Complex Versus Simple Ankle Movement Training in

Stroke Using Telerehabilitation: A Randomized

Controlled Trial
Huiqiong Deng, William K. Durfee, David J. Nuckley,
Brandon S. Rheude, Amy E. Severson, Katie M.
Skluzacek, Kristen K. Spindler, Cynthia S. Davey and
James R. Carey
PHYS THER. 2012; 92:197-209.
Originally published online November 17, 2011
doi: 10.2522/ptj.20110018

The online version of this article, along with updated information and services, can be
found online at: http://ptjournal.apta.org/content/92/2/197

Online-Only Material http://ptjournal.apta.org/content/suppl/2012/01/20/92.2.19

7.DC1.html
Collections This article, along with others on similar topics, appears
in the following collection(s):
Gait and Locomotion Training
Injuries and Conditions: Ankle
Patient/Client-Related Instruction
Randomized Controlled Trials
Stroke (Geriatrics)
Stroke (Neurology)
e-Letters To submit an e-Letter on this article, click here or click on
"Submit a response" in the right-hand menu under
"Responses" in the online version of this article.
E-mail alerts Sign up here to receive free e-mail alerts

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014

 Research Report

Complex Versus Simple Ankle

Movement Training in Stroke Using
Telerehabilitation: A Randomized
Controlled Trial H. Deng, MD, MS, Program in
Rehabilitation Science, Univer-
Huiqiong Deng, William K. Durfee, David J. Nuckley, Brandon S. Rheude, sity of Minnesota, Minneapolis,
Amy E. Severson, Katie M. Skluzacek, Kristen K. Spindler, Cynthia S. Davey, Minnesota.
James R. Carey W.K. Durfee, PhD, Department of
Mechanical Engineering, Univer-
sity of Minnesota.
Background. Telerehabilitation allows rehabilitative training to continue
D.J. Nuckley, PhD, Program in
remotely after discharge from acute care and can include complex tasks known to Physical Therapy, University of
create rich conditions for neural change. Minnesota.

B.S. Rheude, PT, DPT, United Hos-

Objectives. The purposes of this study were: (1) to explore the feasibility of using pital, St Paul, Minnesota.
telerehabilitation to improve ankle dorsiflexion during the swing phase of gait in
people with stroke and (2) to compare complex versus simple movements of the A.E. Severson, PT, DPT, Aegis
Therapies, St Cloud, Minnesota.
ankle in promoting behavioral change and brain reorganization.
K.M. Skluzacek, PT, DPT, North
Design. This study was a pilot randomized controlled trial. Memorial Hospital, Robbinsdale,
Minnesota.
Setting. Training was done in the participant’s home. Testing was done in sepa- K.K. Spindler, PT, DPT, St Lucas
rate research labs involving functional magnetic resonance imaging (fMRI) and Care Center/Back in Action Rehab,
multi-camera gait analysis. Faribault, Minnesota.

C.S. Davey, MS, Biostatistical

Patients. Sixteen participants with chronic stroke and impaired ankle dorsiflexion Design and Analysis Center, and
were assigned randomly to receive 4 weeks of telerehabilitation of the paretic ankle. Clinical and Translational Science
Institute, University of Minnesota.

Intervention. Participants received either computerized complex movement J.R. Carey, PT, PhD, Program in
training (track group) or simple movement training (move group). Physical Therapy, MMC 388, Uni-
versity of Minnesota, 420 Delaware
St SE, Minneapolis, MN 55455
Measurements. Behavioral changes were measured with the 10-m walk test and (USA). Address all correspondence
gait analysis using a motion capture system. Brain reorganization was measured with to Dr Carey at: carey007@umn.
ankle tracking during fMRI. edu.

[Deng H, Durfee WK, Nuckley DJ,

Results. Dorsiflexion during gait was significantly larger in the track group com- et al. Complex versus simple
pared with the move group. For fMRI, although the volume, percent volume, and ankle movement training in stroke
intensity of cortical activation failed to show significant changes, the frequency count using telerehabilitation: a ran-
of the number of participants showing an increase versus a decrease in these values domized controlled trial. Phys Ther.
2012;92:197–209.]
from pretest to posttest measurements was significantly different between the 2
groups, with the track group decreasing and the move group increasing. © 2012 American Physical Therapy
Association
Limitations. Limitations of this study were that no follow-up test was conducted Published Ahead of Print:
and that a small sample size was used. November 17, 2011
Accepted: August 5, 2011
Submitted: January 24, 2011
Conclusions. The results suggest that telerehabilitation, emphasizing complex
task training with the paretic limb, is feasible and can be effective in promoting
further dorsiflexion in people with chronic stroke. Post a Rapid Response to
this article at:
ptjournal.apta.org

February 2012 Volume 92 Number 2 Physical Therapy f 197

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
A
mid an increasingly difficult
health care economy, hospital
lengths of stay for people
with stroke have declined,1,2 which
affects the intensity of their rehabil-
itation. This scenario compels clini-
cians and scientists to be creative in
finding ways to promote better qual-
ity of life for these individuals,
which, in terms of walking, would
include safety, speed, and energy
expenditure. As communication
technology has advanced, a new
method of rehabilitation is emerging
that may allow rehabilitative training
to continue remotely following dis-
charge from acute care. This method
is called telerehabilitation,3 defined
here as therapy from a distance
directed by a computer and telecom-
munication. A variety of studies
exploring telerehabilitation in peo-
ple with stroke have shown value for
improving upper-limb function,3– 8
and lower-limb function.9 –11 None of
these studies has explored whether
telerehabilitation can improve the
range of ankle dorsiflexion (DF) dur-
ing the swing phase of gait, which
is commonly impaired after stroke12
and can lead to increased tripping
potential13 and energy expendi-
ture.14 Thus, the first objective of
this study was to explore the feasi-
bility of using telerehabilitation to
improve ankle DF during the swing
phase of gait in people with stroke.
This study also explored an impor-
tant topic related to the structuring
of the telerehabilitation sessions (ie,
complex versus simple movement
training). Research suggests that com-
Available With
This Article at
ptjournal.apta.org
• eFigure 1: Median pretest-
posttest change values for
outcome measures
• eFigure 2: Median values for
laterality index
198 f Physical Therapy Volume 92 Number 2
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
plex task training has advantages over tial processing to achieve accuracy,
simple task training. In rodents, Kleim versus (2) repetitive simple move-
et al15 showed that rats undergoing ments of the ankle, requiring no
repetitive movement training on a attention to accuracy. We hypothe-
complex reaching task demonstrated sized that complex movements would
greater functional and structural plas- yield greater improvements in ankle
ticity of relevant cortical regions com- DF during gait than simple move-
pared with animals undergoing train- ments and that a different pattern of
ing on a simple reaching task. In brain reorganization would emerge
primates, 2 combined studies16,17 have between the 2 training forms.
shown that repetitive training on a
complex manual task produced func- Method
tional reorganization of cortical maps, Design Overview
whereas repetitive training on a sim- This study was a pilot randomized
ple manual task did not. In humans controlled clinical trial comparing 2
who were healthy, Pascual-Leone et treatment strategies: complex move-
al18 showed that repetitive training on ment training and simple movement
a complex piano task demonstrated training.
greater motor learning and changes
in cortical excitability than repetitive Setting and Participants
training on a simple piano task. Nineteen participants were assigned
Accordingly, our second objective randomly to either a track group or a
was to compare the effects of move- move group. Sixteen participants, 8
ment complexity in 2 telerehabilita- in each group, completed the study
tion training strategies: (1) repeti- (Fig. 1). Inclusion criteria included
tive, complex movements (tracking) poststroke duration of at least 5
of the ankle, requiring temporospa- months, at least 10 degrees of active
The Bottom Line
What do we already know about this topic?
Telerehabilitation allows rehabilitation training to continue remotely after
discharge from acute care and can include complex tasks that are known
to create rich conditions for neural change. Research suggests that com-
plex task training has advantages over simple task training.
What new information does this study offer?
According to this study, a telerehabilitation program that emphasizes
complex task training with the paretic limb is feasible and can be effective
in promoting further dorsiflexion. Improved dorsiflexion will help
decrease tripping in people with chronic stroke.
If you’re a patient, what might these findings mean
for you?
After traditional one-on-one training, your physical therapist may provide
you with a telerehabilitation program that will allow you to take more
responsibility for completing designated tasks or exercises in your own
home, with guidance from a computer and periodic check-in from your
physical therapist.
February 2012

