RESEARCH ARTICLE

Effects of Brain–Computer Interface-controlled

Functional Electrical Stimulation Training on Shoulder
Subluxation for Patients with Stroke: A Randomized
Controlled Trial
Yun Young Jang1, Tae Hoon Kim1,2 & Byoung Hee Lee1*
1
Graduate School of Physical Therapy, Sahmyook University, Seoul, Korea
2
The Post-Professional Doctor of Physical Therapy Program, Stockton University, Galloway, NJ USA

Abstract
The purpose of this study was to investigate the effects of brain–computer interface (BCI)-controlled functional
electrical stimulation (FES) training on shoulder subluxation of patients with stroke. Twenty subjects were ran-
domly divided into two groups: the BCI-FES group (n = 10) and the FES group (n = 10). Patients in the BCI-FES
group were administered conventional therapy with the BCI-FES on the shoulder subluxation area of the paretic
upper extremity, five times per week during 6 weeks, while the FES group received conventional therapy with FES
only. All patients were assessed for shoulder subluxation (vertical distance, VD; horizontal distance, HD), pain (visual
analogue scale, VAS) and the Manual Function Test (MFT) at the time of recruitment to the study and after 6 weeks of
the intervention. The BCI-FES group demonstrated significant improvements in VD, HD, VAS and MFT after the
intervention period, while the FES group demonstrated significant improvements in HD, VAS and MFT. There were
also significant differences in the VD and two items (shoulder flexion and abduction) of the MFT between the two
groups. The results of this study suggest that BCI-FES training may be effective in improving shoulder subluxation
of patients with stroke by facilitating motor recovery. Copyright © 2016 John Wiley & Sons, Ltd.

Received 10 September 2015; Revised 30 November 2015; Accepted 30 November 2015

Keywords
brain–computer interface; rehabilitation; shoulder subluxation; stroke

*Correspondence
Dr Byoung Hee Lee, PT, PhD, Department of Physical Therapy, Graduate School of Physical Therapy, Sahmyook University, 815 Hwarang-ro,
Nowon-gu, Seoul 139-742, Korea.
†
Email: 3679@syu.ac.kr

Published online 15 February 2016 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/oti.1422

Introduction various applications have been reported to improve its

Glenohumeral joint subluxation of stroke patients has
been recognized as a major and frequent complication
and may be associated with shoulder pain, peripheral
nerve damage and autonomic dysfunction, resulting
in interference with functional activities (Andersen,
1985; Dursun et al., 2000). As conventional treatments,
complications, such as an electrical stimulation (Linn
et al., 1999; Ada and Foongchomcheay, 2002), arm
sling (Acar and Karatas, 2010) or a robotic treatment
(Dohle et al., 2013).
Functional electrical stimulation (FES) has been
used to stimulate motor function by activating nerves.

Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd. 175
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors Jang et al.

