computer games and spinal injuries

Game-based Exercises for DynamicShort-Sitting Balance Rehabilitation ofPeople With Chronic Spinal Cord andTraumatic Brain InjuriesAimee L Betker, Ankur Desai, Cristabel Nett, Naaz Kapadia, Tony Szturm

Background and PurposeGoal-oriented, task-specific training has been shown to improve function; however,it can be difficult to maintain patient interest. This report describes a rehabilitationprotocol for the maintenance of balance in a short-sitting position following spinalcord and head injuries by use of a center-of-pressure–controlled video game–basedtool. The scientific justification for the selected treatment is discussed.

Case DescriptionsThree adults were treated: 1 young adult with spina bifida (T10 and L1–L2), 1middle-aged adult with complete paraplegia (complete lesion at T11–L1), and1 middle-aged adult with traumatic brain injury. All patients used wheelchairsfull-time.

OutcomesThe patients showed increased motivation to perform the game-based exercises andincreased dynamic short-sitting balance.

DiscussionThe patients exhibited increases in practice volume and attention span duringtraining with the game-based tool. In addition, they demonstrated substantial im-provements in dynamic balance control. These observations indicate that a videogame–based exercise approach can have a substantial positive effect by improvingdynamic short-sitting balance.

AL Betker, MSc, is a PhD candi-date, Department of Electrical andComputer Engineering, Universityof Manitoba, Winnipeg, Mani-toba, Canada.

A Desai, BPhysio, is a student inthe School of Medical Rehabilita-tion, University of Manitoba.

C Nett, BMR (PT), is a physicaltherapist in public practice and aclinical lecturer with the School ofMedical Rehabilitation, Universityof Manitoba.

N Kapadia, BPhysio, is a student inthe School of Medical Rehabilita-tion, University of Manitoba.

T Szturm, PhD, is Associate Profes-sor, Division of Physical Therapy,School of Medical Rehabilitation,University of Manitoba, R106-771McDermot Ave, Winnipeg, Mani-toba, Canada R3E 0T6. Address allcorrespondence to Dr Szturm at:[email protected].

[Betker AL, Desai A, Nett C, et al.Game-based exercises for dy-namic short-sitting balance re-habilitation of people with chronicspinal cord and traumatic brain in-juries. Phys Ther. 2007;87:1389–1398.]

© 2007 American Physical TherapyAssociation

Case Report

Post a Rapid Response orfind The Bottom Line:www.ptjournal.org

October 2007 Volume 87 Number 10 Physical Therapy f 1389

Balance of the human body re-quires timely control of the po-sition and motion of the center

of body mass relative to the base ofsupport. Maintaining balance in ashort-sitting position at rest; duringvoluntary head, arm, and body move-ments; and during transfers andwheelchair use (both indoors andoutdoors) involves many essentialsensory and motor processes. Feed-forward predictive controls, whichinitiate preparatory postural adjust-ments (goal-directed voluntarymovements), are required to main-tain balance during these move-ments and to anticipate potential fu-ture disturbances.1 Sensory feedbackprocesses are essential for respond-ing in a timely fashion to unexpecteddisturbances or to correct for move-ment errors.

Restoration and maintenance of in-dependent dynamic short-sitting bal-ance* are priorities for many peoplewho use wheelchairs because of aspinal cord lesion or an acquired ortraumatic brain injury. As in standingposture, poor balance in a short-sitting position will increase the fearof falling, fall risk, and mobility limi-tations, creating greater patient de-pendency in basic and instrumentalactivities of daily living. Poor posturealso can have an effect on a person’sself-confidence in dealing with otherpeople.2 In turn, these issues cancause reduced levels of physical ac-tivity, participation in sports, and,more generally, quality of life.