Dropout (n=1) Initial Move
because of fatigue Group (n=9)
Completed
Move Group
(n=8)
Figure 1.
Flow chart showing the number of participants at each stage of the study. MRI⫽magnetic resonance imaging.
dorsiflexion/plantar flexion (DF/PF)
at the paretic ankle, ability to under-
stand the tasks, and ability to ambu-
late 30 m. Exclusion criteria included
indwelling devices incompatible
with magnetic resonance imaging
and currently receiving therapy. The
Table summarizes the demographics
of each group. All participants
signed a statement of informed
consent.
Participant Orientation to
Telerehabilitation
Before enrollment, a randomization
listing of the 2 interventions was
generated electronically. As partici-
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
Phone screen
for eligibility
(n=62)
On-site screen
(n=23)
Total enrolled
(n=19)
Randomization
Analysis
pants enrolled in the study, they
were assigned to the next interven-
tion on the randomized list: track
group or move group. This assign-
ment occurred at the end of the pre-
test visit. Next, the investigator
responsible for training modified the
laptop computer (Dell VOSTRO
1,000, Round Rock, Texas) with the
customized training software so that
the proper program (track versus
move) would be started automati-
cally after pressing the power but-
ton. Then, the participant and a fam-
ily member were oriented in how to
set up the computer/camera and
apply the customized electrogoni-
Excluded (n=39)
Not enough active ankle motion (n=8)
Not compatible with MRI (n=5)
No ankle motor deficit (n=4)
Not interested after further information
(n=22)
Rejected (n=4)
Not enough active
ankle motion
Initial Track Dropout (n=2)
Group (n=10) because of fatigue
Completed
Track Group
(n=8)
ometer with an attached potentiom-
eter (ETI Systems Inc, Carlsbad, Cal-
ifornia) to the paretic ankle. Several
training trials then were observed to
ensure that the participant could fol-
low the procedures before being dis-
missed to repeat the setup at home.
Figure 2 (left panel) shows the home
setup for a participant in the track
group.
Training: Track Group
After pressing the power button on
the computer, the participant then
followed prompts to record the
active DF/PF range, which was used
to set the amplitudes for the tracking
Volume 92 Number 2 Physical Therapy f 199
Ankle Movement Training in Stroke Using Telerehabilitation

Table.
Participant Characteristics by Groupa

Stroke Paretic Preferred Foot

Participant Group Sex Age (y) Duration (mo) Stroke Location Side Prestroke MMSE

1 Track F 62 52 L subcortical (fronto-parietal) R R 30

2 Track F 65 124 L cortical/subcortical R R 29

(occipital)

3 Track M 51 80 R cortical/subcortical L R 30
(fronto-parieto-temporal)

4 Track F 28 110 R cortical (parietal) L R 30

5 Track F 46 12 R cortical/subcortical L R 30
(fronto-parieto-temporal)

6 Track M 58 94 L subcortical (frontal) R NP 28

7 Track F 53 12 R cortical/subcortical L R 29
(fronto-parieto-temporal)

8 Track M 48 52 R, L subcortical (parietal) L R 29

Frequency, median, n⫽8 3M 52.0 (47, 60) 66 (32, 102) 3R 7R 29.5 (29, 30)
(Q1, Q3) 5F 5L 0L
1 NP

9 Move M 62 15 L subcortical (parietal) R R 30

10 Move M 32 5 L subcortical (frontal) R R 29

11 Move M 53 35 R subcortical (frontal) L R 30

12 Move M 67 33 R subcortical (fronto-parietal) L R 28

13 Move M 55 18 R cortical/subcortical L L 28
(fronto-parietal)

14 Move M 79 5 R cortical/subcortical L R 29
(fronto-parietal)

15 Move M 59 32 R subcortical (parietal) L R 30

16 Move M 57 5 R cortical/subcortical L L 28
(fronto-parietal)

Frequency, median, n⫽8 8 M* 58.0 (54, 64) 16.5** (5, 32.5) 2R 6R 29.0 (28, 30)
(Q1, Q3) 0F 6L 2L
a
Foot preference was defined by the Modified Lateral Preference Inventory.49 MMSE⫽Mini-Mental State Examination,50 M⫽male, F⫽female, R⫽right,
L⫽left, NP⫽no preference, Q1⫽first quarter, Q3⫽third quarter. *⫽significant between-group difference (P⫽.026). **⫽significant between-group difference
(P⫽.039).

waveforms. The electrogoniometer
captured the movements of DF/PF
only; inversion/eversion movements
were not investigated in this initial
study.
The program involved 60 training
blocks per day, with 3 trials per
block, for 20 days for a total of 3,600
trials. This dosage of training dou-
bled the amount in our earlier tele-
rehabilitation study.4 Participants
selected their own daily schedule,
but the program limited the partici-
pant to no more than the 180 trials
per day to capitalize on offline con-
solidation processes occurring with
sleep and the passage of time.19 To
continually challenge participants,
the 60 training blocks per day were
randomly ordered from a host set of
100 blocks. Blocks varied across mul-
tiple parameters. Target waveforms
included square, left sawtooth, right
sawtooth, triangle, and sine. Fre-
quencies were 0.2, 0.25, 0.6, and 1
Hz. Trial durations were 5, 10, 15,
and 20 seconds. Thus, the number
of movement cycles (repetitions) of
DF/PF in a trial varied according to
the waveform frequency and dura-
tion. For example, a 0.6-Hz wave-
form during a 10-second trial would
generate 6 DF/PF movement repeti-
tions. The PF amplitude (lower
peaks) was set at 0%, 15%, 30%, 50%,
or 85% of full range (0%⫽full PF)
and the DF amplitude (upper peaks)
was set at 50%, 70%, 85%, 100%,
110%, 120%, or 125% of full range
(100%⫽full DF). The conditions
of 110%, 120%, and 125% were
included to promote further DF
range. Because “attention” to the
joint being trained has been shown
to be important in stimulating brain
reorganization,20 the training required
that participants pay close attention
to the task and occasionally problem-
solve spatial conflict blocks. These
blocks included a stimulus-response

