Feasibility, Motivation, and Selective Motor Control:Virtual Reality Compared to Conventional Home Exercise

in Children with Cerebral Palsy

C. BRYANTON, B.Sc. (PT),1 J. BOSSÉ, B.Sc. (PT),1 M. BRIEN, B.Sc. (PT),2 J. MCLEAN, M.D.,2A. MCCORMICK, M.D.,2 and H. SVEISTRUP, Ph.D.1

ABSTRACT

Children with cerebral palsy (CP) have difficulty controlling and coordinating voluntarymuscle, which results in poor selective control of muscle activity. Children with spastic CPcompleted ankle selective motor control exercises using a virtual reality (VR) exercise systemand conventional (Conv) exercises. Ankle movements were recorded with an electrogo-niometer. Children and their parents were asked to comment on their interest in the exerciseprograms. Greater fun and enjoyment were expressed during the VR exercises. Childrencompleted more repetitions of the Conv exercises, but the range of motion and hold time inthe stretched position were greater during VR exercises. These data suggest that using VR toelicit or guide exercise may improve exercise compliance and enhance exercise effectiveness.

123

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

INTRODUCTION

CEREBRAL PALSY (CP), a non-progressive but notunchanging disorder of movement and/or pos-

ture due to an insult or anomaly of the developingbrain, affects up to one in three premature babies,and over 50,000 Canadians have CP.1 CP is classifiedbased on how much of what parts of the body areaffected, as well as by the movement anomaly. Thedegree of severity varies greatly, and each case isoften described as mild, moderate or severe.2

With CP, the inability to control and coordinatevoluntary muscle results in poor selective controlof muscle activity. The “orderly phasing in and outof muscle activation, co-activation of muscles withsimilar biomechanical functions, and limited co-activation of antagonists during phasic or freemovement”3 is disrupted and leads to coordina-tion, balance, and ambulation deficits. Specifically,the child with CP cannot appropriately contract thetibialis anterior (TA) muscle, creating functional

problems such as the inability to achieve heel strikeduring ambulation.

Training of selective motor control of the TAmuscle in children with CP is a significant compo-nent of physiotherapy intervention3 and may helpwith prevention of plantar flexor contractures. Inearly childhood, children with CP participate inregular rehabilitation sessions where ankle mobi-lization, stretching, strengthening, and gait train-ing are used to help encourage selective motorcontrol and decrease spasticity. School age andolder children are often provided with a home ex-ercise program and no regular therapy.

Unfortunately, children are often not compliantin following a conventional home exercise programbecause they find the exercise meaningless and un-interesting. Although there is minimal research re-garding the efficacy of a virtual reality trainingsession in children with cerebral palsy for motor re-habilitation,4 studies have demonstrated high lev-els of interest, fun and motivation with VR.4–6

1Rehabilitation Sciences, Health Sciences, University of Ottawa, Ottawa, Ontario, Canada.2Ottawa Children’s Treatment Centre, Ottawa, Ontario, Canada.

14246c03.pgs 4/14/06 10:03 AM Page 123

In this study, we asked children to completeankle selective motor control exercises in a conven-tional (Conv) manner and using a virtual reality(VR) system. We determined whether children withCP were able to appropriately interact with the sys-tem and recorded levels of interest and fun ex-pressed by the children and their parents duringthe VR and Conv exercises. We also recorded themovement kinematics as the children completedthe exercises.

METHODS

Participants

Ten children with CP (four male, six female) andsix children without CP (two male, four female),7–17 years old, participated (Table 1). The childrenwith CP included eight children with spastic hemi-plegia and two with spastic diplegia. Children withCP had Gross Motor Functional Classification Sys-tem (GMFCS) scores7 of 1 or 2, indicating indepen-dent ambulation with or without an assistivedevice.

Equipment

The VR system, Interactive Rehabilitation andExercise Systems, Ottawa, Ontario, Canada (IREX),consisted of a large television monitor, camera, andcomputer. The child sat in front of the camera that

recorded his or her image. The computer combinedthe child’s image with an exercise scenario, a gamein which the child could keep score. The child sawhis or her image as part of the scenario on the largemonitor and could interact with virtual objects inthe environment.