Evidence from human studies showsthat goal-oriented, task-specific train-ing improves function and that in-creased amounts of training producebetter outcomes3–5 (for a completereview, see Kwakkel6). One problemwith task-specific treatment ap-

proaches, however, is maintainingpeople’s interest in performing re-petitive tasks and ensuring that theycomplete the treatment program. Alack of interest or a short attentionspan also can impair the potentialeffectiveness of therapeutic exer-cises. Conversely, the use of reward-ing activities has been shown toimprove people’s motivation to prac-tice.7–11 Various approaches havebeen put forth to couple motivatingexperiences with rehabilitation exer-cises. Biofeedback, in which a bio-logical signal is recorded and pre-sented to people, has long been usedclinically to create and strengthenthe awareness of a given task orperformance.12–16

Novel and promising methods of ap-plying biofeedback to rehabilitationare virtual reality and videogames.17–22 In a study by Webster etal,23 a virtual environment was cre-ated to help people with the controland mobility of their wheelchairs,and participants had to navigatethrough a virtual obstacle course. Af-ter treatment, the participants exhib-ited a decrease in wheelchair acci-dents and falls and showed betterperformance on an actual obstaclecourse compared with subjects whodid not have training with the virtualcourse. Video games were used byO’Connor et al24 in an attempt toincrease the physiologic responsesof people using manual wheelchairsand to examine their effects on themotivation of the people to performtheir exercises. The GAMEWheelssystem interfaced commercial videogames with rollers, allowing station-ary propulsion of the wheelchairs.The observations showed that 87%of the subjects found that the gamesmotivated them to perform theirexercises.

On the basis of these ideas and re-sults, 3 interactive video game-basedexercises that are controlled by useof center-of-pressure (COP) signal

biofeedback were created. This in-teractive exercise tool has been ap-plied to standing balance.25,26 Betkeret al25 administered a questionnaireto 15 subjects (7 with balance disor-ders and 8 without balance disor-ders) after they played a 10-minutesession of each game. The resultswere encouraging. The subjects indi-cated that the games were challeng-ing and fun and would be a welcomeaddition to current treatment pro-grams. Subsequently, Betker et al26

reported on the feasibility and bene-fits of interactive standing balanceexercises carried out with the COP-controlled video game system for 3people who had chronic neurologi-cal deficits. The postexercise obser-vations demonstrated that the peo-ple exhibited few falls, decreasedCOP excursion limits for some tasks,and increased attention span duringtraining.

In our treatment program, COP-controlled video game-based exer-cises were used to attempt to im-prove dynamic short-sitting balancein people with central nervous sys-tem injuries. We thought that theinclusion of motivational and func-tional gaming in rehabilitation andsports training might increase thepeople’s desire to perform their ex-ercises and therefore result in im-proved dynamic balance control af-ter the exercises.

Case DescriptionsPatient HistoriesThree people consented to betreated and provided the followinginformation.

Patient 1 was a 26-year-old man withspina bifida (myelomeningocele) ex-tending from T10 to L1–L2 and re-sulting in complete paraplegia andpoorly developed lower extremities.At the time of initial assessment, hedemonstrated good static and dy-namic short-sitting balance and wasindependent with all transfers, activ-

* Defined as maintaining an upright positionof the torso while sitting on the buttocks orthighs (or both), with the shank hanging overthe sitting surface.

Game-based Exercises for People With Chronic Spinal Cord and Traumatic Brain Injuries

1390 f Physical Therapy Volume 87 Number 10 October 2007

ities of daily living, and work. As aperson who participated in Paralym-pic sports, he actively raced forTeam Canada and was actively train-ing to improve dynamic balance con-trol, an important requirement forhigh-speed wheelchair racing.

Patient 2 was a 52-year-old man withcomplete paraplegia (T11-L1) and atransfemoral amputation; these inju-ries resulted from a motor vehicleaccident 10 months before recruit-ment into our treatment program.After the accident, he received in-patient rehabilitation for 6 months.At the time of initial assessment, be-fore the current treatment program,he demonstrated complete motorand sensory loss below the T11 level,demonstrated dependent short-sitting balance (he sat with a ky-photic posture with bilateral upper-extremity support and was unable toperform any functional activity withthe upper extremities in an unsup-ported short-sitting position), and re-quired moderate assistance fromone person for transfers. His primarytreatment goal was to regain inde-pendent short-sitting balance for re-turn to office work.