200 f Physical Therapy Volume 92 Number 2 February 2012

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014

(S-R) noncompatibility condition in
which upward DF movement pro-
duced downward cursor move-
ment on the computer screen. Such
noncompatibile spatial processing
has been shown to be a potent factor
influencing cortical excitability in
primates.21 Stimulus-response non-
compatibility blocks (interspersed
randomly) comprised 15% of the total
of 1,200 training blocks. Each partic-
ipant’s training position, prompted
by the computer screen, varied
between sitting (60% of training
blocks) with the knee flexed or
extended or standing (40% of train-
ing blocks) on the paretic or nonpa-
retic leg.
During a trial, the cursor swept auto-
matically from left to right across the
screen at a constant rate while the
participant adjusted the vertical posi-
tion of the cursor to track the target
as accurately as possible with DF/PF
movements. At the end of a trial, a
pause occurred for a time equal to
the duration of the preceding trial.
Knowledge of results (KR)22 was
provided at the end of each trial for
the track group, with a computer-
calculated score modified from the
accuracy index (AI) described ear-
lier.23 This modification was done so
that participants would not experi-
ence any negative scores in describ-
ing their performance. Although
Anderson et al24 found better reten-
tion of motor skill in participants
who were healthy with a less fre-
quent schedule of KR, we could not
be certain that a reduced or faded
feedback schedule would apply sim-
ilarly for people with stroke. Thus,
we chose to give KR after every trial
because preliminary field testing on
people with stroke indicated that
they enjoyed competing with them-
selves for higher scores, and we
believed this competition would
help to keep our participants moti-
vated as they trained independently
at home.
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
Figure 2.
Left panel: Training at home for a participant in the track group. The training system
consisted of a laptop computer (A) (in this instance, showing 1:17 min/s into a
participant-selected pause from tracking), an ankle electrogoniometer brace (B), a web
camera (C), a single-button control box (in participant’s hands, not shown), and a
wireless modem (D). Right panel: Telecommunication setup at therapist’s location. The
therapist could see the participant’s tracking performance in one window. Simultane-
ously in another window, with the participant’s camera directed on the ankle, the
therapist also could see the actual ankle movement. With the therapist’s camera
directed on the therapist, the participant could see the therapist. Audio was transmitted
in both directions.
Teleconferencing between the par- and number of telecommunication
ticipant and the therapist occurred sessions. The main difference in
to reinforce human interaction and conditions between groups was that
the therapeutic relationship (Fig. 2, the track group saw the command
right panel). A remote desktop appli- “track” on their computer screen
cation (LogMeIn) and a video confer- along with the target waveform and
encing application (Skype, Skype their tracking response, whereas the
Ltd, Luxembourg, Belgium) were move group saw only the command
used for the communication, along “move” with no target, no response
with webcams (Logitech Webcam line and no KR at the end of each
C905). The Internet connection was trial. The move group was instructed
made using a cellular modem at orientation to produce repetitive
(USB720, Verizon Wireless, Verizon, DF/PF movements through their
New York, New York) communicat- full range at a comfortable, self-
ing with a cellular phone network selected frequency. Although the
(Verizon). The therapist contacted participants could not see their
the participant 2 times per week. In movement record, investigators
addition, the participant’s computer could see and confirm the move-
automatically e-mailed daily perfor- ment through the performance files
mance records to the laboratory that the participant’s computer auto-
computer, allowing the therapist to matically e-mailed to our laboratory
monitor compliance. regularly and through the telecom-
munication sessions.
Training: Move Group
The move group followed the same We did not control for movement
setup procedures. As much as possi- amplitude or frequency (ie, number
ble, the dosing of the training was of DF/PF movements per trial) in the
matched between groups. That is, move group, as we did for the track
both groups completed the same group, to eliminate the cognitive
number of training blocks, training processing associated with produc-
position within each block, number ing complex movements, which
of trials per block, duration per trial, formed the basis of our second
Volume 92 Number 2 Physical Therapy f 201
Ankle Movement Training in Stroke Using Telerehabilitation
objective. However, we did estimate
the frequency in each group across
all trials through an algorithm that
analyzed the data after they were col-
lected and counted the mean num-
ber of cursor band-crossings. For the
track group, the band was the mid-
dle third of the ankle motion
required to track each target. For the
move group, it was the middle third
of their full ankle motion. Then, for
each trial, the algorithm counted the
times that the cursor crossed the full
band (ie, either the minimum band
value followed by the maximum
band value or vice versa). The algo-
rithm also calculated the mean peak-
to-peak DF/PF amplitude within
each training trial. We deemed this
analysis to be important to rule out
amplitude and frequency of move-
ment as possible confounders, as
described further in the “Discussion”
section.
Outcome Measures
As this was a pilot study exploring
the feasibility of using telerehabilita-
tion in stroke, our measurement
points occurred at pretest and post-
test. We did not include a follow-up
test. The testers were blinded as to the
treatment group for each participant.
Gait assessment. Quantitative gait
kinematic parameters were mea-
sured using surface markers and an
8-camera motion capture system
(Vicon Inc, Los Angeles, California).
A modified Helen Hayes marker
arrangement was captured by the 4
megapixel cameras and analyzed using
the dynamic Plug-in-Gait model for
kinematic results. Lower-extremity
retroreflective markers were affixed
using the Helen Hayes convention25
and International Society of Biome-
chanics joint coordinate system
descriptions.26 Participants were
instructed to ambulate at their self-
selected speed without wearing an
ankle-foot orthosis (AFO). The 3-
dimensional (3D) marker data were
sampled at 120 Hz, smoothed with a
202 f Physical Therapy Volume 92 Number 2
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
forward/backward fourth-order low-
pass Butterworth filter, and used to
reconstruct limb segment motion
and joint angle kinematics. Ten full
gait cycles with complete data were
analyzed with ensemble averaging
for each pretest and posttest.
We analyzed ankle DF/PF angle, toe
clearance, gait temporal symmetry
ratio (GTSR), and stride length.
These metrics have been shown to
be critical for both predicting the
risk of falls27,28 and the improvement
of energy expenditure in patients
with stroke.29 –32 Maximum DF and
PF were the peak and trough points
of ankle angle during the swing
phase, and toe clearance was
defined as the minimum vertical dis-
placement of the great toe during
the swing phase.27 The GTSR,
defined below, quantified the ratio of
time spent in each phase of gait for
each limb, with a value of 1.0 repre-
senting healthy symmetric gait.29
Paretic Swing Time/
Paretic Stance Time
GTSR ⫽
Nonparetic Swing Time/
Nonparetic Stance Time
Ten-meter walk tests. Partici-
pants performed 2 trials of a 10-m
walk at their self-selected (“comfort-
able”) speed and then 2 trials at max-
imum speed. If participants normally
used a cane and AFO, they were
allowed to use the cane, but not
their AFO.
fMRI. Functional magnetic reso-
nance imaging (fMRI) occurred
inside a 3-Tesla magnet (Magnetom
Trio, Siemens, Munich, Germany).
The task for both groups consisted of
alternating phases of rest and paretic
ankle DF/PF tracking. Electrogoni-
ometers were attached to the paretic
ankle to perform the tracking and to
the nonparetic ankle to monitor for
mirror movements. Data from partic-
ipants who showed mirror move-
ments or head movements greater
than 3 mm were not included in the
fMRI analysis. The possibility of mir-
ror muscle activations below the
threshold for producing movement
is a limitation.
The functional task during fMRI con-
sisted of seven 1-minute phases alter-
nating between rest (4 phases) and
track (3 phases) conditions. For all
phases, the computer displayed a
random sine wave target (0.4 Hz)
along with the corresponding
prompt, “Rest” or “Track,” at the
bottom of the screen. The lower (PF)
peaks were set at 15% of the partic-
ipant’s predetermined ankle active
range of motion, with the partici-
pant’s full DF defined as 100%. The
upper (DF) peaks were set at 85%.
For each track phase, the cursor
swept from left to right across the
screen and the participant attempted
to track the target as accurately as
possible with careful DF/PF move-
ments. For each rest phase, the par-
ticipant watched the cursor sweep
across the screen but executed no
ankle movements. Two 10-second
practice trials occurred inside the
magnet. Tracking accuracy was
quantified with an AI.23 The maxi-
mum score is 100%.
Anatomical and functional images
were acquired using a whole-body
3-Tesla magnet (Magnetom Trio, Sie-
mens) equipped with a standard
12-channel head coil. A high-
resolution (1-mm3), longitudinal
relaxation time (T1)-weighted, 3D
anatomical image data set (3D
FLASH, repetition time [TR]⫽20
milliseconds, flip angle [FA]⫽30°,
total acquisition time⫽5:00 minutes)
was acquired over the entire brain to
identify appropriate landmarks and
serve as a template upon which
functional images would be overlaid.
Functional images were obtained
while the participant performed the
tracking test described above and
consisted of T2*-weighted functional
February 2012

