorehabil Neural Repair 2006;20:355-360.

18. Kottink AI, Hermens HJ, Nene AV, et al. A randomizedcontrolled trial of an implantable 2-channel peronealnerve stimulator on walking speed and activity in post-stroke hemiplegia. Arch Phys Med Rehabil 2007;88:971-978.

19. Hausdorff JM, Ring H. Effects of a new radio frequencycontrolled neuroprosthesis on gait symmetry and rhyth-micity in patients with chronic hemiparesis. Am J PhysMed Rehabil 2008;87:4-13.

20. Hausdorff JM. Gait variability: Methods, modeling andmeaning. J Neuroengineering Rehabil 2005;20(2):19.

21. Yogev G, Plotnik M, Peretz H, et al. Gait asymmetry inpatients with Parkinson’s disease and elderly fallers:When does the bilateral coordination of gait requireattention? Exp Brain Res 2007;177:336-346.

22. Wolly SM. Characteristics of gait in hemiplegia. TopStroke Rehabil 2001;7:1-18.

23. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-mentalstate’’: A practical method for grading the cognitive stateof patients for the clinician. J Psychiatr Res 1975;12:189-198.

24. Sheridan PL, Solomont J, Kowall N, et al. Influence ofexecutive function on locomotor function: Divided atten-tion increases gait variability in Alzheimer’s disease.J Am Geriatr Soc 2003;51:1633-1637.

25. Donker SF, Beek PJ. Interlimb coordination in prostheticwalking: Effects of asymmetry and walking velocity.Acta Psychol (Amst) 2002;110:265-288.

26. Sackley CM. The relationship between weight-bearingasymmetry after stroke, motor function and activities ofdaily living. Physiother Theory Pract 1990;6:179-185.

FES VS. AFO FOR FOOTDROP 47

27. Springer S, Giladi N, Peretz C, et al. Dual-tasking effectson gait variability: The role of aging, falls, and executivefunction. Mov Disord 2006;21:950-957.

28. Hill KD, Goldie PA, Baker PA, et al. Retest reliability ofthe temporal and distance characteristics of hemiplegicgait using a footswitch system. Arch Phys Med Rehabil1994;75:577-583.

29. Harris-Love ML, Forrester LW, Macko RF, et al. Hemi-paretic gait parameters in overground versus tread-mill walking. Neurorehabil Neural Repair 2001;15:105-112.

30. Maki BE. Gait changes in older adults: Predictors of fallsor indicators of fear. J Am Geriatr Soc 1997;45:313-320.

31. Bunnell KL, Goldie PA. The difference effects of externalankle support on postural control. J Orthop Sports PhysTher 1994;20:287-295.

32. Pohl PS, Duncan PW, Perera S, et al. Influence of stroke-related impairments on performance in 6 minute walktest. J Rehabil Res Dev 2002;39:439-444.

33. Eng JJ, Chu KS, Dawson AS, et al. Functional walk tests inindividuals with stroke: Relation to perceived exertionand myocardial exertion. Stroke 2002;33:756-761.