Many researchers reported the effects of FES on facili- Participants

tating normal movement in patients with paralysis
Thirty patients who were diagnosed with hemiparetic
due to upper motor neuron diseases such as stroke,
stroke participated in this study. The inclusion
multiple sclerosis and spinal cord injury (Rushton,
criteria were as follows: (1) had onset within
2003). Also, it has been studied that it is not only its
6 months post-stroke; (2) diagnosed with shoulder
peripheral mechanisms that stimulate the patient’s
subluxation by X-ray finding; (3) scored more than
remaining motor units to enhance muscular strength,
24 points on the Mini-Mental State Examination
increase range of motion (ROM) and reduce stiffness
(MMSE); and (4) had unimpaired visual and
(Granat et al., 1993; Sabut et al., 2011; Lo et al., 2012;
somatosensory function. Patients with a cardiac
Sahin et al., 2012; Dirks et al., 2014) but also the central
pacemaker, a continuing neurological deficit due to
mechanisms that occur to help control movement by a
past contralateral stroke or a shoulder pathology
re-organization of the cortex based on neurophysiolog-
not related to the stroke were excluded from the
ical responses (Wassermann, 1995). On the basis of
study. After baseline measures, 7 out of 30 patients
these mechanisms, the effect of FES on shoulder sub-
were excluded by low MMSE score and long onset
luxation has been reported (Faghri et al., 1994;
date (>6 months), and 23 patients fulfilled the selec-
Koyuncu et al., 2010). However, individuals applied with
tion criteria. All patients were provided with a full
FES treatment have not voluntarily attempted the con-
explanation of the experimental procedure and pro-
traction of involved muscles; therefore, there is a limita-
vided written consent signifying voluntary participa-
tion for motor re-education (Vafadar et al., 2015).
tion. The Sahmyook University Human Studies
Recently, a brain–computer interface (BCI) has been
Committee approved this study. The patients were
proposed in the fields of rehabilitation as a form of
randomly divided into the BCI-FES group (n = 11)
neuro-feedback (Silvoni et al., 2011). The BCI measures
and the FES group (n = 12). Randomization was
neuronal signals by using an electroencephalogram
conducted using a computer-based 1:1 randomiza-
(EEG), magneto-encephalography or real-time magnetic
tion program by a researcher who was not involved
resonance imaging and processes the information
in participant recruitment. There were no differ-
through amplification and classifies it by an algorithm
ences between the groups in primary outcome mea-
(Birbaumer et al., 2008). In the BCI system, such classi-
sures such as shoulder subluxation (vertical distance,
fied information stimulates the motor system through
VD; horizontal distance, HD), visual analogue scale
biofeedback equipment such as FES or robotic assist de-
(VAS), Modified Ashworth Scale (MAS) and Manual
vices (Do et al., 2011). Therefore, the BCI system can
Function Test (MFT) before training began (independent
make up for the limitation of FES treatment by inducing
t-tests, all p > 0.05).
a subject’s active participation for various tasks. To our
knowledge, however, no study has investigated the effi-
cacy of the BCI-controlled FES on shoulder subluxation
Experimental protocols
in stroke patients.
Therefore, the objective of this study was to investi- The patients in both groups participated in the
gate the effects of the BCI-controlled FES on structural scheduled programmes by two trained occupational
and functional improvements of shoulder subluxation therapists to minimize errors. The patients in the
in patients after stroke. BCI-FES group were administered 20 minutes of
FES under the BCI programme with conventional
Material and methods occupational therapy and required to complete 30
sessions during 6 weeks, while those in the FES
Design
group received FES with conventional occupational
This study was a single-blind, randomized clinical trial therapy for the same time. Twenty-three patients
that included 20 people who had shoulder subluxation who fulfilled the selection criteria participated in
after stroke and were randomly allocated to one of two the study, and 20 patients completed the follow-up
groups (the BCI-FES group or the FES group). Out- assessment as shown in Figure 1. The main reasons
come measurements were obtained before and after for three dropouts from the study were a discharge
the intervention period. before the intervention period and a personal issue.

