A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients

Bel

en Rubio Ballester

, Alica Lathe

, Esther Duarte

, Armin Duff

and Paul F. M. J. Verschure

1,3

Laboratory of Synthetic Perceptive, Emotive, and Cognitive Systems (SPECS), Departament de Tecnologies

de la Informaci

oi la Comunicaci

o(DTIC), Universitat Pompeu Fabra, Roc Boronat, 138 08018 Barcelona, Spain

Servei de Medicina F

ısica I Rehabilitaci

o, Hospitals del Mar I l’Esperanc¸a,

Institut Hospital del Mar d’Investigacions M

ediques, Barcelona, Spain

Instituci

o Catalana de Recerca i Estudis Avanc¸ats (ICREA), Barcelona, Spain

Keywords:

Stroke, Learned Non-use, Wearables, Motor Rehabilitation, Hemiparesis.

Abstract:

After stroke, many patients experience hemiparesis or weakness on one side of the body. In order to com-

pensate for this lack of motor function, they tend to overuse their non-affected limb. This so called learned

non-use may be one of the most relevant contributors to functional loss after post-stroke hospital discharge.

We hypothesize that frequent exposure to movement related feedback through a wearable bracelet device may

1) increase the patient’s intrinsic motivation for using the paretic limb, and 2) counteract learned non-use,

therefore inducing motor recovery. First, to validate the accelerometers-based measurement of arm use, we

recruited 10 right-handed volunteers without neurological impairments. Second, we explored the acceptability

and clinical impact of a low-cost wearable system on 4 chronic stroke patients with hemiparesis. Our results

suggest that frequent exposure to direct feedback about arm use promotes the integration of the paretic limb

in the performance of instrumental activities of daily living (iADLs). In addition, results from questionnaires

revealed that the use of wearable devices may inﬂuence positively the patient’s intrinsic motivation for using

the affected arm. To the best of our knowledge, this is the ﬁrst study suggesting the beneﬁts of wearable-based

feedback as an intervention tool for counteracting learned non-use.

1 INTRODUCTION

After hospital discharge, up to 55% to 75% of stroke

patients experience persistent motor impairments (Lai

et al., 2002) and may even suffer substantial declines

in function in the following 6 months. A number of

studies suggest that this loss may be due to the lack

of use of the paretic limb (Lai et al., 2002), a phe-

nomenon that has been called learned non-use. Re-

cent work on studying the dynamics of motor recov-

ery after stroke have shown that learned non-use may

emerge as a consequence of decision making for mo-

tor optimization, therefore being dependent on two

main factors: the expected success and the expected

cost of using either effector (Hidaka et al., 2012; Han

et al., 2008; Ballester et al., 2015a). On these basis,

there may exist different strategies for counteracting

learned non-use. For instance, Constrained-Induced

Movement Therapy (CIMT) proposes to reduce the

probability of success and increases the cost of us-

ing the non-affected limb by restricting its movement

and tactile feedback using a mitt (Taub and Uswatte,

2003). Recently, we have shown that Reinforcement-

Induced Movement Therapies (RIMT) may be com-

plementary to CIMT (Ballester et al., 2015a; Ballester

et al., 2015b). In RIMT, visual manipulations during

training increase the probability of success and reduce

the cost of using the paretic limb. However, these re-

habilitation protocols are usually limited to short ses-

sions of intervention and may not be suitable for unsu-

pervised domiciliary environments. In light of these

limitations, the use of wearable devices could be spe-

cially suitable for the persistent monitoring and treat-

ment of leaned non-use.

The current state of research and technological de-

velopment shows a tendency towards the gamiﬁca-

tion of rehabilitation tools, combining various types

of sensors to capture motion and posture. Several

studies have tested the reliability and validity of us-

ing accelerometers for measuring arm use in activities

of daily living (Noork

oiv et al., 2014; Uswatte et al.,

2005). In this vein, signiﬁcant effort has been made in

evaluating the acceptability of wearable devices that

incorporate accelerometers for the quantiﬁcation of

motor performance and recovery (Wang et al., 2014).

The application of wearable devices to the rehabilita-

Ballester, B., Lathe, A., Duarte, E., Duff, A. and Verschure, P..

A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients.