magnetic resonance images of the
blood oxygen level– dependent
(BOLD) signal with a slice thickness
of 3 mm, which were obtained in
the transverse plane using a gradient
echo planar imaging sequence
(echo time [TE]⫽30 milliseconds,
TR⫽3,000 milliseconds, FA⫽80°,
field of view⫽192 ⫻ 192 mm with a
matrix size of 64 ⫻ 64 leading to a
resolution of 3 ⫻ 3 ⫻ 3 mm). The
total imaged volume extended from
the superior pole of the cortex to a
depth of 108 mm in 36 interleaved
slices. A block fMRI design was used
whereby 145 magnetic resonance
scans were acquired, for a total scan
time of 7:15 minutes, which covered
the time for all the alternating rest/
track phases.
Brain Voyager (Brain Innovation BV,
Maastricht, the Netherlands) soft-
ware was used for fMRI data prepro-
cessing and analysis. Functional
images were preprocessed to correct
for head motion artifacts, differences
in slice scan time acquisition, and
temporal linear trends. The 3D func-
tional volume was aligned with the
corresponding 3D anatomical vol-
ume, and both were normalized to
standard Talairach space.33
For each participant, regions of inter-
est (ROIs) were drawn manually on
anatomical images in each hemi-
sphere for the primary motor area
(M1), primary somatosensory area
(S1), premotor cortex (PMC), and
supplementary motor area (SMA)
according to specified landmarks34
(Fig. 3). We used a general linear
model (GLM) with 7 predictors to
record the active voxel count (vol-
ume) and the average BOLD signal
intensity, calculated as the average
t statistic of the difference between
the paretic ankle tracking and rest
conditions for each participant in
specified ROIs. One predictor was
the track condition. Three predictors
accounted for translational move-
ment of the head in sagittal, coronal,
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
Figure 3.
Regions of interest in sagittal, coronal, and transverse images of one participant’s brain
with left hemisphere stroke. A⫽anterior, P⫽posterior, R⫽right, L⫽left.
and transverse planes, and the
remaining 3 predictors accounted
for rotational movement in the same
3 planes. These last 6 predictors
were entered as covariates in the
model and served to exclude the
effect of any movement artifact in
the variability of BOLD signal. The
GLM analysis created an activation
map showing active voxels with sig-
nificantly different signal intensity
between paretic ankle tracking and
rest using a false discovery rate
(FDR) of q (FDR)⬍0.01.35
For each ROI, the number (volume)
of active voxels, percentage of total
ROI voxels (% volume), and average
signal intensity of active voxels was
recorded. Also, the volume and inten-
sity were compared between hemi-
spheres for each ROI by using the
laterality index (LI)36 calculated as:
共Voxel countIpsilesional兲 ⫺
共Voxel countContralesional兲
LIVolume ⫽
共Voxel countIpsilesional兲 ⫹
共Voxel countContralesional兲
共IntensityIpsilesional兲 ⫺
共IntensityContralesional兲
LIIntensity ⫽
共IntensityIpsilesional兲 ⫹
共IntensityContralesional兲 .
Data Analysis
For data sets that were normally dis-
tributed (ie, AI), we used paired t
tests to evaluate within-group differ-
ences from pretest to posttest and
Volume 92 Number 2
2-sample t tests for between-group
differences. For data sets that were
not normally distributed (ie, DF, toe
clearance, GTSR, stride length, 10-m
walk speed, and fMRI variables), we
used Wilcoxon signed rank tests to
evaluate within-group differences
and Wilcoxon rank sum tests for
between-group differences. We also
were interested in the number of
participants (ie, frequency count)
showing an increase or a decrease in
each fMRI variable from pretest to
posttest. We analyzed for a differ-
ence in these frequency counts
between groups using the Fisher
exact test. No adjustments to the sig-
nificance level were made for the
multiple comparisons because of
small sample size. Accordingly, sig-
nificant results should be considered
with caution for possible spurious
effects stemming from multiple com-
parisons. Beyond analyzing the test-
ing data, we also analyzed all partic-
ipants’ training data to compute the
mean number of movement repeti-
tions per training trial and the mean
peak-to-peak amplitude of move-
ment. The alpha level was set at
P⬍.05 (2-tailed). All analyses were
conducted using SAS statistical soft-
ware (version 9.2, 2008, SAS Insti-
tute, Cary, North Carolina).
Role of the Funding Source
This project was funded by the
following National Institutes of
Health grants: R03HD051615, P41
Physical Therapy f 203
Ankle Movement Training in Stroke Using Telerehabilitation

Figure 4.
Median (first quartile, third quartile, minimum, maximum) values of gait variables: (A) dorsiflexion, (B) toe clearance, (C) gait
temporal symmetry ratio (GTSR), and (D) stride length.