176 Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
Jang et al.
Figure 1. Flow of participants through the training programme. BCI, brain–computer interface; FES, functional electrical stimulation;
MMSE, Mini-Mental State Exam
The brain–computer interface-controlled
functional electrical simulation training
The BCI equipment consisted of EEG sensors
(brainwave-sensing headset, NeuroSky, San Jose, CA,
USA, 2101) to receive brainwave information, a moni-
tor screen, a laptop to record and process brainwave
activity, a USB output board to link the brain waves
to FES when the brain activity was high through con-
centration and FES (EMG FES 1000-Walking Man,
Cybermedic Inc., Iksan, Korea, 2009).
Patients in the BCI-FES group were provided with pri-
vate space without external stimuli that influence perfor-
mance and brain activity. Patients were seated on a
comfortable chair with their hands and forearms resting
on a desk. A trained examiner explained the tasks
through simple instructions. Patients were provided with
action observational tasks related to shoulder move-
ments by watching a computer monitor screen, and they
performed a sequence of movements with their own arm
including (1) shoulder abduction without weight, (2)
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
shoulder adduction without weight, (3) shoulder abduc-
tion with 500 g of weight and (4) shoulder adduction
with 500 g of weight. The patients were advised to keep
focusing on tasks.
Electroencephalogram patterns underlying shoulder
motions were detected in real time. The level of concen-
tration on tasks was analysed using the concentration
index. In a state of concentration, the sensory motor
rhythm (SMR) and mid-beta rhythms increase while
the theta rhythm decreases. Increased SMR indicates
unfocused attention, whereas mid-beta rhythms imply
focused attention. SMR was in the range of 12–15 Hz,
and mid-beta rhythm was in the range of 16–20 Hz.
SMR was observed when a subject focused on solving
problems without being nervous or stressed. The con-
centration index was quantified by a relative ratio involv-
ing SMR, mid-beta and theta rhythms as expressed by
the formula (SMR + mid-beta)/theta. EEG patterns
during concentration were subsequently used to trigger
FES on the supraspinatus and posterior deltoid muscle
177
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
so as to achieve shoulder reduction. Detailed parameters of
the FES were 15-second stimulation time with a 7-second
rest, frequency 35 Hz, pulse width 150 microseconds, rise
time 0.5 seconds, square wave and fall time 0.5 seconds.
The intensity was adjusted from 1 to 50 mA according to
the response from the patients’ shoulder joint. To calibrate
the focused threshold, focused observation was imple-
mented 10 times before the training to build an average
threshold. The concentration index threshold was input
into the computer algorithm, and then the patients were
instructed to focus on the movement of the shoulder on
the monitor screen. When the measured concentration
index surpassed the threshold of the concentration index,
this information was transferred to the USB output board
to activate the FES, whereas when the measured
concentration index did not exceed the threshold, this
information was transferred to turn off the FES. Therefore,
FES activation occurred based upon the degree of
concentration of the patient (Figure 2).
The functional electrical simulation training
The FES group was provided with FES application for
20 minutes. This equipment has an adjustable frequency,
Figure 2. Diagramming BCI-controlled FES for real-time feedback. BCI, brain–computer interface; FES, functional electrical stimulation;
EEG, electroencephalogram
178
Jang et al.
contraction and relaxation times, pulse width, a pair of
surface electrodes measuring 50 × 50 mm, and a stimula-
tor. Detailed parameters were under identical conditions
with those of the BCI-FES group. In order to avoid muscle
fatigue due to electrical stimulation, the off time was set to
last 7 seconds, pulse frequency was 35 Hz, pulse width was
150 microseconds and current was 1–50 mA.
Conventional therapy
Conventional therapy was supervised by trained ther-
apists in the hospital. Both groups were engaged in a for-
mal occupational therapy protocol consisting of an ROM
exercise, strengthening exercise and hand and finger ex-
ercises for 30 minutes day 1 during the same period of
time as the intervention. To minimize the therapeutic ef-
fects from different therapists, subjects had instruction
seven times before and during the intervention period
to strengthen adherence to the designated protocol.
Outcome measures
The data were measured by the same blinded evaluator
before training began and at the end of the 6-week
training period.
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
Jang et al.
Shoulder subluxation
Shoulder subluxation was evaluated with pre-
rehabilitation and post-rehabilitation anteroposterior
(A-P) shoulder X-rays using the classification devel-
oped by Park et al. (2007). Radiographs of the
affected and unaffected shoulders were taken in an
A-P plane with the patient seated and arms depen-
dent in a neutral position. Three reference points
were initially identified: the central points (the
glenoid fossa and the humeral head) and the most
inferolateral point (the acromial surface). The VD
of the glenohumeral joint was measured from the
inferior acromial point to the central point of the
humeral head, and the HD was measured between
the central points (Figure 3). All measurements were
independently performed by the same investigator,
and X-ray was not taken until at least 24 hours after
the intervention to eliminate any possible short-
term effects.
Pain intensity
The 10-cm VAS was used to assess pain intensity of
subluxation in the shoulder (Bijur et al., 2001). The
patients were instructed to mark their subjective pain in-
tensity through the line at the appropriate represented
point on the scale.
Figure 3. Measurement of shoulder subluxation
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
Spasticity
The MAS was used to assess shoulder spasticity
(Bohannon and Smith, 1987; Gregson et al., 1999),
which is a 6-point scale grading the resistance of a re-
laxed limb to rapid passive stretch, with grades of 0
(no increase in muscle tone), 1 (slight increase in muscle
tone, manifested by a catch and release, or by minimal
resistance at the end of ROM), 2 (slight increase in muscle
tone manifested by a catch, followed by minimal resis-
tance throughout less than half of the ROM), 3 (more
marked increase in muscle tone through most of the
ROM, but the affected part is still easily moved), 4 (con-
siderable increase in muscle tone, passive movement is dif-
ficult through range in time indicated) and 5 (affected
part is rigid in flexion or extension). The examiner eval-
uated spasticity at the affected shoulder by performing
passive abduction and adduction of shoulder joint with
the patient in a sitting position.
Upper-limb function test
As an upper-limb function test, we used the MFT
developed to evaluate unilateral manual performance
in hemiparetic patients, which consisted of eight items
including shoulder flexion, abduction, touching the oc-
ciput with the palm, touching the back with the palm,
grasping, pinching, carrying of cubes and pegboard
179
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors Jang et al.