Patient 3 was a 41-year-old man whohad had a severe traumatic brain in-jury more than 5 years before thecurrent treatment program. He hadreceived physical therapy interven-tion several times during those 5years for trunk and lower-extremitymotor control and balance re-education. His upper-limb functionwas good bilaterally, but he had poortrunk and lower-limb motor controland high muscle tone (velocity-dependent resistance to stretch),which fluctuated from extensor toneto flexor tone, depending on his po-sitioning. He had a progressiveplantar-flexion contracture of theright ankle secondary to spasticity(hypertonity of the plantar flexors).He was unable to maintain short-sitting balance without the use of his

hands for support because of im-paired balance and trunk control. Asa result (and because of his size), hehad to be transferred with a Hoyerlift. He used a powered wheelchairfor mobility indoors and outdoors.He had no sensory loss, and his in-tellectual and memory functions alsowere intact. However, he was easilydistracted during most activities andtherapy, requiring constant cuingand verbal commands to stay fo-cused on the task at hand.

COP-Controlled VideoGame–based Exercise ToolThe COP position signal has longbeen used as an indicator of balanceperformance.27–30 We developed theCOP-controlled video game-basedexercise tool for use with the Force-Sensitive Applications (FSA) soft-ware† and pressure mat.† The COPposition signal input is acquired via aflexible pressure mat measuring53 � 53 � 0.036 cm and containinga 16 � 16 grid of piezoelectricity-resistive sensors spaced 2.8575 cmapart (other mat sizes are available).The flexibility of the pressure matpermits games to be performed onsolid, fixed surfaces and allows pro-gression to compliant surfaces, withthe FSA pressure mat being placedbetween the patient and the surface.The position of the COP is calculatedfrom the pressures produced by thepatient seated on the pressure mat.This COP position signal then ismapped as an input to each of 3different games (Under Pressure,Memory Match, and Balloon Burst),which are described below.

In Under Pressure (Fig. 1a), gameplayers shift their weight to move areceptacle in order to “catch” an ob-ject. The game comprises 3 modes:in the horizontal mode, players mustshift their weight side to side; in thevertical mode, players must shift

their weight back and forth; and inboth modes together, players mustshift their weight in all directions.Thus, movement range and speed inall or targeted directions are exer-cised. Difficulty levels can be config-ured through the receptacle size, theobject speed, the number of objects,and the option of multiple objectsappearing at specified intervals.

In Memory Match (Fig. 1b), the goalis to select 2 matching cards from a3 � 3 or 4 � 4 array of squares.Game players select a card (square)by shifting their weight to move theon-screen COP indicator to 1 of the 9or 16 possible cards (squares). Oncethe COP is held still in a square for aduration selected by the player, thecard is revealed. The second cardthen is selected in a similar manner;if the cards match, they remain faceup. This process is repeated until allof the card pairs are selected. Diffi-culty levels can be configuredthrough the number of seconds theplayers have to select their cards andthe number of cards displayed (9 or16).

In Balloon Burst (Fig. 1c), a newlycreated game, the goal is to “pop”balloons. Stationary balloons appearat random locations on the screen.Game players must shift their weightin all directions in order to move theon-screen COP marker over the bal-loon to pop it. The difficulty levelcan be configured through the sizeof the balloon.

In order to allow a customized andgraded protocol for each player, theinteractive video game system of-fered the following features. The ad-justable difficulty levels within thegame software helped to ensure thateach player was competitive andcould successfully play the videogames while exercising his full rangeand speed of voluntary movement.This feature is important to prevent aplayer from becoming frustrated and

† Vista Medical, 3–55 Henlow Bay, Winnipeg,Manitoba, Canada R3Y 1G4.



quickly losing interest. The gamesoftware allows the player’s move-ment range to be determined dynam-ically or manually and can be scaled,allowing even people who are se-verely disabled to play and becompetitive.

Evaluations and OutcomeMeasuresTwo different test protocols wereused to obtain quantitative out-come measurements: (1) a question-naire that was administered after theexercises and (2) stability measure-ments that were obtained during aset of 6 tasks performed under 2conditions (before and after exer-cise). The 2 protocols are describedbelow.