RR008079, and M01-RR00400. No
US federal agencies provided fund-
ing for this study. The funding
source had no involvement in the
design, conduct, or reporting of
information.
Results
Behavioral Outcomes
The data for the dependent measures
below (except AI) were not nor-
mally distributed. Thus, we used the
median (50th percentile) to describe
the central tendency and the 1st
quartile (Q1 or 25th percentile) and
3rd quartile (Q3 or 75th percentile)
to describe the variability of the dis-
tribution. The primary outcome mea-
sure (ie, the paretic ankle DF during
the swing phase of gait) showed a
significant within-group increase in
the median (Q1, Q3) from pretest at
6.75 (4.84, 8.99) degrees to posttest
at 12.86 (8.20, 14.15) degrees for the
track group (P⫽.008) and a trend
toward an increase from pretest at
6.61 (5.80, 10.31) degrees to post-
test at 8.99 (5.61, 10.95) degrees in
the move group (P⫽.055) (Fig. 4A).
Furthermore, whereas the between-
group difference in DF at pretest was
nonsignificant, the between-group
difference in change in DF from pre-
test to posttest was significant
(P⫽.017), favoring the track group,
Toe clearance exhibited a trend
toward a median increase in the
track group (P⫽.055), whereas the
move group showed no change
(Fig. 4B). Also important, the average
variance of toe clearance in the track
group decreased from 0.48 mm at
pretest to 0.29 mm at posttest
(P⫽.014), whereas in the move
group it decreased from 0.47 to 0.34
mm (P⫽.082). Median GTSR showed
a significant decrease toward 1.0 in
both the track (P⫽.008) and move
(P⫽.008) groups (Fig. 4C), indicat-
ing improved symmetry between
the paretic and nonparetic limbs.
Median stride length increased sig-
nificantly in both the track (P⫽.008)
and move (P⫽.016) groups (Fig. 4D).
Except for DF, as noted above,
there were no statistically significant
between-group differences.
For the 10-m walk at comfortable
speed, in the track group, the
median was 0.91 (0.53, 1.00) m/s at
pretest and 0.89 (0.71, 1.09) m/s at
posttest. In the move group, it was
0.92 (0.62, 1.14) m/s at pretest, and
0.90 (0.55, 1.15) m/s at posttest. For
the 10-m walk at maximum speed, in
the track group the median was 1.09
(0.83, 1.26) m/s at pretest and 1.06
(0.95, 1.24) m/s at posttest. In the
move group, it was 1.28 (0.74, 1.59)
m/s at pretest and 1.22 (0.72, 1.52)
m/s at posttest. All within- and

204 f Physical Therapy Volume 92 Number 2 February 2012

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014

between-group differences were
nonsignificant.
The mean (SE) AI for the track group
at pretest was 18.95% (7.35), increas-
ing to 39.92% (⫾6.80) at posttest,
and this change was significant
(P⫽.041). The move group also
improved from 34.37% (14.10) at
pretest to 53.78% (7.20) at posttest.
Although the magnitude of improve-
ment was equivalent to that of the
track group, the variance was higher
in the move group and so this
change was not significant. All of the
tracking scores are considerably less
than those reported by LaPointe et
al37 for young men (72.0% [⫾2.49])
and women (73.6% [⫾1.54]) who
were healthy using the same track-
ing protocol, which confirms the
impaired ankle function in the cur-
rent participants.
The mean (SD) number of move-
ment repetitions per trial for the
3,600 training trials in the track
group was 7.64 (1.36) compared
with 11.61 (2.80) for the move
group (P⫽.003). The peak-to-peak
amplitude of DF/PF movements aver-
aged 89.79 (0.69) degrees for the
track group and 93.62 (1.25) for the
move group (P⫽.018).
fMRI
The fMRI data for 2 participants in
the track group were excluded
because of excessive head motion.
Also, several ROIs for some partici-
pants in both groups showed no
active voxels at the specified thresh-
old for activation, resulting in
reduced sample sizes for signal inten-
sity data of 4 to 6 for the track group
and 6 to 8 for the move group.
Because of the low sample sizes,
only between-group analysis on
median change from pretest to post-
test was done on volume, percent
volume, and signal intensity.
None of the ROIs showed a signifi-
cant median change in these mea-
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
sures (eFig. 1, available at ptjournal.
apta.org). However, trends toward
between-group differences in median
change from pretest to posttest were
observed for contralesional M1 vol-
ume (P⫽.067), contralesional M1%
volume (P⫽.077), ipsilesional S1
intensity (P⫽.055), and contra-
lesional PMC volume (P⫽.067). For
each of these trends, the track group
showed decreased values from pre-
test to posttest, whereas the move
group showed increased values. A
conspicuous pattern in the change
of cortical activation from pretest to
posttest between the 2 groups is
shown in eFigure 1. Close inspection
reveals that, for the track group, the
direction of change for the median
values of the fMRI variables nearly
always showed decreases from pre-
test to posttest, whereas the move
group nearly always showed
increases. Of the 8 volume and the
8 percent volume measures (2
sides ⫻ 4 ROIs), all but one showed
a median increase in the move
group, whereas all values showed a
median decrease in the track group
(Fisher exact test, P⫽.0014). For the
8 intensity measures, 7 of 8 showed
a median increase in the move
group, whereas 7 of 8 showed a
median decrease in the track group
(Fisher exact test, P⫽.01). Figure 5
demonstrates these changes in the
cortical activation from pretest to
posttest for one participant from
each group.
Regarding laterality, the only signifi-
cant between-group difference in
the change from pretest to posttest
was for the PMC in LI%Volume, which
increased (shifted toward ipsile-
sional activation) in the track group
and decreased (shifted toward con-
tralesional activation) in the move
group (P⫽.039) (eFig. 2, available
at ptjournal.apta.org). For LIvolume,
there were no significant between-
group differences; however, PMC
demonstrated a trend (P⫽.053)
toward a difference in the same
Volume 92 Number 2
direction as LI%Volume. The LIIntensity
showed no significant changes or
trends.
Training Adherence and
Participant Feedback
Of the 19 participants who enrolled,
16 finished all the training trials,
which were confirmed by data files
automatically sent to the therapist
daily from the participant’s training
computer. Three participants, 1
from the move group and 2 from the
track group, dropped out of the
study, citing fatigue during training
as the reason. Beyond the report of
fatigue in these 3 participants, there
were no other adverse events.
Overall, participant feedback at the
completion of the study was very
favorable for telerehabilitation. The
difficulty most frequently reported
was donning the electrogoniometer;
some participants required partial
assistance from the primary care-
giver, particularly at the beginning
and less so toward the end. The tele-
conferencing connection was gener-
ally successful in allowing 2-way
video and audio communication;
however, for 4 participants the wire-
less signal was inadequate in their
homes. In these cases, telephones
were used for audio communication
and data were retrieved from the
computer’s hard drive after they fin-
ished the training.
Discussion
The important findings of this study
were that: (1) telerehabilitation to
improve ankle dorsiflexion was fea-
sible for people with chronic stroke,
and (2) complex movement training
(track group) produced greater
improvement in ankle DF during the
swing phase of gait than simple
movement training (move group).
As described further below, we
believe that the cognitive processing
during the complex movement train-
ing of the track group was responsi-
Physical Therapy f 205
Ankle Movement Training in Stroke Using Telerehabilitation