In Proceedings of the 3rd International Congress on Neurotechnology, Electronics and Informatics (NEUROTECHNIX 2015), pages 24-31

ISBN: 978-989-758-161-8

tion ﬁeld offers a number of advantages, such as im-

proved objectivity, sensitivity, and ease of measure-

ment of therapy outcomes. However, because of their

ubiquity, wearable devices may be also useful as inter-

vention tools. Their design and interface enables the

frequent delivery of multimodal feedback during the

performance of iADLs, which may facilitate the re-

alignment of attention towards the affected limb, thus

encouraging the selection of the weaker arm.

Recently, Markopoulos et. al (Markopoulos et al.,

2011) developed a credibility and usability study on

an experimental wearable device that monitors the

patient’s behavior and displays feedback about the

use of the affected versus unaffected arm. So far,

previous work on wearable devices for rehabilitation

describe prototypes and techniques for the integra-

tion of monitoring hardware in wearable garments

as well as communications systems (Uswatte et al.,

2005). Uswatte et al. conducted a clinical experi-

ment in which 20 stroke patients wore an accelerom-

eter on each arm, the chest, and the more affected

leg. Recordings from each sensor were used to es-

timate the duration of movement as a percentage of

the total recording period. Results revealed a strong

correlation between the accelerometers-derived mea-

surements and the Motor Activity Log (Uswatte et al.,

2006). More recent studies have validated and ex-

tended these ﬁndings showing strong correlations be-

tween triaxial accelerometry-derived measurements

and the Quality of Movement scores (van der Pas

et al., 2011) or the National Institutes of Health Stroke

Scale (NIHSS) (Gubbi et al., 2013). However, there

is no evidence yet about the clinical impact of these

types of devices.

The aim of this study is to evaluate the potential

of a wearable system for measuring the amount of

use of the paretic arm in iADLs in chronic stroke pa-

tients with upper extremity hemiparesis. Speciﬁcally,

we hypothesize that frequent exposure to movement

related feedback through a wearable device may 1)

counteract learned non-use, and 2) increase the pa-

tient’s intrinsic motivation for using the paretic limb.

This work presents results from a pilot study explor-

ing stroke patients’ acceptance of a wearable device

for independent usage in their home setting and its ef-

fectiveness as an intervention tool for promoting the

use of the paretic limb.

2 METHODS

2.1 Equipment

The RGS-Wear is a wearable system for the con-

Figure 1: Prototype wearable bracelets integrating the

MetaWear board and Velcro straps. the MetaWear board

includes a low-power, 3-axial capacitive micromachined ac-

celerometer, RGB LEDs, and a coin vibrating motor.

tinuous monitoring of arm use in hemiparetic stroke

patients. It is composed by a pair of bracelets

and a smartphone (Sony Xperia Z3 Compact).

The bracelets include a coin-sized Bluetooth-Board

(MetaWear, MbientLab, San Francisco, CA.) with in-

tegrated accelerometer, a vibrating motor, an ultra

bright RGB LED, a battery, and a wristband (Fig.

1). The accelerometer is a Freescale MMA8452Q:

a smart low-power, three-axis, capacitive micro-

machined accelerometer with a resolution of 12 bits.

Data recordings from each accelerometer are con-

tinuously monitored and sent through Bluetooth to

a paired smartphone, which demands the patient to

carry the smartphone with him or her throughout the

day. For this purpose the participants were equipped

with a holding bag for the phone to be placed at the

waist.

2.2 Quantiﬁcation of Arm Use

For monitoring purposes, the acceleration data from

each device (left bracelet, right bracelet, and smart-

phone) is sampled at 50Hz for each directional di-

mension. In order to derive from these data some

meaningful quantiﬁcation of arm use, we followed a

number of steps. First, over a one minute epoch, we

computed the mean squared sum of the acceleration:

β =

+ α

(1)

This measurement represents a rough index of the

amount of movement of the object to which the ac-

celerometer is attached. This method has been shown

to be an adequate approximation of Energy Expen-

diture (EE) in comparison to measurements derived

from the heart’s electrical activity, muscle activation,

and oxygen consumption (van der Pas et al., 2011;

A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients

Tsurumi et al., 2002). Next, in order to provide feed-

back to the user, we deﬁned the change in the EE of

the paretic arm as:

δ =

− γ

· 100 (2)

where γ

refers to the mean activity of the paretic

arm per hour, γ

corresponds to the mean activity

of the paretic arm at baseline, and arm activity is

given by the difference between the activity of the

corresponding arm β

and the activity of the body β

γ = β

− β

(3)

where [γ]

γ, if γ > 0

0, if γ < 0

In addition, in order to monitor arm balance, we

compute the ratio between the daily mean activity of

the paretic γ

, and non-paretic arm γ

. Thus arm bal-

ance is given by:

θ =

+ γ

∗ 100 (4)

Figure 2: Graphical interface for feedback delivery. A.