Questionnaire. After exercise, aquestionnaire that included the fol-lowing questions was administered:Were the video game-based exer-cises fun to play? Did the video

games increase your motivation toperform your exercises? Were thevideo game-based exercises chal-lenging? Did the difficulty levels ofthe video games enhance the exer-cises? and Do you prefer videogame-based balance exercises to tra-ditional balance exercises? The re-sponse options were: “strongly dis-agree,” “disagree,” “agree,” and“strongly agree.”

Dynamic balance assessment. Inkeeping with the Sensory Organiza-tion Test concept, Shumway-Cookand Horak31 devised a clinical toolfor testing the sensory componentof balance: the Clinical Test of Sen-sory Interaction and Balance. In theClinical Test of Sensory Interactionand Balance, a compliant foam padis used as an unstable support baseto simulate the Sensory OrganizationTest in terms of somatosensory dis-tortion, with an added advantagethat it is not limited to the pitch

plane; the disturbance can be multi-directional.32 For the purpose of ourtreatment program, an air bladderwas used to distort and produce anunstable support surface, in a man-ner similar to the compliant foampad used during standing. The airbladder modified the surface reac-tion forces under the seat; thus, thesurface could not completely recip-rocate the normal forces beneath theseat as the center of body massmoved. The result was an increasein the magnitude and frequencyof involuntary (unpredictable) bodysway. To prevent a loss of balance,a fall, or both, an individual must beable to sense and respond to thiscondition. This condition consti-tutes a demand on whole-bodybalance reactions, and continuousautomatic postural adjustments arerequired to maintain upright short-sitting balance and postural stability.The degree of difficulty of the bal-ance tasks could be adjusted by se-

Figure 1.Screenshots of games. (a) Under Pressure during horizontal mode. The game player must move the flower under the bee. The totalnumber of bees, the number of bees caught, and mediolateral (ML) and anteroposterior (AP) movement ranges (in centimeters) aredisplayed. (b) Memory Match. The game player must select cards in order to find the pairs. The number of pairs found and the MLand AP movement ranges (in centimeters) are displayed. (c) Balloon Burst. The game player must move the cursor over the balloonto pop it. The total number of balloons, the number of balloons popped, and the ML and AP movement ranges (in centimeters) aredisplayed.



lecting different shapes and sizesfor the air bladder, just as differentthicknesses and densities could beselected for the foam pad used dur-ing standing.

A SwisDisk‡ was used for patients 2and 3, and a deflated (80%–90% ofthe air removed) yellow Physio Gym-nic§ ball (a more difficult and unsta-ble surface) was used to challengepatient 1. Patients were transferredfrom their wheelchairs to a low treat-ment plinth for all testing. To mini-mize any skin irritation during test-ing (and treatment), the patientswere seated on their regular seatcushions (foam-type cushions de-signed to help distribute forcesevenly, away from bony promi-nences, thus reducing the risk of ul-ceration). Each patient was in-structed to perform 6 tasks (Tab. 1),each 20 seconds in duration, under 2different conditions: first while sit-ting on their regular seat cushionsand then while sitting on the air blad-ders. Hand support was not permit-ted for this test.

The 4 movements (tasks 3–6) werepaced by the beat of a metronome,set to a frequency of 0.4 Hz. Thesemovements were selected becausethey represent important functionalactivities of daily living and work.The metronome frequency was se-lected to represent relatively slowself-paced movement speeds. For all6 tasks and both surfaces (cushionand air bladder), a fall was recordedif the patients could not maintain in-dependent balance for 20 seconds orif they could not perform the move-ments without holding on with theirhands. A physical therapist was po-sitioned directly behind the patientsto provide assistance, if needed.

InterventionAll treatments were performed at anoutpatient physical therapy clinicoperated by the Division of PhysicalTherapy, School of Medical Rehabil-itation, University of Manitoba; theprogram was designed partially forthe clinical training of undergraduatephysical therapist students under su-pervision. Each patient attendedtwelve 30- to 45-minute exercise ses-sions 2 or 3 times per week. Theexercise regimen consisted solely ofour COP-controlled video game-

based exercises; the patients did notreceive any other balance training orphysical therapy intervention duringthe treatment period. The patientswere transferred from their wheel-chairs to a low treatment plinth forall treatments.

