Virtual Reality in the Rehabilitation of the Upper Limbafter Stroke: The User’s Perspective

J.H. CROSBIE, M.Sc.,1 S. LENNON, Ph.D.,1 M.D.J. MCNEILL, Ph.D.,2 and S.M. MCDONOUGH, Ph.D.1

ABSTRACT

Our group has developed a relatively low-cost virtual reality (VR) system for rehabilitationof the upper limb following stroke. Our system is immersive in that the participant views arepresentation of their arm and hand, reaching and retrieving objects in the virtual environ-ment (VE), through a head-mounted display (HMD). This is thought to increase the partici-pant’s sense of presence in the VE and may lead to improved rehabilitation outcomes.However, use of immersion, particularly with our low-cost system, may increase the inci-dence of side effects reported. Therefore, the aim of this project was to assess the interactionof healthy users and those following stroke, in terms of their experience of presence in theVE and the rate of self-reported side effects. Differences in rates of perceived exertion, levelsof enjoyment, and sense of control between both groups were also explored.

137

CYBERPSYCHOLOGY & BEHAVIORVolume 9, Number 2, 2006© Mary Ann Liebert, Inc.

INTRODUCTION

RECOVERY OF UPPER LIMB function is a major prob-lem, with 30–66% of stroke survivors no longer

being able to use the affected arm.1 This can be ex-plained in part by the site of injury in the cortex,2which can cause limb paresis, which limits activepractice with the arm in the real world.3 Other fac-tors are low levels of interaction between the pa-tient and the environment,4,5 ineffective therapytechniques,6 and the very small percentage of timeactually spent practicing tasks.4

The use of virtual environment (VE) technologyhas been advocated for this problem, as it can bemanipulated to avoid the physical constraints thatcan prevent a patient practicing in the real world.Several other groups have developed virtual reality(VR) for rehabilitation of the upper limb,7–9 withthe main difference between our system and othersystems being the immersive element coupled withrelatively low cost hardware, especially the data-

glove and head-mounted display (HMD). How-ever, these aspects of our system may increase theincidence of side effects, as it is well known that aproportion of healthy VR users will report side ef-fects during and after exposure to an immersiveVE. These side effects are similar to motion sickness(e.g., sweating, nausea, headaches, disorientation,and balance disturbances) and can be attributed todelays between the sampling of head and limb po-sitions, and the presentation of an appropriateimage through the HMD, leading to an incongruitybetween visual cues through the HMD and vestib-ular motion cues.10 Considerable variations in theextent to which healthy people report these symp-toms have been identified,10 and to our knowledge,there are no data reported in users followingstroke.

Therefore, although there is much potential forthe use of immersive VE in clinical applications,there are problems that could limit their ultimateusability. Use of immersion, particularly with our

1Health and Rehabilitation Sciences Research Institute, University of Ulster, Newtownabbey, Co. Antrim, Northern Ireland.2Faculty of Engineering, University of Ulster, Coleraine, Co. Londonderry, Northern Ireland.

14246c06.pgs 4/14/06 10:04 AM Page 137

low-cost system, may increase the incidence of sideeffects reported.

The aim of this project was to assess the interac-tion of the user, in both healthy and stroke popula-tions, in terms of their experience of presence in theVE and the rate of self-reported side effects. Differ-ences in rates of perceived exertion, levels of enjoy-ment, and sense of control between both groupswere also explored.

METHODS

Equipment: University of Ulster Virtual RealityRehabilitation System (UUVRR)

Our research group has built a system for use instroke rehabilitation, which gives the user the abil-ity to move their arm and hand in a VE and interactwith familiar objects (e.g., cup, pen, cylinder,blocks) by touching, grasping, and moving theirupper limb. The VE has been created using World-up Software (<www.sense8.cierraweb.net>) andconsists of a home environment with objects on akitchen table. Tracking of the users arms and hand,and manipulation of objects in the VE is achievedwith four electromagnetic sensors (<www.ascen-sion-tech.com/products/netsofbirds/index.html>)(Flock of Birds [FOB], which provide real-time 6-degrees-of-freedom (position and orientation)movement and a dataglove (<www.isense.com/products/pro/index.htm>). The user views a styl-ized representation of their arm and hand in theVE, through a HMD (<www.5dt.com/products/pdataglove5.html>) that provides visual and audiocues during the tasks (Figs. 1 and 2).