Figure 5.
Cortical activation at one coronal slice (Y⫽⫺30) and 3 transverse slices (Z⫽65, 60, and 55) for one participant in move group with
right hemisphere stroke tracking with paretic left ankle and one participant in track group with left hemisphere stroke tracking with
paretic right ankle. Change from pretest to posttest shows reduction in activation with greater focus to the ipsilesional hemisphere
in the track group participant compared with the move group participant. R⫽right, L⫽left. FDR⫽false discovery rate.

ble for the greater improvement in
the track group. However, potential
confounders did exist. Greater corti-
cal activation is associated with
higher frequency movements38 – 40
and higher amplitude movements,41
but it was impossible for us to con-
trol these 2 factors in the training
movements of the move group with-
out giving them guiding feedback.
They were instructed during orienta-
tion to produce repetitive DF/PF
movements at a comfortable rate
using their full range of motion.
However, they could have self-
selected a lower number of repeti-
tions and lower amplitude of DF/PF
movements within each training trial

206 f Physical Therapy Volume 92 Number 2 February 2012

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014

than what the training software
directed for the track group. Thus,
the greater improvement in DF dur-
ing the swing phase of gait in the
track group could have stemmed
from either their cognitive process-
ing of complex movements or
greater cortical activation associated
with greater repetitions and ampli-
tude of their DF/PF movements
within each training trial compared
with the move group. We found that
the move group produced more rep-
etitions and amplitude of training
movements per trial, indicating that
the possible confounder did not
materialize.
Another possible confounder was
the difference in duration of stroke
between groups (Table). However,
as recovery potential is generally
considered to be stronger early after
stroke rather than later, we believe it
is doubtful that the longer duration
of stroke in the track group could
account for their greater improve-
ment in dorsiflexion compared with
the move group.
We speculate that the larger increase
in DF in the track group may stem
from more focused cortical activa-
tion, involving an overall reduction
in cortical activation and shifting
toward the ipsilesional hemi-
sphere.42 In our study, although the
magnitude of volume, percent vol-
ume, and intensity values for the
track group failed to show significant
within-group decreases, the signifi-
cant finding was that the direction of
change for the median values of
these variables showed a decrease
for nearly all the ROIs for this
group, compared with the increases
observed for nearly all the ROIs in
the move group. Thus, although the
magnitude of individual fMRI
changes may have been subtle, the
direction of change (ie, decreased
activation plus a shift toward the
ipsilesional hemisphere) may be
important and is consistent with
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
other reports extolling the value of
focused cortical activation.43,44 The
mechanism by which cognitive pro-
cessing of a complex movement
task, which includes close attention
to feedback and the associated inter-
nalized reward system, stimulates
greater neuroplastic change above
physical activity alone15–18 is uncer-
tain, but studies suggest that the
increased liberation and synaptic
effects of neurotrophins during com-
bined cognitive processing and phys-
ical activity could be important.45,46
The important criterion that defines
motor learning is not improved per-
formance on the training task but
improvement on a transfer task.47
Indeed, the findings of improved dor-
siflexion during gait as well as the
improved ankle tracking in an altered
environment (inside the MR scanner)
in the track group, but not in the move
group, indicate that the telerehabili-
tation tracking training was effective
in promoting motor learning.
Our gait measurements assessed
functions critical for predicting trip-
ping potential27,28 and improvement
of energy expenditure29 –32 in people
with stroke. Tripping potential is
embedded within the ankle DF and
toe clearance data, with both exhib-
iting improvements in the track
group. Additionally, as increased toe
clearance variance can predict falls
in elderly adults,28 the significant
decrease in variance found here in
the track group suggests this group
could be safer walkers. Energy expen-
diture is embedded within the tempo-
ral symmetry of gait, with symmetric
gait patterns (ie, GTSRs approaching
1.0) being more energy efficient.30
Decreases in GTSR toward 1.0 were
observed for both the track and
move groups, suggesting decreased
energy expenditure after engaging in
both forms of training.
We did not observe improved walking
speed, which is contrary to the find-
Volume 92 Number 2
ings of Mirelman and colleagues.10,11
They showed that improved walking
speed could be accomplished in peo-
ple with stroke with clinic-based vir-
tual reality training of ankle move-
ments in all planes. Their results
suggest that for the next generation of
telerehabilitation software and hard-
ware, it may be beneficial to add
tracking training in multiple ankle
movement directions, rather than
DF/PF only, to achieve even further
gains marked by increased walking
speed. Their results also invite fur-
ther applications of virtual reality for
the lower limb, beyond that of
Holden et al5 for the upper limb, into
telerehabilitation. Finally, with our
low number of participants per
group, our study was underpow-
ered, and future studies would ben-
efit from having more participants.
We believe that telerehabilitation
offers a valuable method for promot-
ing further recovery from stroke.
Our earlier telerehabilitation study,
which also involved track versus
move groups, but using the paretic
index finger,4 did not produce a
clear advantage in the track group.
However, the training in that study
lasted only 2 weeks, compared with
4 weeks here, and so dosage of train-
ing likely is an important factor.
This finding highlights a potentially
important advantage of telerehabili-
tation. It avails extended training
time so that a large number of move-
ment repetitions, considered to be
crucial for cortical reorganization,48
can occur. Because of health care
economics and the declining length
of inpatient rehabilitation stay,1 alter-
native forms of rehabilitation, includ-
ing telerehabilitation, may become
increasingly important to approach
the dosage of training needed to pro-
mote neuroplasticity and a higher
level of recovery. In this way, we do
not envision telerehabilitation as a
replacement for the hands-on inter-
action between patients and thera-
pists so crucial for establishing trust,
Physical Therapy f 207
Ankle Movement Training in Stroke Using Telerehabilitation
motivation, and guidance of therapy
early on. Rather, it is envisioned as
therapy at a less dependent stage in