manipulation. MFT is graded on a four-level scale. The
total MFT score can range from 0 (severely impaired) to
32 (full function) (Miyamoto et al., 2009). The exam-
iner provided patients with verbal instructions and
demonstration before testing. All patients tried three
times and were assigned the best score in each item.
Statistical analysis
SPSS version 18.0 (IBM Corp., Armonk, NY, USA) was
used for all statistical analyses. Descriptive statistics
were used to analyse the demographic variables at base-
line, and the Shapiro–Wilk test was used in determin-
ing the distribution of the general properties and
outcome measures of the subjects. The independent
t-test was performed for comparison of VD, HD, VAS
and MFT between the two groups. The Mann–Whitney
test was performed to compare the data of the MAS
between the two groups. To detect changes in the
pre-test and post-test results within each group, the
paired t-test was used for comparison of VD, HD,
VAS and MFT, and Wilcoxon’s signed-rank test was
performed for the MAS. Results were accepted as statis-
tically significant at p < 0.05.
Results
Baseline characteristics of the 20 patients who com-
pleted the experiment are shown in Table 1. There were
no significant differences in age, height, weight,
affected side and time since onset between the groups
at baseline.
The results of VD, HD, VAS, MAS and MFT within
group comparisons are shown in Tables 2 and 3. When
comparing the differences in outcomes before and af-
ter the intervention within each group, the BCI-FES
group showed statistically significant changes in VD
(p < 0.001), HD (p = 0.002), VAS (p = 0.009) and
MFT (p < 0.000) after the intervention period, while
the FES group showed significant changes in HD
(p = 0.034), VAS (p = 0.003) and MFT (p = 0.007).
There were significant differences between groups
on the three outcome measures, VD (p = 0.004) and
two items of the MFT (shoulder flexion, p = 0.029,
and shoulder abduction, p = 0.016), compared with
the FES group. There was a statistically significant re-
duction in pain VAS within both groups, but not be-
tween groups. For MAS, there were no significant
differences within or between groups.
Discussion
Shoulder subluxation is a frequent complication in pa-
tients with post-stroke hemiplegia and may lead to
sympathetic reflex dystrophy, rotator cuff injury, adhe-
sive capsulitis, tendonitis or rupture of ligaments (Paci
et al., 2005). Therefore, the present study evaluated the
effects of BCI-controlled FES on shoulder subluxation
of the paretic upper extremity in patients with stroke.
We found that the BCI-controlled FES training helped
reduce vertical subluxation and improve shoulder flex-
ion and abduction more than in the FES-alone group.
In the literature of hemiplegic shoulder, FES has
been reported as an effective neuromuscular approach

Table 1. IDemographic data of groups (n = 20)

2
BCI-FES group (n = 10) FES group (n = 10) χ /t(p)

Sex Male 6 (60%) 4 (40%) 0.80 (0.371)

Female 4 (40%) 6 (60%)
Affected side Right 5 (50%) 8 (80%) 1.978 (0.160)
Left 5 (50%) 2 (20%)
Reason Haemorrhage 4 (40%) 4 (40%) 0.00 (1.000)
Infarction 6 (60%) 6 (60%)
Age (year) 61.10 (13.77) 61.70 (12.09) 0.104 (0.919)
Height (cm) 166.00 (5.58) 162.30 (8.62) 1.140 (0.272)
Weight (kg) 65.10 (7.91) 67.40 (6.42) 0.714 (0.485)
Time since onset (month) 4.40 (0.97) 4.10 (0.74) 0.780 (0.446)

Note.
1
n (%).
2
Mean (SD).
MMSE = Mini-Mental State Exam – Korean; BCI = brain–computer interface; FES = functional electrical stimulation.

180 Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
Jang et al. BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors

Table 2. IIComparison of shoulder subluxation, pain intensity and spasticity within groups and between groups (n = 20)

Parameters BCI-FES group (n = 10) FES group (n = 10) t(p)

X-ray VD (mm) Before 46.57 (7.06) 46.17 (9.07)

After 37.66 (6.32) 44.86 (10.32)
Before–after 8.90 (4.91) 1.31 (5.35) 3.308 (0.004)
t(p) 5.729 (0.000) 0.773 (0.459)
X-ray HD (mm) Before 26.87 (3.16) 23.98 (3.51)
After 21.68 (4.01) 20.59 (4.11)
Before–after 5.19 (3.67) 3.39 (4.29) 1.010 (0.326)
t(p) 4.467 (0.002) 2.500 (0.034)
VAS (cm) Before 4.40 (3.50) 4.20 (3.16)
After 2.00 (1.70) 3.00 (2.40)
Before–after 2.40 (2.27) 1.20 (0.92) 1.549 (0.148)
t(p) 3.343 (0.009) 4.129 (0.003)
MAS Before 0.60 (0.70) 0.60 (0.70)
After 1.00 (0.82) 1.10 (0.74)
Before–after 0.40 (0.70) 0.50 (0.71) 1.643 (0.100)
t(p) 1.633 (0.102) 1.890 (0.059)

Values are means (SD).