Types of feedback according to the change in EE by hour

(Hourly Feedback). B. Example of Review Feedback show-

ing the change in Energy Expenditure (EE) achieved dur-

ing the morning (from 10 to 13 hours). C. Example of Re-

view Feedback delivered at the end of the day, showing the

change in EE by hour along the daily session (from 10 to 19

hours).

2.3 Feedback Design

The main objective of this study was to evaluate the

inﬂuence of feedback of performance on arm use.

Therefore, we ﬁrst explored how to deliver this type

of information to the patient in an efﬁcient and mean-

ingful manner. The design principles shaping RGS-

Wear were derived from the Self-Determination The-

ory proposed by Ryan et al. (Ryan et al., 2008).

This theory deﬁnes three main behavioral mediators,

which determine a patient’s self-engagement within

the process of Health Behavior Change:

1. Autonomy: the patient’s degree of Self-

engagement as the willingness to change due to

the self-referenced value of the targeted behavior.

2. Competence: the individual’s capacity to af-

ford a change.

3. Relatedness: the relation of the patient with

the practitioner, who facilitates the other two me-

diators.

Based on these principles, the RGS-Wear was de-

signed to serve as a rehabilitation device for self-

reinforcement. The incoming feedback messages are

delivered through the vibration of both the smart-

phone and bracelets. Additionally, the phone displays

a sound signal, whereas the bracelets’ LEDs blink in

red, green or blue, according to the improvement cat-

egory achieved (Table 1). Every message has a con-

ﬁrmation button, which should be pressed to conﬁrm

its reception. The following subsections describe in

detail these feedback messages.

2.3.1 Hourly Feedback

The Hourly Feedback was designed to be simple and

effective as the user derives all information about im-

provement in use with a short glance at his smart-

phone (Fig. 2A). A percentage indicates the mean

change in the EE of the paretic limb in respect to

baseline that was achieved within the previous hour

(mean δ, see Eq.2). This type of messages are accom-

panied by an illustration of the upper-limbs. Beneath,

a comment window shows different motivational say-

ings (e.g. ”Stay active along the day.”). Considering

not to generate pressure, the text appears as a general

request to stay active within the day, regardless of the

numerical result shown.

This feedback was designed to support the pa-

tient’s self-engagement and increase the value of us-

ing the paretic arm in the performance of iADLs. This

type of feedback is thus tightly related to the con-

cept of Autonomy proposed by the Self-Determination

Theory (SDT). In this line, the display of a percentage

NEUROTECHNIX 2015 - International Congress on Neurotechnology, Electronics and Informatics

Table 1: Feedback Categories.

Result Above 10% Between 10% and -10% Below -10%

Category Positive Neutral Negative

Signal in bracelets Green blinking light Blue blinking light Red blinking light

Signal in smartphone Positive green operator Positive blue operator Negative red operator

Figure 3: Averaged activity of the non dominant arm, for

each subject, achieved during the execution of three iADLs

(Wash de dishes, eating, and buttoning up a shirt) under two

different conditions. During the Restricted condition, the

use of the left limb was limited. During the Balanced condi-

tion, the participant was encouraged to use both arms. Red

horizontal lines indicate the inactivity threshold.

improvement in use provides a target oriented self-

regulation.

2.3.2 Review Feedback

The Review Feedback displays a summary of the

Hourly Feedback in session intervals (Fig. 2B), thus

providing knowledge of progress. In this study, RGS-

Wear was pre-programmed to monitor 9 consecutive

hours a day. Daily recordings were partitioned in a

morning, afternoon, and night sessions, and each of

them had a duration of 3 hours. Review Feedback was

provided at the end of each session and at the end of

the day, displaying the hourly mean activity level of

the paretic arm in a graphical chart. The rational for

this feedback was to meet the patients’ psychologi-

cal need of Competence by presenting an overview of

performance over time.

2.3.3 Instructions Slides

The RGS-Wear daily protocol is initialized at 10 a.m.

by presenting a number of welcoming slides accom-

panied by an alarm sound that signalizes the begin-

ning of the monitoring.