Patient 1 played Under Pressure in allmodes only. Patient 2 played UnderPressure 80% of the time, playedMemory Match 19% of the time, andtried the new game, Balloon Burst,for the remaining 1% of the time.Patient 3 played Under Pressure 70%of the time and Memory Match forthe remainder of the time. Thegames were played with the patientssitting on the treatment plinth andprogressed (as appropriate) to sittingon a deflated Physio Gynmic ball orSwisDisk; the FSA pressure mat wasplaced between the patient and thesurface (Fig. 2). The ball or diskadded uncertainty to the system, as itwould randomly modify the surfacereaction forces; for people with sen-sation, it would distort or delay thepressure information from seat-to-surface contact.

As improvements in game playscores were noted and as improve-

‡ PI Professional Therapy Products Inc, POBox 1067, Athens, TN 37371.§ Ledraplastic Spa, Via Brigata Re 1, Osoppo,Udine, Italy 33010.

Table 1.Task Descriptions

Task Description

1 Maintain erect short-sitting balance with eyes open and looking straight ahead, as partof the CTSIB.a

2 Maintain erect short-sitting balance with eyes closed, as part of the CTSIB.

3 Perform rhythmic left and right horizontal head rotations to visual targets placed 120°apart.

4 Perform a rhythmic arm lifting and lowering task while holding a 50-cm lightweightwooden pole, 1.91 cm in diameter, with the hands kept shoulder width apart.Raise the pole to eye level and then back down to the legs, keeping elbowsextended.

5 Perform rhythmic left and right horizontal trunk rotations to approximately 30° ineach direction.

6 Perform rhythmic forward trunk bending and extension to return to the upright(erect) short-sitting position. The amplitude of trunk flexion should beapproximately 30°.

a CTSIB�Clinical Test of Sensory Interaction and Balance.



ments in balance and head-arm-trunkcontrol were observed, the treat-ment program progressed. In gen-eral, the minimum game play scorewas set at 50% success—for exam-ple, catching the object 50% of thetime in Under Pressure.

There were a number of game pa-rameters and task conditions thatcould be adjusted and modifiedwhen appropriate in order to permitthe treatment program to progressand to challenge the patients. Theseincluded the following 5 items:

1. A scaling factor was used to mapthe magnitude of COP excursion(movement range) to the excur-sion of the game cursor on the

computer display. Initially, theprogram was set so that a rela-tively small COP excursion pro-duced a moderate to large gamecursor movement. As game playscores improved and as balanceor trunk control improved, scal-ing was increased so that largerand larger COP excursions wererequired to move the gamecursor.

2. The speed of the game targets(objects) was adjusted. Initially,the speed was set to slow; thissetting permitted more time forthe patients to move and posi-tion the game cursor (COP) tocatch the object. Increasing thegame speed required faster COP

movement. Alternately, very slowspeeds required the patients tohold the COP position at the de-sired locations for longer periodsof time. For example, when a lat-eral or anterior trunk tilt was re-quired to catch the object at avery slow target speed, the pa-tient would have to hold the tiltedposition for a few seconds. Gamespeed was adjusted, scaling wasadjusted, or both as game playscores improved and as balanceimproved.

3. The exercise interval was in-creased by increasing the numberof game targets, that is, the num-ber of objects. Initially, the in-terval duration was set to be-

Figure 2.System setup. The patient sits on the pressure mat (1), which is connected to the laptop by the interface box (2). The laptop currentlydisplays the game Balloon Burst. The pressure mat is currently placed on top of the SwisDisk (3); the Physio Gymnic (4) ball also isdepicted.



tween 15 and 30 seconds ofgame play. As tolerated, this dura-tion was increased to 60 to 90seconds in order to increase thenumber of repetitions and tobuild endurance.

4. Reliance on hand support for bal-ance progressed to less reliance,from using both hands to usingone hand and then using no handsupport.

5. Air bladders were used to intro-duce a destabilizing compliantsupport surface. Once the pa-tients were able to play the gameswithout hand support, a compli-ant support surface was intro-duced. By changing the amountof air in the Physio Gymnic ball orSwisDisk, an appropriate traininglevel was achieved and progresswas made. For patients 2 and 3,within 3 treatment sessions, airbladders were being used for theentire treatment session. For pa-tient 1, an air bladder was usedimmediately, as this type of sup-port surface was required to chal-lenge his balance control.