VD = vertical distance; HD = horizontal distance; VAS = visual analogue scale; MAS = Modified Ashworth Scale; BCI = brain–computer interface;
FES = functional electrical stimulation.

and has shown beneficial effects on subluxation, pain
and mobility (Faghri et al., 1994; Chantraine et al.,
1999; Vuagnat and Chantraine, 2003). Both groups
given FES in this study showed improvements in shoul-
der reduction, pain and functional mobility. These
results are in tune with previous findings by Koyuncu
et al. (2010) who investigated the effect of FES for
shoulder subluxation and pain in hemiplegic patients.
They randomly divided 50 hemiplegic patients with
shoulder subluxation into the study and control groups
and measured the shoulder subluxation levels with
Van Langenberghe’s classification, VAS during resting,
passive ROM and active ROM before and after FES appli-
cation on supraspinatus and posterior deltoid muscles.
They found significant changes in shoulder subluxation
in favour of the FES group and concluded that applying
FES treatment in addition to conventional treatment
might be more beneficial than conventional treatment by
itself. However, all studies for FES on shoulder subluxation
have a limitation that subjects remained passive and did
not perform any activity or receive any other form of treat-
ment while FES was administered (Vafadar et al., 2015).
Therefore, our study demonstrated the efficacy of BCI-
controlled FES to compensate for this limitation of FES.
Recently, BCI-controlled FES systems have been in-
creasingly recommended as potential neuro-rehabilitation
interventions for improving impaired upper-extremity
function in individuals with stroke (Daly et al., 2009).
The BCI-controlled FES system enables the direct brain
control of paralytic muscles and allows the user to control
stimulation settings (Gollee et al., 2010). In this study,
when compared with the FES group, the BCI-FES group
showed significant improvements in VD of shoulder lux-
ation measurement and two abilities (shoulder flexion
and abduction) in the MFT score. To our knowledge, no
previous study has evaluated the impact of BCI-FES train-
ing on hemiplegic shoulder subluxation. Although no
literature compares with the outcome of our study for
subluxation, there is limited published evidence to com-
pare for the improvements of upper-limb function. Daly
et al. (2009) determined the feasibility of BCI-FES for mo-
tor learning after stroke. In their study, the participant
who was 10 months post-stroke was unable to produce
isolated movement of any of the digits of her involved
hand. With the intervention, brain signals from the le-
sioned hemisphere were used to trigger FES for finger
movement practice. The BCI-FES consisted of trials of ei-
ther attempted finger movement with relaxed conditions
or imagined finger movement with relaxed conditions.
The participant achieved a highly accurate brain signal
control to trigger the FES device for isolated finger move-
ments. Cincotti et al. (2012) proposed the clinical applica-
tion of a BCI-based rehabilitation device to promote
motor recovery after stroke and reported that a BCI-based
device such as FES provides the clinicians with a monitor-
ing instrument to assess the patient’s rehabilitative

Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd. 181
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors Jang et al.

Table 3. IIIComparison of MFT upper-limb function within groups and between groups (n = 20)

Parameters BCI-FES group (n = 10) FES group (n = 10) t(p)

Shoulder flexion Before 1.60 (1.65) 1.60 (1.71)

After 3.80 (0.42) 2.40 (1.27)
Before–after 2.20 (1.55) 0.80 (1.03) 2.378 (0.029)
t(p) 4.491 (0.002) 2.449 (0.037)
Shoulder abduction Before 1.70 (1.49) 1.90 (1.52)
After 3.80 (4.42) 2.60 (1.35)
Before–after 2.10 (1.29) 0.70 (1.06) 2.656 (0.016)
t(p) 5.161 (0.001) 2.090 (0.066)
Touching the occiput with the palm Before 1.60 (1.43) 1.40 (1.35)
After 3.20 (1.23) 2.30 (1.06)
Before–after 1.60 (1.27) 0.90 (0.99) 1.376 (0.186)
t(p) 4.000 (0.003) 2.862 (0.019)
Touching the back with the palm Before 1.00 (1.63) 1.10 (1.66)
After 1.50 (1.58) 1.80 (1.40)
Before–after 0.50 (0.71) 0.80 (0.92) 0.818 (0.424)
t(p) 2.236 (0.052) 2.090 (0.066)
Grasping Before 0.80 (1.32) 0.90 (1.45)
After 1.50 (1.58) 1.10 (1.37)
Before–after 0.70 (1.25) 0.00 (0.82) 1.481 (0.156)
t(p) 1.769 (0.111) 1.500 (0.168)
Pinching Before 0.50 (1.08) 0.70 (1.25)
After 0.80 (1.23) 1.20 (1.32)
Before–after 0.30 (0.48) 0.50 (0.97) 0.583 (0.567)
t(p) 1.964 (0.081) 1.627 (0.138)
Carrying of cubes Before 0.50 (1.27) 0.40 (0.97)
After 0.70 (1.16) 0.90 (1.29)
Before–after 0.30 (0.48) 0.50 (0.97) 0.818 (0.424)
t(p) 0.802 (0.443) 1.861 (0.096)
Pegboard manipulation Before 0.40 (0.97) 0.20 (0.63)
After 0.60 (0.97) 0.80 (1.14)
Before–after 0.20 (0.42) 0.60 (0.84) 1.342 (0.202)
t(p) 1.500 (0.168) 2.250 (0.051)
Total MFT score (0–32) Before 8.10 (9.19) 8.20 (9.68)
After 15.90 (6.85) 13.10 (8.32)
Before–after 7.80 (4.26) 4.90 (4.48) 1.482 (0.156)
t(p) 5.785 (0.000) 3.456 (0.007)