An electrogoniometer consisting of two metalarms and a parallelogram frame attached to a po-tentiometer was used to measure ankle joint rangeof motion. The electrogoniometer was calibrated atintervals of 5 degrees. The output was linear, andthe final electrogoniometer records of ankle jointrange of motion were converted to degrees usingthe slope of the calibration curve.

Exercises

Ankle dorsiflexion movements in chair-sittingand long-sitting were completed in both the Convand VR programs. In chair sitting, the child sat on astool with the hip, knee, and ankle as close to 90 de-grees as possible (Fig. 1A,B). In long sitting, thechild sat on the floor with the hip at 90 degrees,knee at 0 degrees, and ankle in a relaxed position(Fig. 1C,D). For both movements, the child was in-structed to dorsiflex the ankle to the end of theiravailable range, hold the maximal position for 3sec, relax, and then repeat. Depending on the de-gree of spasticity of the lower extremity, childrenwith CP were more or less able to attain the startingposition. The range of ankle dorsiflexion motion

124 BRYANTON ET AL.

TABLE 1. PARTICIPANT CHARACTERISTICS

Age Height Weight Affected GMFCS Home Program(yrs) Sex (cm) (kg) side score program adherence

9 F 132 25 Left Hemiparesis 1 Yes Minimal10 M 135 29 Left 1 Yes None17 M 163 71 Right 1 Yes None17 F 165 50 Left 2 Yes None

9 F 122 34 Right 2 Yes Minimal11 F 145 33 Right 2 Yes None15 M 170 77 Left 1 Yes Moderate11 F 145 33 Left 2 Yes Moderate14 M 165 59 Diplegia 2 Yes None16 F 168 68 Diplegia 2 Yes Moderate15 M 175 62 N/A N/A N/A N/A

7 M 134 25 N/A N/A N/A N/A10 M 139 26 N/A N/A N/A N/A13 M 152 50 N/A N/A N/A N/A11 F 150 36 N/A N/A N/A N/A

7 F 135 60 N/A N/A N/A N/A

14246c03.pgs 4/14/06 10:03 AM Page 124

was thus calculated from the initial position foreach child.

Two specific applications were created to elicitthe ankle movements in the virtual environment:Coconut Shooters and Ninja Flip. Coconut Shootersconsisted of a coconut being ejected from thechild’s toes. In Ninja Flip, a ninja appeared on thechild’s toe, cuing ankle dorsiflexion, and thenjumped onto one of four platforms. The degree ofankle dorsiflexion was modified on an individualbasis, allowing each child to reach the maximumscore. For example, for a child with CP and limitedankle dorsiflexion range, each degree of difficultymay have corresponded to a 2–3-degree move-ment, while for a child without CP each degree ofdifficulty may have corresponded to a 5-degreemovement.

Protocol

Each participant completed one 90-min exercisesession. Ankle dorsiflexion exercises were thencompleted in 10-min blocks with each block con-sisting of a set of Conv and a set of VR exercises. AnAB-BA design was used with exercise order coun-terbalanced between children. In each 10-minblock, the child completed 2 min of each exercise(chair- or long-sitting), with a 1-min rest period be-tween exercise type (Conv or VR). Children withCP used the most affected lower extremity, whereaschildren without CP used the preferred leg.

Outcome measures

Following each type of exercise, the child indi-cated how interesting and how fun the exercise was

using a visual analog scale (VAS). Once a childcompleted a set of VR and a set of Conv exercise,their parent completed similar VASs for their per-ception of their child’s fun and interest during eachexercise type. Comments were also recordedthroughout the exercise session.

Post-acquisition processing of the electrogo-niometer data provided starting and finishingankle position for each repetition, time to completeeach repetition, repetition hold time, and the num-ber of repetitions completed for each 2-min exercisebout.

RESULTS

Perceptions of exercise programs

All children responded with higher fun and in-terest (Fig. 2A,B) scores for VR than Conv exercises.All parents indicated their child had more fun andthat they would be more likely to do the VR thanConv exercises at home (Fig. 2C,D).