OutcomesQuestionnaireThe questionnaire results were verypositive, with all patients answering“strongly agree” to all 5 questions.All of the patients indicated that theyenjoyed the video game-based tool,preferring it over exercise programsthat they had performed in the past,and indicated that they would like tocontinue the treatment. The adjust-able parameters and different modesof the tool offered sufficient diffi-culty levels; even patient 1, who par-ticipated in Paralympic sports, foundthe games to be challenging. In ad-dition, patient 2 particularly enjoyedthe new game, Balloon Burst, prefer-ring it over the other games.

Dynamic Balance AssessmentThe results of the dynamic balanceassessment are shown in Table 2.

Before exercise, patient 1 main-tained independent short-sitting bal-ance for the full 20 seconds duringall 6 tasks when he sat on his regularwheelchair cushion; in addition,short-sitting balance was maintainedfor the eyes-open, head rotation, andarm lifting tasks when he sat on thedeflated Physio Gymnic ball. How-ever, for 3 other conditions, whenpatient 1 sat on the deflated PhysioGymnic ball, he clearly lost short-sitting balance, and therapist inter-vention was required to prevent afall. After exercise, patient 1 main-tained independent short-sitting bal-ance for the full 20 seconds duringall 6 tasks on both surfaces.

For patient 2, 9 falls were recordedbefore exercise. Patient 2 was able tomaintain independent short-sittingbalance (without the use of hishands for support) only while sittingon the wheelchair cushion in theeyes-open, head rotation, and armlifting tasks. After exercise, patient 2was able to maintain independentshort-sitting balance for the full 20seconds during all tasks on both sur-faces (cushion and disk).

Before training with the COP-controlled video game-based system,patient 3 typically would attend onlyto balance exercises for 20 to 30 sec-onds at a time, with the training ses-sions typically lasting for only 10 to15 minutes. After practice with theCOP-controlled video game-basedsystem, patient 3 was able to main-tain concentration during the games(balance exercises) for up to 2 to 3minutes at a time and would repeatthis activity 10 to 15 times. The du-ration of the exercises increasedfrom short-interval training (approx-imately 20 seconds for 10–15 min-utes) to 2-minute interval training for20 to 30 minutes. Twelve falls were

recorded before exercise for patient3. In addition, hand support was re-quired during all 6 tasks on both sur-faces (cushion and disk). After exer-cise, patient 3 was able to maintainindependent short-sitting balance for20 seconds during all tasks on bothsurfaces.

DiscussionHere we report on the feasibility andbenefits of interactive COP-controlledvideo game-based exercises for short-sitting balance rehabilitation. Our ob-servations demonstrate that improvedrehabilitative interventions, which in-corporate a functional approach totraining and graded balance condi-tions or disturbances (ie, sensory feed-back and increased muscle activity),can produce substantial improve-ments in dynamic short-sitting bal-ance. Complete spinal cord lesionsbelow T10, T11, or T12 will abolishproprioceptive and cutaneous orpressure sensation in the hip jointsand in the pelvis structures andthereby will reduce the available spa-tial information, which is needed tomaintain short-sitting balance in theunsupported upright position. This ef-fect is amplified without vision—thatis, in dark or low-light conditions—and during sitting on different compli-ant surfaces. Learning a new balancesense is an important objective duringrehabilitation for people with com-plete thoracic spinal cord lesions andtraumatic brain injuries. Functionally,during game play, interactive move-ments are random, varying in direc-tion, amplitude, and precision; thus,during game play, people need tomake slow, maintained goal-directedmovements or quick shifts in the COPtrajectory. At moderate to high target(object) speed settings, these bodymovements require active mediolat-eral and anteroposterior weightshifts—for example, acceleration ofthe center of mass toward the in-tended target, followed quickly bybody deceleration to stop themovement.