Values are means (SD). The score for each test can vary from 0 (lowest possible score) to 4 (highest possible). The total MFT score can range from
0 (severely impaired) to 32 (full function).
MFT = Manual Function Test; BCI = brain–computer interface; FES = functional electrical stimulation.

cognitive activities and assists the patient in the practice of
motor imagery (MI). They also found that the rehabilita-
tion training based on BCI-FES-mediated feedback mech-
anisms can facilitate a better engagement of motor areas
and can promote neuroplasticity in brain regions affected
by stroke.
Their finding is similar to that of Prasad’s (2010)
study, in which the BCI system was used to provide a
computer game-based neuro-feedback to stroke pa-
tients during the MI and measured the upper-limb
functional recovery using the action research arm test
and grip strength. They concluded that the BCI is a
useful intervention in a post-stroke rehabilitation
protocol combining both physical practice and MI
practice of rehabilitation tasks. In the present study,
the BCI-FES group concentrated on the shoulder
movements on the DVDs as the MI and followed the
movements as physical practice. As subjects concen-
trated on the shoulder movement, BCI-FES neuro-
feedback enhanced movements of shoulder flexion
and abduction by stimulating the supraspinatus and
posterior deltoid muscle relating to the vertical reduc-
tion (Faghri et al., 1994). Therefore, such neuro-
feedback in the BCI-FES group might have the more