Specification of tasks

Functional tasks have been designed to incorpo-rate a range of levels of difficulty for reach, grasp,release and manipulative components. The tasksengage both individual joints and the whole arm,and some require fine upper limb and hand motorcontrol. A wrist extension task has also been de-signed to focus on the movement at this joint,which is likely to be functionally useful to promotewrist extensor activity, as it impacts on gripstrength and the dynamic positioning of the handfor grasping objects. The tasks can be chosen de-pending on the patient’s motor dysfunction; for ex-ample, if the patient has difficulty reachingforward with the arm, the target for reaching can beincreased in size and the distance to reach reduced.The level of difficulty can be adjusted in accordancewith the individual’s performance. A specially de-signed graphical user interface has been incorpo-rated to allow straightforward initialisation of thesystem, configuration of individual tasks, andscripting of whole sessions, with the goal beingthat this complex system should be usable by phys-iotherapists and other non-computing personnelwith a minimum of training.

Subjects

Healthy volunteers and people after stroke wererecruited in order to assess their experience ofusing the UUVRR system. Inclusion criteria forthose following stroke were first stroke with amotor impairment of the upper limb as a primarydeficit, muscle strength greater than 2/5 on theMedical Research Council (MRC) scale; stable med-ical condition; ability to communicate; good cog-nitive ability as indicated by a score greater than 15

138 CROSBIE ET AL.

FIG. 1. Dataglove.

FIG. 2. Head-mounted display.

14246c06.pgs 4/14/06 10:04 AM Page 138

on the Mini Mental State Examination (MMSE),11

Motricity Index score12 greater than 26, indicatingeither activity of the shoulder or elbow and begin-nings of prehension in the hand. Exclusion criteriawere patients with significant dysphasia; severesymptomatic arm pain; poor muscle strength (asdetermined by a score of less than 2/5 on the MRCscale of shoulder girdle muscle strength), an unsta-ble medical condition, poor cognitive status (as de-fined by the MMSE)11 and severe visual-spatialneglect (The Line Cancellation Test).13 Healthyadults were recruited from staff at the University ofUlster. Informed written or witnessed verbal con-sent was obtained from all recruited patients andhealthy adults.

Description of user session

Each person undertook one session in whichthey worked through a short series of exercises forthe upper limb, which included reach to targettasks, a wrist extension task and a reach and re-trieve of an object task. Features that have been de-signed into the system were evaluated: the audibleinstructions and pre-recorded demonstrations; theoptions to add or remove objects and to adjust dis-tances and the height of objects in relation to theuser; and finally the type of view of the VE whenusing the HMD could be set as “moving” (i.e., theVE was updated in synchrony with the FOB sensoron the HMD and thus was responsive to the usershead position) or “fixed” (i.e., the VE remained sta-tic and only the arm moved within it). On average,sessions lasted 30–40 min, with a tendency forlonger sessions in healthy users.

Assessment of the user experience

i. The Immersive Tendencies Questionnaire (ITQ).The ITQ14 was used in this study to determinedifferences in the tendencies of individuals toexperience presence. The ITQ is comprised of 29questions, the answers to which are presentedas a horizontal visual analogue scale with leftand right anchors, representing negative andpositive immersive tendencies, respectively. Auser with a positive immersive tendency asscored on the ITQ is likely to experience higherlevels of presence in the VR and this has beenassociated with better task performance.14

ii. Task Specific Feedback Questionnaire (TSFQ). TheTSFQ15 assessed the users perceived difficulty oftask completion, levels of enjoyment and senseof control in the VE. This was used to evaluate

the experience of using the system. The instru-ment comprised six questions, scored on a Lik-ert scale of 1–5, giving a total score of 1–30.Higher scores indicated that the user had a morefavorable experience using the virtual system.