2.4 Experimental Paradigm

In order to assess the reliability of the RGS-Wear

for capturing differences in arm use, we ﬁrst con-

ducted an experiment on humans with no neurolog-

ical impairments. We instructed participants to per-

form four iADLs (washing the dishes, eating, button-

ing up a shirt, and walking) while wearing the RGS-

wear bracelets. Each activity had a duration of 3 min-

utes. All activities were performed twice: ﬁrst, partic-

ipants were asked to use their left arm with reduced

intensity as they would normally do (non-dominant

condition), and second, bi-manual execution was en-

couraged (balanced condition).

After validating RGS-Wear as a monitoring tool,

we designed an experimental paradigm to explore the

potential of RGS-Wear for promoting the use of the

paretic limb in stroke patients. Participants were in-

structed to use the RGS-Wear system at home for ﬁve

consecutive weekdays, from 10 to 19 o’clock, except

when bathing. The experiment was divided in three

phases: pre-test, intervention, and post-test. Due to

the ubiquitous presence of the device, the sensation

of being under observation could be cued and may

A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients

Figure 4: Quantifying behavioral changes. A. Difference between arm activity during baseline (day 1). B. Arm Use Balance

between the paretic (red) and non-paretic arm (green) per day exhibited by each patient. C. Mean change in the activity of the

paretic limb with respect to baseline, across the three days of intervention (day 2-4), and post-test (day 5). The y-axis refers

to the change from baseline averaged across subjects.

lead to an over-encouraged behavior. In order to con-

trol for this effect, patients were instructed to wear

the system everyday but did not receive any type of

feedback at day 1 (i.e. pre-test or baseline) and day

5 (i.e. post-test) of the experimental protocol. From

day 2 to day 4 (i.e. intervention phase), the RGS-

wear system provided Hourly Feedback and Review

Feedback to the patient. Before (day 1) and after the

experiment (day 5), participants fulﬁlled an Intrinsic

Motivation Questionnaire (IMQ). The IMQ consisted

in 7 statements designed to capture changes in the pa-

tient’s perceived competence and effort when using

the paretic limb (see APPENDIX, Questionnaire on

Intrinsic Motivation). Answers were reported using a

7-point Likert Scale, ranging from Strongly Disagree

to Strongly Agree. In addition, a Usability Question-

naire (UQ) was administered at the end of the experi-

mental protocol (day 5) to assess the system’s accept-

ability in terms of its hardware design, graphical user

interface (GUI), interaction design, and perceived ef-

ﬁcacy (see APPENDIX, Questionnaire on Usability).

In this questionnaire, answers were reported using a

5-point Likert Scale. The ethics committee of clinical

research of the Parc de Salut Mar approved experi-

mental guidelines.

2.5 Participants

For the validation of the accelerometers-based mea-

surement of arm use, we recruited 10 right-handed

volunteers without neurological impairments (5 fe-

males, mean age = 26.6 ± 2.59 years old). Sec-

ondly, in order to explore the clinical impact of the

RGS-Wear, ﬁve chronic stroke patients were ﬁrst ap-

proached by a doctor from the rehabilitation depart-

ment of Hospital Esperanc¸a in Barcelona to determine

their interest in participating in this research project.

Selected patients met the following inclusion criteria:

1) Ischaemic strokes (Middle cerebral artery territory)

and hemorrhagic strokes (intra-cerebral). 2) Mild-to-

moderate upper-limbs hemiparesis. 3) Age between

45 and 85 years old. 4) Absence of any major cogni-

tive impairments. 5) Frequent smartphone user. One

patient refused to participate. The remaining four pa-

tients (4 males, 70.5 ± 6.76 years old) were included

in the study. Prior to the experiment, all participants

signed informed consent.

3 RESULTS

3.1 Accelerometer-based Measurement

of Arm Use

In order to evaluate the reliability and validity

of accelerometry for measuring arm use in non-

impaired subjects, we examined the subjects’ non-

dominant arm activity under 2 different conditions

(non-dominant and balanced), in four different iADLs

(washing the dishes, eating, buttoning up a shirt, and

walking). As we expected, in the non-dominant con-

dition, performance of iADLs was characterized by

the decreased activity of the left hand (Fig. 3). In

the walking task, activity measures fell below the in-

activity level in both conditions, indicating that the

mean acceleration of each hand was not superior to

the mean body acceleration. These results validate

the reliability of the RGS-Wear system for capturing

the amount of use of the upper-limbs in iADLs.