Movement kinematics

On average, significantly more repetitions of theConv than the VR exercises were completed by thechildren in both groups (p < 0.04, all comparisons).However, participants with and without CP took asignificantly longer average time to complete onerepetition of a VR exercise compared to the similarConv exercise (p < 0.01, all comparisons). Separateanalysis of hold time showed similarly longer aver-age times for the VR versus the Conv exercises (Fig.3A). When completing VR exercises, the children

VR COMPARED TO CONVENTIONAL EXERCISE IN CHILDREN WITH CP 125

A

C

B

D

FIG. 1. Pictures of chair and long-sitting in VR (A,C) and Conv (B,D) exercises.

14246c03.pgs 4/14/06 10:03 AM Page 125

126 BRYANTON ET AL.

Children with CP

0

2

4

6

8

10

Conventional VR Conventional VR

Block 1 Bloc k 2

VA

S (

cmChildren without CP

0

2

4

6

8

10

Conventional VR Conventional VR

Block 1 Block 2

VA

S (

cm

Was the exercise interesting?

A B

Did your child have fun?

0

2

4

6

8

10

Conventional Virtual Reality

VA

S (

cm

Would your child exercise at home?

0

2

4

6

8

10

Conventional Virtual Reality

VA

S (

cm

C D

FIG. 2. Scores on visual analog scales (VAS) from children (A,B) and parents (C,D). Note that the lines correspondto the scores of individual children and parents.

Repetition Hold Time

0

1

2

3

Chair Sit Long Sit Chair Sit Long Sit

Block 1 Block 2

Tim

e (s

)

CP Conventional CP VR Non CP Conventional Non CP VR

Range of Motion into Dorsiflexion

0

10

20

30

40

50

Chair Sit Long Sit Chair Sit Long Sit

Block 1 Block 2

Deg

rees

CP Conventional CP VR Non CP Conventional Non CP VR

A

B

FIG. 3. Mean repetition hold time (A) and mean range of motion into dorsiflexion (B).

14246c03.pgs 4/14/06 10:03 AM Page 126

had to maintain ankle dorsiflexion at the maximalposition in order to generate an action, such as theninja flipping, before starting the next repetitionand having a new ninja appear. When completingthe Conv exercise, there was no task-oriented in-centive other than verbal instruction to hold the ex-treme position. Most importantly, participants withand without CP recorded consistently albeit not al-ways significantly greater mean ankle active rangesof motion into dorsiflexion during VR versus Convexercise (Fig. 3B: children without CP, p < 0.03, allcomparisons; children with CP, p = 0.09). With theVR exercises, the children had a goal to attain (i.e.,scoring the greatest number of points), whereaswith the Conv exercises, children received no feed-back on ankle range.

DISCUSSION

Even the youngest children were able to com-plete the tasks within the virtual environment andgenerate movements necessary to interact with thevirtual objects. The children clearly expressedgreater interest and fun with the VR exercises com-menting that they were “a lot more fun” and“cool.” One child expressed excitement from scor-ing stating “Whoa, mommy, I got 300. Can I try itwith my left foot?” Less effusive comments wererecorded with the Conv exercises where childrenstated “Can I have a nap? I’m tired” and “These areboring.” Parents were equally positive with the VRexercises with statements “She was much more mo-tivated and interested in them as a game,” “muchmore motivated to try harder and score higher,” “Iknow the computer exercises would increase his in-terest and get him involved in doing his stretchingon a daily basis.” Similar experiences and demon-strations of enthusiasm have been reported foradults6 and children4 using similar technology.

Improvements in selective motor control havebeen reported in individuals with spastic CP fol-lowing different intervention paradigms, includ-ing technologies such as feedback training,8 whilesupplementing regular therapy with an intense pe-riod of increased exposure to physiotherapy hasbeen reported to accelerate acquisition of motorskills in some children with CP.9 The single sessionwith virtual reality clearly resulted in differentmotor patterns from the children when completingthe VR versus the regular exercises. It is unclearwhether an intervention trial using VR would re-sult in a change in selective motor control andwhether a change would be retained over an ex-tended period.