iii. Side effects and levels of exertion. This was mea-sured using question 7 on the TSFQ in additionto user self-report on the nature of any discom-fort or simulator sickness symptoms followingthe use of the system. The Borg Scale ofPerceived Exertion16 was used to assess the indi-vidual’s perception of any physical exertion ex-perienced whilst exercising their arm in the VE.Perceived exertion is the overall effort or dis-tress of the body during exercise, the scaleranges from 0, representing no perceived exer-tion or discomfort, to 10, representing the great-est amount of perceived exertion the users hasever experienced.

iv. User comments on design aspects of the system.These were collated during each session. TheITQ was administered prior to the subject partic-ipating in the VR session. The TSFQ, side effectsand the Borg Scale of Perceived Exertion was ad-ministered immediately after the VR session.

RESULTS

The demographic details of all subjects (10healthy and 10 stroke users) can be seen in Table 1.The mean age differed between the two groups(healthy users, 42 years [±SD, 12.7]; stroke users 60years [±SD, 17.9]). The time since stroke rangedfrom 6 days to 20 years (Table 1), with greater num-bers with chronic stroke. Results are displayed inTable 2, and it can be seen that there was no differ-ence in the sense of presence, or likelihood to be-come immersed, between healthy users and thosefollowing stroke, with equal numbers scoring posi-tively or negatively on the ITQ.

In general, the scores on the TSFQ were similarbetween groups, with higher scores indicating highlevels of enjoyment, control, presence and few sideeffects. However, it is worth noting that there weretwo users with stroke, with higher levels of disabil-ity, who had low TSFQ scores, which suggests aless positive experience in the VE.

With regards to perceived levels of exertion,there was a difference between the groups (healthyusers, Borg scores of 0–4; stroke users, Borg scoresof 3–7). This is not an unexpected finding that indi-cated that the tasks were quite easy to execute forhealthy users whereas for the stroke user the tasks

VR FOR REHABILITATION OF UPPER LIMB AFTER STROKE 139

14246c06.pgs 4/14/06 10:04 AM Page 139

were likely to require an increase in motor activityand thus effort to complete them.

It was notable that a greater number of healthyusers (n = 5) reported side effects (transient symp-toms of headache, dizziness, discomfort, and nau-sea) in the VE compared to stroke users (n = 3); thismay have been linked to increased time of use in theVE by healthy users. Moreover, fewer healthy usersliked the “moving view” in the VR compared to the“fixed view,” which was the only notable differencein user report of the design aspects of the system.

CONCLUSION

The main finding of this user study was that ourlow-cost immersive VR system was acceptable to themajority of healthy and stroke users. There were

some notable differences between the groups—someexpected (e.g., the levels of perceived exertion) andsome unexpected (e.g., the greater report of side ef-fects and preference for the “fixed view” in the VE byhealthy users). These differences may be explainedby the increased time spent in the VE, which was notupdated in synchrony with the head position; in-deed, there was a significant time lag between thetracking of the sensor on the HMD (and thus headposition) and the update of the VE. This is potentiallya major drawback with our system, which we pro-pose to control with shorter VE exposures.

ACKNOWLEDGMENTS

We would like to thank the Northern IrelandChest, Heart and Stroke Association (6th Floor, 22

140 CROSBIE ET AL.

TABLE 1. DEMOGRAPHIC DETAILS OF HEALTHY AND STROKE USERS

Healthy Age Age Side ofuser (years) Stroke user (years) stroke Time since stroke

1 70 1 65 Left 3 years2 48 2 83 Right 6 days3 39 3 78 Right 2 years4 44 4 62 Right 9 years5 42 5 25 Right 20 years6 48 6 62 Left 4 years7 37 7 76 Left 3 years8 23 8 41 Left 5 years9 44 9 61 Right 11 years