182 Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
Jang et al.
positive impact on shoulder vertical reduction with
shoulder flexion and abduction than the control
group. Based on the results, we hypothesize that the
application of the BCI-controlled FES system may
have additional therapeutic effects on shoulder
reduction by improving shoulder active movement.
Possible mechanisms involved in the effect of the
BCI system could include the activation of the brain
through learning mechanisms involving MI. MI pro-
vided by task monitoring can be used to generate dis-
crete patterns of cortical activation in the EEG, thus
controlling the BCI system. As discrete patterns, an
SMR was classified according to the BCI control per-
formance. In a functional magnetic resonance imaging
study by Halder et al. (2011), when subjects performed
an MI, motor observation and motor execution task,
high-aptitude BCI users achieved significantly higher
activation of the supplementary motor areas for the
MI and the motor observation tasks. Furthermore,
Hermes et al. (2011), who investigated the brain areas
involved in MI, found that the primary motor area
may be more reliably activated during MI and the
pre-motor cortex may be a better area to implant a
BCI system.
In the present study, we found that the use of the
BCI-FES system had a positive influence on shoulder
reduction and active movements. These results are in
line with previous studies that have demonstrated the
effects of BCIs on motor recovery from stroke by stim-
ulating brain activation (Daly et al., 2009; Gollee et al.,
2010; Prasad et al., 2010; Cincotti et al., 2012). How-
ever, this is the first BCI research to investigate its effi-
cacy for hemiplegic shoulder subluxation and broaden
the application domain of the BCI to include the most
prevalent of neurological injuries, stroke. Based on
these results, we suggest that the BCI-controlled FES
system may have a greater beneficial effect on reduction
of shoulder subluxation than FES alone. Because of
the benefits of shoulder reduction and improved ac-
tive ROM, the use of BCI-FES system in the field of
occupational therapy may facilitate functional use in
one’s affected upper extremity. These improvements
could give rise to increased engagement in meaningful
activities requiring the use of the affected limb such as
self-care, leisure or work pursuits. Also, when a spe-
cific patient is being considered, the occupational
therapists can take into consideration the patient’s
preferences by modifying the action observational
programme.
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
This study has certain limitations. First, only partic-
ipants who had a certain level of upper-limb function
were examined. Future studies should consider broader
populations according to upper-limb function. Second,
follow-up results in the experiment group were not in-
vestigated; thus, it is unclear whether the effects will be
maintained for a sufficient period of time to produce
lasting improvements in shoulder reduction. Third,
the sample size was small; some differences in outcome
may not have been statistically detectable, and others
may have been fortuitous.
Conclusions
Our results indicate that BCI-FES training with con-
ventional therapy may be effective in enhancing
improvement in shoulder subluxation of patients with
stroke by facilitating motor recovery. In addition,
information on the details underlying the beneficial
effects of BCI on arm performance after stroke may
stimulate the development of innovative behavioural
concepts in rehabilitation. Future studies, with longer
intervention periods and larger samples, may result in
meaningful improvements by determining the addi-
tional contributions of BCI in improving shoulder sub-
luxation and impaired motor function due to stroke.
Conflicts of interest
The authors declare no conflict of interest.
REFERENCES
Acar M, Karatas GK (2010). The effect of arm sling on
balance in patients with hemiplegia. Gait and Posture
32: 641–644.
Ada L, Foongchomcheay A (2002). Efficacy of electrical
stimulation in preventing or reducing subluxation of the
shoulder after stroke: a meta-analysis. The Australian
Journal of Physiotherapy 48(4): 257–267.
Andersen LT (1985). Shoulder pain in hemiplegia. The
American Journal of Occupational Therapy 39(1): 11–19.
Bijur PE, Silver W, Gallagher EJ (2001). Reliability of the vi-
sual analog scale for measurement of acute pain. Academic
Emergency Medicine 8(12): 1153–1157.
Birbaumer N, Murguialday AR, Cohen L (2008). Brain–
computer interface in paralysis. Current Opinion in
Neurology 21: 634–638.
Bohannon RW, Smith MB (1987). Interrater reliability of
a modified Ashworth scale of muscle spasticity. Physical
Therapy 67(2): 206–207.
183
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
Chantraine A, Baribeault A, Uebelhart D, Gremion G
(1999). Shoulder pain and dysfunction in hemiplegia:
effects of functional electrical stimulation. Archives of
Physical Medicine and Rehabilitation 80(3): 328–331.
Cincotti F, Pichiorri F, Arico P, Aloise F, Leotta F, de Vico FF,
Millán Jdel R, Molinari M, Mattia D (2012). EEG-based
brain–computer interface to support post-stroke motor re-
habilitation of the upper limb. Conference Proceedings: An-
nual International Conference of the IEEE Engineering in
Medicine and Biology Society 2012: 4112–4115.
Daly JJ, Cheng R, Rogers J, Litinas K, Hrovat K, Dohring M
(2009). Feasibility of a new application of noninvasive brain
computer interface (BCI): a case study of training for