3.2 Effects on Amount of Use

After exploring the use of wearable devices for arm

use monitoring, we proceeded to investigate its appli-

cability as an intervention tool. Since amount of use

and recovery are tightly coupled, using wearable de-

vices to induce an increase in arm use could have a

positive impact in motor recovery. One approach to

pursue this idea is to use wearable devices to expose

the patient to arm movement related feedback, thus

increasing the intrinsic motivation for integrating the

paretic limb in the performance of iADLs.

NEUROTECHNIX 2015 - International Congress on Neurotechnology, Electronics and Informatics

Figure 5: Responses from questionnaires. A. Average scores quantifying the patient’s intrinsic motivation for using the paretic

limb in the performance of iADLs. B. Mean scores for each category assessed by the usability test.

To address this question, we ﬁrst compared the

mean levels of Energy Expenditure (EE) estimated at

baseline (day 1) for each arm and subject (β in Eq.

1). As expected, we observed that all patients reached

higher EE values when using the less affected limb

(Figure 4A). Patient 3, who presented with mild hemi-

paresis, showed a highly balanced arm use, reaching a

mean EE value of 7.53 for the paretic limb, and 8.85

for the non-paretic limb. These preliminary results

support the use of accelerometry for quantifying arm

use in hemiparetic stroke patients. Next, we analyzed

the change in arm use balance respect to baseline (day

1). Although we observed differences between pa-

tients, the estimation of arm balance values remained

stable within subjects (Fig. 4B). Overall we found

a general increase in the Arm Use Balance, suggest-

ing an increased integration of the affected limb in

the performance of iADLs. However, since Arm Use

Balance is a relative measurement (see Eq. 4), it does

not express the amount of movement. A patient could

therefore achieve positive improvements in Arm Use

Balance by only limiting the movement of the non-

paretic limb. In order to take into account the patient’s

amount of arm movement, we analyzed the change in

the activity of the paretic limb with respect to baseline

(day 1). Interestingly, results revealed an increase in

activity which accumulated along the three days of in-

tervention (Fig. 4C). Even though we observed a drop

in activity at day 5 (post-test), when no feedback was

delivered any more, arm use improvements were still

partially retained.

3.3 Effects on the Patients’ Intrinsic

Motivation

We analyzed the inﬂuence of the RGS-Wear paradigm

on intrinsic motivation by comparing the scores re-

ported by the four patients before (day 1) and after

the treatment (day 5). We observed that, after treat-

ment, 3 out of 4 patients exhibited higher intrinsic

motivation to use the paretic limb (Fig. 5A). Accord-

ing to the Self-Determination Theory, this subjective

improvement may emerge from the repetitive expo-

sure to knowledge of progress, a factor tightly linked

to the behavioral mediator Autonomy.

3.4 Usability

We studied the usability aspects of the RGS-Wear

through a questionnaire that was divided into 4 cat-

egories (5 questions each): hardware, graphical inter-

face, feedback, and perceived efﬁcacy. Overall, the

patients’ ratings were above 3 (neutral), suggesting

that the system’s design was generally accepted. In-

terestingly, we noticed that the rating of hardware fea-

tures was notably lower in comparison to the other

categories. When we explored the patient’s answers

in detail, we found that those statements referring to

the comfort of putting the bracelets on received lower

scores from most of the patients.

4 CONCLUSION AND

DISCUSSION

We have presented results from a pilot study support-

ing the beneﬁts of wearable-based feedback on arm

use. Our results suggest that frequent exposure to

direct feedback about arm use promotes the integra-

tion of the paretic limb in the performance of iADLs.

In addition, results from questionnaires revealed that

the use of wearable devices may inﬂuence positively

the patient’s intrinsic motivation for using the affected

arm.

The work we presented in this article is the con-

tinuation of our previous work on the use of new

technologies for counteracting learned non-use. In

a recent study, we used a neurologically grounded

A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients

computational model of motor recovery (Han et al.,

2008) that can predict the positive inﬂuence of

reinforcement-based training on arm use (Ballester

et al., 2015a). In this work, we proposed that hand

selection is modulated by two main parameters: ex-

pected success and effort. We conducted two clini-

cal experiments that suggested that by increasing the

value of using the paretic limb (expected success) and

decreasing its cost (effort) we can promote its spon-

taneous use (Ballester et al., 2015b) and boost recov-

ery (Ballester et al., 2015a). Based on these ﬁndings,

we now explore how wearable devices could allow

the ubiquitous delivery of a variant of Reinforcement-

Induced Movement Therapy (RIMT).