10 28 10 47 Right 14 years

TABLE 2. RESULTS OF THE IMMERSIVE TENDENCIES QUESTIONNAIRE (ITQ), TASK SPECIFIC FEEDBACK

QUESTIONNAIRE (TSFQ), AND BORG SCALE OF PERCEIVED EXERTION FOR HEALTHY AND STROKE GROUPS

Health group Stroke group

Side SideSubject ITQ Borg TSFQ effects Subject ITQ Borg TSFQ effects

1 � 2 23 No 1 + 7 24 Yes2 � 2 24 Yes 2 � 5 27 No3 � 1 27 Yes 3 + 3 26 No4 + 1 28 Yes 4 NS 5 15 No5 + 0 24 No 5 NS 5 8 No6 � 0.5 27 No 6 + 10 28 No7 + 2 30 Yes 7 + 2 28 No8 + 0.5 25 No 8 � 6 22 Yes9 � 4 19 Yes 9 � 7 22 No

10 + 3 25 No 10 + 6 23 Yes

NS, no score.

14246c06.pgs 4/14/06 10:04 AM Page 140

Great Victoria Street, Belfast BT2 7LX; applicationnumber 2002004) for support.

REFERENCES

1. van der Lee, J.H., Wagenaar, R.C., & Lankhorst, G.(1999). Forced use of the upper extremity in chronicstroke patients. Stroke 30:2369–2375.

2. Kandel, E.R., Schwartz, J.H., & Jessell, T.M. (1991).Principles of neural science, 3rd ed. London: Prentice-Hall International Inc.

3. Chae, J., Berthoux, F., & Bohine, T. (1998). Neuromus-cular stimulation for upper extremity motor andfunctional recovery in acute hemiplegia. Stroke29:975–979.

4. Tinson, D.J. (1989). How stroke patients spend theirdays. International Disability Studies 11:45–49.

5. Mackey, F., Ada, L., & Heard, R. (1986). Stroke reha-bilitation: are highly structured units more con-ducive to physical activity than less structures units?Archives of Physical Medicine and Rehabilitation77:1066–1070.

6. Kraft, G.H., Fitts, S.S., & Hammond, M.C. (1992).Techniques to improve function of the arm and handin chronic hemiplegia. Archives of Physical Medicineand Rehabilitation 73:220–227.

7. Merians, A.S., Jack, D., Boian, R., et al. (2002). Virtualreality–augmented rehabilitation for patients follow-ing stroke. Physical Therapy 82:898–915.

8. Holden, M.K., & Dyar, T. (2002). Virtual environmenttraining: a new tool for neurorehabilitation. Neurol-ogy Report 26:62–71.

9. Broeren, J., Rydmarck, M., & Sunnerhagen, K.S.(2004). Virtual reality and haptics as a training devicefor movement rehabilitation after stroke: a single-

case study. Archives of Physical Medicine and Rehabili-tation 85:1247–1250.

10. Kolanski, E.M. (1995). Simulator sickness in virtualenvironments. U.S. Army Research Institute for theBehavioural and Social Sciences Virginia, TechnologyReport 1027.

11. Folstein, M.F., Folstein, S.E., & McHugh, P.R. (1975).Mini Mental state. A practical method for gradingthe cognitive state of patients for the clinician. Jour-nal of Psychiatric Research 12:189–198.

12. Wade, D.T. (1992). Measurement in neurological rehabil-itation. Oxford: Oxford University Press.

13. Albert, M.L. (1973). A simple test of visual neglect.Neurology 23:658–664.

14. Witmer, B.G., & Singer, M.J. (1998). Measuring pres-ence in virtual environments: a presence question-naire. Presence 7:225–240.

15. Kizony, R., Raz, L., Katz, N., et al. (2003). Using avideo-projected VR system for patients with spinalcord injury, Presented at the 2nd InternationalWorkshop on Virtual Rehabilitation, Piscataway,NJ.

16. Borg, G.V. (1982). Psychophysical bases of perceivedexertion. Medicine and Science in Sport and Exercise14:377–381.

Address reprint requests to:Dr. S.M. McDonough

Health and Rehabilitation Sciences Research InstituteUniversity of Ulster

Shore RoadNewtownabbey, Co. Antrim

BT37 0QB, Northern Ireland

E-mail: s.mcdonough@ulster.ac.uk

VR FOR REHABILITATION OF UPPER LIMB AFTER STROKE 141

14246c06.pgs 4/14/06 10:04 AM Page 141