recovery of volitional motor control after stroke. Journal
of Neurologic Physical Therapy 33(4): 203–211.
Dirks ML, Wall BT, Snijders T, Ottenbros CL, Verdijk LB,
van Loon LJ (2014). Neuromuscular electrical stimulation
prevents muscle disuse atrophy during leg immobiliza-
tion in humans. Acta Physiologica 210(3): 628–641.
Do AH, Wang PT, King CE, Abiri A, Nenadic Z (2011).
Brain–computer interface controlled functional electrical
stimulation system for ankle movement. Journal of
Neuroengineering and Rehabilitation 8: 49.
Dohle CI, Rykman A, Chang J, Volpe BT (2013). Pilot study
of a robotic protocol to treat shoulder subluxation in pa-
tients with chronic stroke. Journal of Neuroengineering
and Rehabilitation 10: 88.
Dursun E, Dursun N, Ural CE, Cakci A (2000).
Glenohumeral joint subluxation and reflex sympathetic
dystrophy in hemiplegic patients. Archives of Physical
Medicine and Rehabilitation 81(7): 944–946.
Faghri PD, Rodgers MM, Glaser RM, Bors JG, Ho C,
Akuthota P (1994). The effects of functional electrical
stimulation on shoulder subluxation, arm function re-
covery, and shoulder pain in hemiplegic stroke patients.
Archives of Physical Medicine and Rehabilitation 75(1):
73–79.
Gollee H, Volosyak I, McLachlan AJ, Hunt KJ, Gräser A
(2010). An SSVEP-based brain–computer interface for the
control of functional electrical stimulation. IEEE Transac-
tions on Bio-medical Engineering 57(8): 1847–1855.
Granat MH, Ferguson AC, Andrews BJ, Delargy M
(1993). The role of functional electrical stimulation
in the rehabilitation of patients with incomplete spinal
cord injury – observed benefits during gait studies.
Paraplegia 31: 207–215.
Gregson JM, Leathley M, Moore AP, Sharma AK, Smith
TL, Watkins CL (1999). Reliability of the Tone Assess-
ment Scale and the modified Ashworth scale as clinical tools
for assessing poststroke spasticity. Archives of Physical
Medicine and Rehabilitation 80(9): 1013–1016.
Halder S, Agorastos D, Veit R, Hammer EM, Lee S, Varkuti
B, Bogdan M, Rosenstiel W, Birbaumer N, Kübler A
184
Jang et al.
(2011). Neural mechanisms of brain–computer interface
control. NeuroImage 55(4): 1779–1790.
Hermes D, Vansteensel MJ, Albers AM, Bleichner MG,
Benedictus MR, Mendez Orellana C, Aarnoutse EJ,
Ramsey NF (2011). Functional MRI-based identifica-
tion of brain areas involved in motor imagery for im-
plantable brain–computer interfaces. Journal of Neural
Engineering 8(2): 025007.
Koyuncu E, Nakipoğlu-Yüzer GF, Doğan A, Ozgirgin N
(2010). The effectiveness of functional electrical stimula-
tion for the treatment of shoulder subluxation and shoul-
der pain in hemiplegic patients: a randomized controlled
trial. Disability and Rehabilitation 32(7): 560–566.
Linn SL, Granat MH, Lees KR (1999). Prevention of
shoulder subluxation after stroke with electrical stimu-
lation. Stroke 30(5): 963–968.
Lo HC, Hsu YC, Hsueh YH, Yeh CY (2012). Cycling exer-
cise with functional electrical stimulation improves pos-
tural control in stroke patients. Gait and Posture 35(3):
506–510.
Miyamoto S, Kondo T, Suzukamo Y, Michimata A, Izumi
S (2009). Reliability and validity of the Manual Function
Test in patients with stroke. American Journal of Phys-
ical Medicine and Rehabilitation 88(3): 247–255.
Paci M, Nannetti L, Rinaldi LA (2005). Glenohumeral
subluxation in hemiplegia: an overview. Journal of Re-
habilitation Research and Development 42(4): 557–568.
Park GY, Kim JM, Sohn SI, Shin IH, Lee MY (2007). Ul-
trasonographic measurement of shoulder subluxation
in patients with post-stroke hemiplegia. Journal of Re-
habilitation Medicine 39(7): 526–530.
Prasad G, Herman P, Coyle D, McDonough S, Crosbie J
(2010). Applying a brain–computer interface to sup-
port motor imagery practice in people with stroke for
upper limb recovery: a feasibility study. Journal of
Neuroengineering and Rehabilitation 7: 60.
Rushton DN (2003). Functional electrical stimulation and
rehabilitation – an hypothesis. Medical Engineering and
Physics 25: 75–78.
Sabut SK, Sikdar C, Kumar R, Mahadevappa M (2011).
Improvement of gait & muscle strength with functional
electrical stimulation in sub-acute & chronic stroke pa-
tients. Conference Proceedings: Annual International
Conference of the IEEE Engineering in Medicine and
Biology Society 2011: 2085–2088.
Sahin N, Ugurlu H, Albayrak I (2012). The efficacy of elec-
trical stimulation in reducing the post-stroke spasticity:
a randomized controlled study. Disability and Rehabili-
tation 34(2): 151–156.
Silvoni S, Ramos-Murguialday A, Cavinato M, Volpato C,
Cisotto G, Turolla A, Piccione F, Birbaumer N (2011).
Brain–computer interface in stroke: a review of prog-
ress. Clinical EEG and Neuroscience 42(4): 245–252.
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
Jang et al.
Vafadar AK, Côté JN, Archambault PS (2015). Effective-
ness of functional electrical stimulation in improving
clinical outcomes in the upper arm following stroke: a
systematic review and meta-analysis. BioMed Research
International 2015: 729768.
Vuagnat H, Chantraine A (2003). Shoulder pain in hemi-
plegia revisited: contribution of functional electrical
Occup. Ther. Int. 23 (2016) 175–185 © 2016 John Wiley & Sons, Ltd.
BCI-controlled FES Improved Shoulder Subluxation in Stroke Survivors
stimulation and other therapies. Journal of Rehabilita-
tion Medicine 35(2): 49–54.
Wassermann E (1995). Changes in motor representation
with recovery of motor function after stroke: combined
electrophysiological and imaging studies. EEG Clinical
Neurophysiol 97: S26.
185