The interactive gaming activities (ex-ercises) were designed around a flex-ible pressure mat for COP recording.This method allows training to beconducted on compliant or unevensurfaces; that is, the mat may beplaced on top of a compliant or ir-regular surface rather than on a forceplatform. The ability to apply agraded compliant support surface,along with the adjustable parametersof the tool, offers a variety of diffi-culty levels. For example, a deflatedPhysio Gymnic ball was required tochallenge the balance of patient 1,who is active in wheelchair racingand team sports. Similarly, a Swis-Disk was used to increase the bal-ance requirements of the exercisesfor patients 2 and 3. Thus, each gameand session could be enhanced tomeet the needs and performance lev-els of each patient. Such flexibilitycan better prepare people to interactand deal with more dynamic envi-ronmental conditions. Flexible pres-sure mats permit accurate COPrecording while eliminating the non-linear distortions and damping ef-

fects produced in the COP trajectoryby different materials.33

A main observation in this case re-port was that the interactive gamingintervention can motivate peoplewith chronic spinal cord and trau-matic brain injuries to practice dy-namic movement tasks. This ap-proach was applied effectively topeople with severe balance and mo-bility limitations and to an individualwho actively participated in sports.All 3 people indicated that they en-joyed the video game-based tool, pre-ferring it to normal treatment regi-mens, and that they would like tocontinue the treatment. These obser-vations showed that our COP-controlled video game system pro-vided a motivational and challengingenvironment. It has been shown thatwith the proper experiences and vol-ume of practice, the spinal cord canestablish new neuronal associationsand demonstrate functional improve-ments.34,35 One limitation of ourtreatment program is the potentiallybiased language in the questionnaire.

In future treatment programs, thequestions used to quantify the levelof motivation or fun during a partic-ular therapy program will be neutralin order to not lead or bias an indi-vidual’s responses.

Another observation was that afterexercise, all patients exhibited de-creased fall rates. In particular, afterexercise, patients 2 and 3 were ableto maintain independent short-sitting balance while performingmany demanding functional tasks.This observation is consistent withthe observation that intense practiceof a motor task following a completespinal cord lesion can result in sub-stantial functional improvements.

During game play, voluntary move-ments were generated in multipledirections and were varied in ampli-tude and speed. The patients pro-duced accurate targeted movements,were competitive at least 50% of thetime, and did not fall. It was evidentthat there was a temporary loss ofbalance and unwanted movements

Table 2.Dynamic Balance Assessment Resultsa

Surface Task Result for:

Patient 1 Patient 2 Patient 3

BeforeExercise

AfterExercise

BeforeExercise

AfterExercise

BeforeExercise

AfterExercise

Cushion 1 — — — — Fall —

2 — — Fall — Fall —

3 — — — — Fall —

4 — — — — Fall —



Air bladder 1 — — Fall — Fall —

2 Fall — Fall — Fall —





a Dashes indicate that no fall occurred.



(because of poor sensory control,motor control, or both and the effectof the compliant support surface);however, corrective balance reac-tions were generated successfully.Thus, both goal-directed voluntarymovements (feedforward control)and corrective balance reactions(feedback control) were evident dur-ing game play.

Like current biofeedback and virtualreality systems, the interactive videogame system provided the patientsand the therapist with instantaneousfeedback about performance andgoal attainment on a moment-to-moment basis. The patients and thetherapist were able to measure theirsuccessful progression to more com-plex tasks and support surfaces inreal time. Performance also could belogged on a trial-by-trial basis by useof the report feature. In future treat-ment programs, the functionality ofthe report feature of the video gamesystem will be expanded to includeadditional outcome measures detail-ing a patient’s performance.

Further motivation might beachieved through the developmentof a universal input device to allowthe pressure mat to be used withcommercial video games. This mod-ification will increase the selectionof games (an important factor inkeeping players motivated and inter-ested) and eliminate the cost of hav-ing to program new games. Masked,randomized clinical trials also are re-quired to confirm these preliminaryobservations and to provide a com-parison of the effects of this treat-ment with the effects of other, con-ventional therapies in parallel groupsof patients.