Importantly, the loss of selective motor controlmay interfere with overall level of functioning evenwhen other impairments are treated,10 since the un-derlying strength and coordination may be limited.There are no specific modalities to treat selectivemotor control but physical and occupational ther-apy in conjunction with a home program may im-prove selective motor control enough to affectfunctioning. Thus repetitive activities guided by atherapist or as in this case, a virtual environment,and a continuation of repetitive activities in dailyfunctioning may improve gross motor skills.

A main goal in therapy is the transfer of theacquired motor control from the training site to dailyliving activities. We have previously shown transferof acquired functional balance and mobility followinga VR-based balance intervention with community-living traumatic brain injury survivors11 as well ashealthy older adults (unpublished data). The currentdata do not allow us to address this question.

CONCLUSION

Children generate a greater range of ankle dorsi-flexion, demonstrate better control of active ankledorsiflexion movement, and report greater interestin doing the same exercise when delivered througha VR system than as a stand-alone exercise. Ournext series of experiments will characterize themuscle activity generated during the two exercisemodes as well as determine retention and transferof effects following an intervention trial.

ACKNOWLEDGMENTS

We thank the children and parents who partici-pated in this study. Funding support was providedby PREA, Ontario; NSERC, Canada Summer Re-search Fellowship (to C.B.); CIHR, Summer Re-search Fellowship (to J.B.); Career Scientist Award,MOHLTC, Ontario (to H.S.). The VR applicationswere generously supplied by IREX.

REFERENCES

1. Cerebral Palsy Association of Canada. Available at:<www.fcip.ca/CP/> accessed May 1, 2005.

2. Olney, S.J., & Wright, M.J. (2000). Cerebral palsy. In:Campbell, S.K. (ed.), Physical therapy for children.Philadelphia: Saunders, pp. 533–570.

3. Campbell, S.K. (2000). Physical therapy for children.Philadelphia: Saunders.

VR COMPARED TO CONVENTIONAL EXERCISE IN CHILDREN WITH CP 127

14246c03.pgs 4/14/06 10:03 AM Page 127

4. Reid, D.T. (2002). Benefits of a virtual play rehabilita-tion environment for children with cerebral palsy onperceptions of self-efficacy: a pilot study. Pediatric Re-habilitation 5:141–184.

5. Thornton, M., Marshall, S., McComas, J., et al. (2005).Benefits of activity and virtual reality based balanceexercise programs for adults with traumatic brain in-jury: perceptions of participants and their caregivers,Brain Injury 19:989–1000.

6. Weiss, P.L., Bialik, P., & Kizony, R. (2003). Virtual re-ality provides leisure time opportunities for youngadults with physical and intellectual disabilities. Cy-berPsychology & Behavior 6:335–342.

7. Palisano, R., Rosenbaumm, P., Walter, S., et al.(1997). Development and reliability of a system toclassify gross motor function in children with cere-bral palsy. Developmental Medicine & Child Neurology39:214–223.

8. Skrotzky, K., Gallenstein, J.S., & Osternig, L.R. (1978).Effects of electromyographic feedback training onmotor control in spastic cerebral palsy. Physical Ther-apy 58: 547–552.

9. Bower, E., & McLellan, D.L. (1992). Effect of in-creased exposure to physiotherapy on skill acquisi-tion of children with cerebral palsy. DevelopmentalMedicine & Child Neurology 34:25–39.

10. Gormley, M.E. (2001). Treatment of neuromuscularand musculoskeletal problems in cerebral palsy. Pe-diatric Rehabilitation 4:5–16.

11. Sveistrup, H., McComas, J., Thornton M., et al. (2003).Experimental studies of virtual reality–deliveredcompared to conventional exercise programs for re-habilitation. CyberPsychology & Behavior 6:245–249.

Address reprint requests to:Dr. H. Sveistrup

Rehabilitation SciencesHealth Sciences

University of Ottawa451 Smyth Rd.

Ottawa, ON, Canada K1H 8M5

E-mail: Heidi.Sveistrup@uottawa.ca

128 BRYANTON ET AL.

14246c03.pgs 4/14/06 10:03 AM Page 128