the recovery process that continues
to emphasizes repetitive movements
and cognitive processing, but with
only periodic rather than continual
therapist guidance and at a location,
intensity, and schedule selected by
the patients themselves. As the
advancement of communication tech-
nology continues to accelerate, as the
need for cost containment in health
care economics continues to deepen,
and as evidence continues to mount
that recovery of function can occur
even in the chronic stages of stroke,
telerehabilitation offers hope for con-
tinued progress.
Conclusion
These results suggest that telereha-
bilitation emphasizing complex task
training with the paretic limb can be
effective in promoting brain reor-
ganization and further recovery in
people with chronic stroke. Further
trials on a larger scale are needed to
determine the cost, effectiveness, and
the duration of retention of benefits
associated with telerehabilitation.
Dr Durfee and Dr Carey provided concept/
idea/research design. Dr Deng, Dr Durfee,
Dr Skluzacek, Ms Davey, and Dr Carey pro-
vided writing. Dr Deng, Dr Durfee, Dr Nuck-
ley, Dr Rheude, Dr Severson, Dr Skluzacek,
Dr Spindler, and Dr Carey provided data col-
lection. All authors provided data analysis. Dr
Deng, Dr Durfee, and Dr Carey provided
project management. Dr Deng and Dr Carey
provided fund procurement and institutional
liaisons. Dr Deng, Dr Durfee, Dr Nuckley,
and Dr Carey provided facilities/equipment.
Dr Skluzacek provided clerical support. Dr
Deng, Dr Durfee, Dr Nuckley, Dr Skluzacek,
and Dr Carey provided consultation (includ-
ing review of manuscript before submission).
The authors acknowledge Brett Burgstahler,
Dillon Hodapp, Dr Haitham Hussein, and
Edison Lin for their valuable contributions to
this study.
The study was approved by the University
of Minnesota’s Committee on the Use of
Human Subjects in Research.
208 f Physical Therapy Volume 92 Number 2
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
This project was funded by the following
National Institutes of Health grants:
R03HD051615, P41 RR008079, and M01-
RR00400. No US federal agencies provided
funding for this study.
This trial has been registered with Clinical
Trials.gov; registration number: NCT01298583.
DOI: 10.2522/ptj.20110018
References
1 Dobrez D, Heinemann AW, Deutsch A,
et al. Impact of Medicare’s prospective
payment system for inpatient rehabilita-
tion facilities on stroke patient out-
comes. Am J Phys Med Rehabil. 2010;
89:198 –204.
2 Granger CV, Markello SJ, Graham JE, et al.
The uniform data system for medical reha-
bilitation: report of patients with stroke
discharged from comprehensive medical
programs in 2000 –2007. Am J Phys Med
Rehabil. 2009;88:961–972.
3 Piron L, Turolla A, Agostini M, et al. Exer-
cises for paretic upper limb after stroke:
a combined virtual-reality and telemedi-
cine approach. J Rehabil Med. 2009;41:
1016 –1102.
4 Carey JR, Durfee WK, Bhatt E, et al. Com-
parison of finger tracking versus simple
movement training via telerehabilitation
to alter hand function and cortical reorga-
nization after stroke. Neurorehabil Neu-
ral Repair. 2007;21:216 –232.
5 Holden MK, Dyar TA, Dayan-Cimadoro L.
Telerehabilitation using a virtual environ-
ment improves upper extremity function
in patients with stroke. IEEE Trans Neural
Syst Rehabil Eng 2007;15:36 – 42.
6 Huber M, Rabin B, Docan C, et al. Feasibil-
ity of modified remotely monitored in-
home gaming technology for improving
hand function in adolescents with cerebral
palsy. IEEE Trans Inf Technol Biomed.
2010;14:526 –534.
7 Piron L, Tonin P, Trivello E, et al. Motor
tele-rehabilitation in post-stroke patients.
Med Informatics Internet Med. 2004;29:
119 –125.
8 Piron L, Turolla A, Tonin P, et al. Satis-
faction with care in post-stroke patients
undergoing a telerehabilitation programme
at home. J Telemed Telecare. 2008;14:
257–260.
9 Deutsch JE, Lewis JA, Burdea G. Technical
and patient performance using a virtual
reality-integrated telerehabilitation sys-
tem: preliminary finding. IEEE Trans Neu-
ral Syst Rehabil Eng. 2007;15:30 –35.
10 Mirelman A, Bonato P, Deutsch JE. Effects
of training with a robot-virtual reality sys-
tem compared with a robot alone on the
gait of individuals after stroke. Stroke.
2009;40:169 –174.
11 Mirelman A, Patritti BL, Bonato P, Deutsch
JE. Effects of virtual reality training on gait
biomechanics of individuals post-stroke.
Gait Posture. 2010;31:433– 437.
12 Kesar TM, Perumal R, Jancosko A, et al.
Novel patterns of functional electrical
stimulation have an immediate effect on
dorsiflexor muscle function during gait for
people poststroke. Phys Ther. 2010;90:
55– 66.
13 Said CM, Goldie PA, Patla AE, Sparrow
WA. Effect of stroke on step characteris-
tics of obstacle crossing. Arch Phys Med
Rehabil. 2001;82:1712–1719.
14 Waters RL, Mulroy S. The energy expendi-
ture of normal and pathologic gait. Gait
Posture. 1999;9:207–231.
15 Kleim JA, Barbay S, Cooper NR, et al.
Motor learning-dependent synaptogenesis
is localized to functionally reorganized
motor cortex. Neurobiol Learn Mem.
2002;77:63–77.
16 Nudo R, Milliken G, Jenkins W, Merzenich
M. Use-dependent alterations of move-
ment representations in primary motor
cortex of adult squirrel monkeys. J Neuro-
sci. 1996;16:785– 807.
17 Plautz EJ, Milliken GW, Nudo RJ. Effects of
repetitive motor training on movement
representations in adult squirrel monkeys:
role of use versus learning. Neurobiol
Learn Mem. 2000;74:27–55.
18 Pascual-Leone A, Nguyen KT, Cohen AD,
et al. Modulation of muscle responses
evoked by transcranial magnetic stimula-
tion during the acquisition of new fine
motor skills. J Neurophysiol. 1995;74:
1037–1045.
19 Doyon J, Korman M, Morin A, et al. Con-
tribution of night and day sleep vs simple
passage of time to the consolidation of
motor sequence and visuomotor adapta-
tion learning. Exp Brain Res. 2009;195:
15–26.
20 Stefan K, Wycislo M, Classen J. Modulation
of associative human motor cortical plas-
ticity by attention. J Neurophysiol. 2004;
92:66 –72.
21 Georgopoulos AP, Lurito JT, Petrides M,
et al. Mental rotation of the neuronal pop-
ulation vector. Science. 1989;243:234 –
236.
22 Schmidt RA, Lee TD. Motor Control and
Learning: A Behavioral Emphasis. Cham-
paign, IL: Human Kinetics Inc; 2005.
23 Carey JR. Manual stretch: Effect on finger
movement control and force control in
stroke subjects with spastic extrinsic fin-
ger flexor muscles. Arch Phys Med Reha-
bil. 1990;71:888 – 894.
24 Anderson DI, Magill RA, Sekiya H. Motor
learning as a function of KR schedule and
characteristics of task-intrinsic feedback. J
Motor Behav. 2001;33:59.
25 Kadaba MP, Ramakrishnan HK, Wootten
ME. Measurement of lower extremity kine-
matics during level walking. J Orthop Res.
1990;8:383–392.
26 Wu G, Siegler S, Allard P, et al; for the
International Society of Biomechanics. ISB
recommendation on definitions of joint
coordinate system of various joints for the
reporting of human joint motion, part I:
ankle, hip, and spine. J Biomech. 2002;35:
543–548.
February 2012