Our results suggest that monitoring the amount of

arm use and providing knowledge of progress could

provide multiple beneﬁts: 1) it may allow the pa-

tient to set-up implicit goals, and 2) it may increase

the value of using the paretic limb, therefore bias-

ing effector selection patterns. Thus, the repetitive

exposure to reinforcement-based feedback after per-

formance may modify both the individual’s goals and

self-representation. While the ﬁrst may provide the

necessary context for the introduction of behavioral

changes, the second may consolidate them. Inter-

estingly, a recent controlled clinical trial including

156 acute stroke patients evaluated the clinical impact

of using wearable triaxial accelerometers at both an-

kles and recording continuously for 8 hours per day

(Dorsch et al., 2015). Once a week, participants in

the experimental group also reviewed the results of

their summary activity graphs with the therapists. Re-

sults indicated that the group receiving the augmented

feedback did not spend a greater amount of time walk-

ing. This ﬁndings seem to be contrary to our re-

sults. This difference can be explained by three fac-

tors: 1) the RGS-Wear provided frequent daily feed-

back about performance and progress, 2) the patients

using RGS-Wear reviewed their activity feedback au-

tonomously and 3) the RGS-Wear was applied on the

upper-extremities while Dorsch, et al. focused in gait

and lower-extremities.

Future work aims at validating the impact of RGS-

Wear in arm use by conducting a controlled longitu-

dinal clinical study on acute stroke patients. In this

study we plan to measure the retention of improve-

ments in arm use induced by the RGS-Wear, and its

consequent inﬂuence on motor recovery.

ACKNOWLEDGEMENTS

We would like to acknowledge all patients who par-

ticipated in this study. Special thanks to Dr. Boza

omez for her assistance in recruiting stroke patients.

This project was supported through ERC project

cDAC (FP7-IDEAS-ERC 341196), EC H2020 project

socSMCs (H2020-EU.1.2.2. 641321) and MINECO

project SANAR (Gobierno de Espaa).

REFERENCES

Ballester, B. R., Maier, M., Duff, A., San Segundo,

R., Casta

neda, V., and Verschure, P. (2015a).

Reinforcement-Induced Movement Therapy: A novel

approach for overcoming learned non-use in chronic

stroke patients. In Proceedings of the International

Conference on Virtual Rehabilitation. In press., pages

1–6.

Ballester, B. R., Nirme, J., Duarte, E., Cuxart, A., Ro-

driguez, S., Verschure, P., and Duff, A. (2015b).

The visual ampliﬁcation of goal-oriented movements

counteracts acquired non-use in hemiparetic stroke

patients. Journal of neuroengineering and rehabili-

tation, 12(1):50.

Dorsch, a. K., Thomas, S., Xu, X., Kaiser, W., Dobkin,

B. H., Emara, T., Edwards, D., Fonzetti, P., Maasch,

J., Lee, S.-G., Owolabi, M. O., Hamzat, T. K.,

LeBlanc, C. J., Morse, R., Swaminathan, N., Karatas,

G. K., Boza, R., Brown, a. W., Miyai, I., Kawano, T.,

Chen, S.-Y., Hanger, H. C., Zucconi, C., Mammi, S.,

Ghislanzoni, C., Juan, F., and Lang, C. E. (2015). SIR-

RACT: An International Randomized Clinical Trial of

Activity Feedback During Inpatient Stroke Rehabili-

tation Enabled by Wireless Sensing. Neurorehabilita-

tion and Neural Repair, 29(5):407–415.

Gubbi, J., Rao, A. S., Fang, K., Yan, B., and Palaniswami,

M. (2013). Motor recovery monitoring using accel-

eration measurements in post acute stroke patients.

Biomedical engineering online, 12:33.

Han, C. E., Arbib, M. A., and Schweighofer, N. (2008).

Stroke rehabilitation reaches a threshold. PLoS com-

putational biology, 4(8):e1000133.

Hidaka, Y., Han, C. E., Wolf, S. L., Winstein, C. J., and

Schweighofer, N. (2012). Use it and improve it or

lose it: interactions between arm function and use

in humans post-stroke. PLoS computational biology,

8(2):e1002343.

Lai, S.-M., Studenski, S., Duncan, P. W., and Perera, S.