27 Begg R, Best R, Dell’Oro L, Taylor S. Min-
imum foot clearance during walking: strat-
egies for the minimisation of trip-related
falls. Gait Posture. 2007;25:191–198.
28 Mills PM, Barrett RS, Morrison S. Toe clear-
ance variability during walking in young
and elderly men. Gait Posture. 2008;28:
101–107.
29 Rotenberg A, Pascual-Leone A. Safety of 1
Hz repetitive transcranial magnetic stim-
ulation (rTMS) in patients with titanium
skull plates. Clin Neurophys. 2009;120:
1417.
30 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.
31 Oken O, Yavuzer G, Ergocen S, et al.
Repeatability and variation of quantitative
gait data in subgroups of patients with
stroke. Gait Posture. 2008;27:506 –511.
32 Patterson KK, Gage WH, Brooks D, et al.
Evaluation of gait symmetry after stroke: a
comparison of current methods and rec-
ommendations for standardization. Gait
Posture. 2010;31:241–246.
33 Talairach J, Tournoux P. Co-planar Stereo-
taxic Atlas of the Human Brain. New York,
NY: Thieme Medical Publishers; 1988.
34 Dassonville P, Lewis SM, Zhu XH, et al.
The effect of stimulus-response compati-
bility on cortical motor activation. Neuro-
image. 2001;13:1–14.
35 Genovese CR, Lazar NA, Nichols T.
Thresholding of statistical maps in func-
tional neuroimaging using the false discov-
ery rate. Neuroimage. 2002;15:870 – 878.
February 2012
Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014
Ankle Movement Training in Stroke Using Telerehabilitation
36 Cramer SC, Nelles G, Benson RR, et al. A
functional MRI study of subjects recov-
ered from hemiparetic stroke. Stroke.
1997;28:2518 –2527.
37 LaPointe KE, Klein JA, Konkol ML, et al.
Cortical activation during finger tracking
vs ankle tracking in healthy subjects. Restor
Neurol Neurosci. 2009;27:253–264.
38 Rao SM, Bandettini PA, Binder JR, et al.
Relationship between finger movement
rate and functional magnetic resonance
signal change in human primary motor
cortex. J Cereb Blood Flow Metab. 1996;
16:1250 –1254.
39 Sadato N, Ibanez V, Campbell G, et al.
Frequency-dependent changes of regional
cerebral blood flow during finger move-
ments: functional MRI compared to PET.
J Cereb Blood Flow Metab. 1997;17:670 –
679.
40 Jenkins IH, Passingham RE, Brooks DJ. The
effect of movement frequency on cere-
bral activation: a positron emission tomog-
raphy study. J Neurol Sci. 1997;151:
195–205.
41 Waldvogel D, van Gelderen P, Ishii K, Hal-
lett M. The effect of movement amplitude
on activation in functional magnetic reso-
nance imaging studies. J Cereb Blood Flow
Metab. 1999;19:1209 –1212.
42 Feydy A, Carlier R, Roby-Brami A, et al.
Longitudinal study of motor recovery after
stroke: recruitment and focusing of brain
activation. Stroke. 2002;33:1610 –1617.
43 Ward NS, Brown MM, Thompson AJ,
Frackowiak RS. Neural correlates of motor
recovery after stroke: a longitudinal fMRI
study. Brain. 2003;126:2476 –2496.
Volume 92 Number 2
44 Muellbacher W, Richards C, Ziemann U,
et al. Improving hand function in chronic
stroke. Arch Neurol. 2002;59:1278 –1282.
45 Klintsova AY, Dickson E, Yoshida R, Gree-
nough WT. Altered expression of BDNF
and its high-affinity receptor TrkB in
response to complex motor learning and
moderate exercise. Brain Res. 2004;1028:
92–104.
46 Gomez-Pinilla F, So V, Kesslak JP. Spatial
learning and physical activity contribute
to the induction of fibroblast growth fac-
tor: neural substrates for increased cogni-
tion associated with exercise. Neurosci-
ence. 1998;85:53– 61.
47 Schmidt RA, Lee TD. Motor Control and
Learning: A Behavioral Emphasis. 3rd
ed. Champaign, IL: Human Kinetics Inc;
1999.
48 Kimberley TJ, Samargia S, Moore LG, et al.
Comparison of amounts and types of prac-
tice during rehabilitation for traumatic
brain injury and stroke. J Rehabil Res Dev.
2010;47:851– 862.
49 Coren S. The lateral preference inventory
for measurement of handedness, footed-
ness, eyedness, and earedness: norms for
young adults. Bull Psychonomic Soc.
1993;31:1–1.
50 Folstein MF, Folstein SE, McHugh PR.
“Mini-mental state”: a practical method for
grading the cognitive state of patients for
the clinician. J Psychiatr Res. 1975;12:
189 –198.
Physical Therapy f 209
 Complex Versus Simple Ankle Movement Training in
Stroke Using Telerehabilitation: A Randomized
Controlled Trial
Huiqiong Deng, William K. Durfee, David J. Nuckley,
Brandon S. Rheude, Amy E. Severson, Katie M.
Skluzacek, Kristen K. Spindler, Cynthia S. Davey and
James R. Carey
PHYS THER. 2012; 92:197-209.
Originally published online November 17, 2011
doi: 10.2522/ptj.20110018

References This article cites 47 articles, 11 of which you can access

for free at:
http://ptjournal.apta.org/content/92/2/197#BIBL
Cited by This article has been cited by 1 HighWire-hosted articles:

http://ptjournal.apta.org/content/92/2/197#otherarticles
Subscription http://ptjournal.apta.org/subscriptions/
Information
Permissions and Reprints http://ptjournal.apta.org/site/misc/terms.xhtml
Information for Authors http://ptjournal.apta.org/site/misc/ifora.xhtml

Downloaded from http://ptjournal.apta.org/ by guest on November 10, 2014