(2002). Persisting consequences of stroke measured

by the Stroke Impact Scale. Stroke, 33(7):1840–1844.

Markopoulos, P., Timmermans, A. A. A., Beursgens, L.,

Van Donselaar, R., and Seelen, H. A. M. (2011). Us’

em: The user-centered design of a device for motivat-

ing stroke patients to use their impaired arm-hand in

daily life activities. In Engineering in Medicine and

Biology Society, EMBC, 2011 Annual International

Conference of the IEEE, pages 5182–5187. IEEE.

Noork

oiv, M., Rodgers, H., and Price, C. I. (2014). Ac-

celerometer measurement of upper extremity move-

ment after stroke: a systematic review of clinical stud-

ies. Journal of neuroengineering and rehabilitation,

11(1):144.

NEUROTECHNIX 2015 - International Congress on Neurotechnology, Electronics and Informatics

Ryan, R. M., Patrick, H., Deci, E. L., and Williams,

G. C. (2008). Facilitating health behaviour change

and its maintenance: Interventions based on self-

determination theory. European Health Psychologist,

10(1):2–5.

Taub, E. and Uswatte, G. (2003). Constraint-induced move-

ment therapy: bridging from the primate laboratory to

the stroke rehabilitation laboratory. Journal of Reha-

bilitation Medicine, 35:34–40.

Tsurumi, K., Itani, T., Tachi, N., Takanishi, T., Suzumura,

H., and Takeyama, H. (2002). Estimation of En-

ergy Expenditure during Sedentary Work with Up-

per Limb Movement. Journal of occupational health,

44(6):408–413.

Uswatte, G., Foo, W. L., Olmstead, H., Lopez, K., Holand,

A., and Simms, L. B. (2005). Ambulatory monitoring

of arm movement using accelerometry: an objective

measure of upper-extremity rehabilitation in persons

with chronic stroke. Archives of physical medicine

and rehabilitation, 86(7):1498–1501.

Uswatte, G., Taub, E., Morris, D., Light, K., and Thompson,

P. A. (2006). The Motor Activity Log-28 assessing

daily use of the hemiparetic arm after stroke. Neurol-

ogy, 67(7):1189–1194.

van der Pas, S. C., Verbunt, J. A., Breukelaar, D. E., van

Woerden, R., and Seelen, H. A. (2011). Assessment

of arm activity using triaxial accelerometry in patients

with a stroke. Archives of physical medicine and re-

habilitation, 92(9):1437–1442.

Wang, Q., Chen, W., and Markopoulos, P. (2014). Literature

review on wearable systems in upper extremity reha-

bilitation. In IEEE-EMBS International Conference

on Biomedical and Health Informatics (BHI), pages

551–555. IEEE.

APPENDIX

Questionnaire on Intrinsic Motivation: 1- Inte-

grating the affected in the performance of activities

of daily living allows me to be more independent.

2- I’m quite competent when I use my affected arm.

3- It’s really tiring to use the affected arm in my

activities of daily living. 4- I feel secure when I use

the affected arm for eating. 5- I feel secure when I

use the affected arm for washing the dishes. 6- I feel

secure when I use the affected arm for dressing up.

7- How much do you use the affected arm?

Questionnaire on Usability: 1- It was easy to put on

the bracelets without help. 2- It was easy to put on

the smartphone without help. 3- It was comfortable

to wear the bracelets. 4- It was comfortable to wear

the smartphone. 5- It was easy to move the affected

arm while wearing the bracelets. 6- It was easy to

hear the messages alarm on the smartphone. 7- It was

easy to notice the bracelets vibrations and lights. 8- It

was easy to understand the messages on the screen of

the smartphone. 9- It was ease to press the buttons on

the screen of the smartphone. 10- It was easy to read

the texts appearing on the screen of the smartphone.

11- I think I missed more than 3 messages a day. 12-

Sometimes the messages were annoying. 13- Some-

times the messages scared me. 14- The hourly feed-

back about the amount of movement was correct. 15-

Feedback about the amount of arm movement across

days was correct. 16- I think the messages were ac-

curate in reporting my activity. 17- I think the levels

of activity reported by the messages were lower than

my real activity level. 18- I think the levels of activity

reported by the messages were higher than my real ac-

tivity level. 19- There were too many messages along

the day. 20- I would like to receive more messages.

A Wearable Bracelet Device for Promoting Arm Use in Stroke Patients