ConclusionsHere we report on the benefits ofour video game-based exercise regi-men. Our observations demon-strated that graded, dynamic balanceexercises on different surfaces can

be coupled effectively with videogame play and that this treatmentoffers the following values for reha-bilitation: goal-directed and intendedbehavior with random target pre-sentation and motion; the ability tomap small to large active COP excur-sions to game cursor excursion onthe computer display; the choice of awide range of game speeds andthus movement speeds; the abilityto select accuracy from small to largetarget (object) sizes; multitasking (in-corporation of gaze control [headand smooth pursuit], attention togame play strategy [target motionand prediction of final location],body movements, and balance con-trol); and rewards, with moment-to-moment feedback about goal attain-ment and positive reinforcement,both visual and audio. In addition tothe training program being enjoy-able, all 3 patients showed decreasedfall rates after the video game-basedexercise therapy. The portability ofthe system affords its use in moni-tored at-home programs, a featurethat makes this therapy approachcost-effective.

Ms Betker, Mr Desai, and Dr Szturm pro-vided concept/idea/project design and writ-ing. Mr Desai, Ms Nett, and Ms Kapadiaprovided data collection. Ms Betker, Mr De-sai, and Ms Kapadia provided data analysis.Ms Betker and Dr Szturm provided projectmanagement and facilities/equipment. MsNett and Dr Szturm provided patients. DrSzturm provided institutional liaisons. All au-thors provided consultation (including re-view of manuscript before submission).

This work was funded by a Manitoba HealthResearch Council Studentship and a NaturalSciences and Engineering Research CouncilFellowship.

This article was submitted August 10, 2006,and was accepted May 23, 2007.

DOI: 10.2522/ptj.20060229

References1 Huxham FE, Goldie PA, Patla AE. Theoret-

ical considerations in balance assessment.Aust J Physiother. 2001;47:89–100.

2 Douglas J. Wheelchair exercises for fitnessand confidence. Diabetes Self Manag.2005;22:47–48, 51–53, 55.

3 Remple MS, Bruneau RM, VandenBerg PM,et al. Sensitivity of cortical movement rep-resentations to motor experience: evi-dence that skill learning but not strengthtraining induces cortical reorganization.Behav Brain Res. 2001;123:133–141.

4 Garr JH, Shepherd RB. Neurological Reha-bilitation: Optimizing Motor Performance.Oxford, United Kingdom: Butterworth-Heinemann Ltd; 1998.

5 Richards CL, Malouin F, Bravo G, et al. Therole of technology in task-oriented trainingin persons with subacute stroke: a ran-domized controlled trial. NeurorehabilNeural Repair. 2004;18:199–211.

6 Kwakkel G. Impact of intensity of prac-tice after stroke: issues for consideration.Disabil Rehabil. 2006;28:823–830.

7 Cogan A, Madey J, Kaufman W, et al. Ponggame as a rehabilitation device. In: WarrenCG, ed. Fourth Annual Conference onSystems and Devices for the Disabled. Se-attle, Wash: University of WashingtonSchool of Medicine; 1977:187–188.

8 Nelson DL, Konosky K, Fleharty K, et al.The effects of an occupationally embed-ded exercise on bilaterally assisted supina-tion in persons with hemiplegia. Am J Oc-cup Ther. 1996;50:639–646.

9 Malone TW. Heuristics for designing en-joyable user interfaces: lessons fromcomputer games. In: Proceedings of the1982 Conference on Human Factors inComputing Systems. New York, NY: ACMPress; 1982:63–68.

10 Cunningham D, Krishack M. Virtual real-ity: a holistic approach to rehabilitation.Stud Health Technol Inform. 1999;62:90–93.

11 Tsang WW, Hui-Chan CW. Effects of exer-cise on joint sense and balance in elderlymen: Tai Chi versus golf. Med Sci SportsExerc. 2004;36:658–667.

12 Glanz M, Klawansky S, Chalmers T.Biofeedback therapy in stroke rehabilita-tion: a review. J R Soc Med. 1997;90:33–39.

13 Dozza M, Chiari L, Chan B, et al. Influenceof a portable audio-biofeedback device onstructural properties of postural sway.J Neuroengineering Rehabil. 2005;2:13.

14 Geiger RA, Allen JB, O’Keefe J, Hicks RR.Balance and mobility following stroke: ef-fects of physical therapy interventionswith and without biofeedback/forceplatetraining. Phys Ther. 2001;81:995–1005.

15 Bourbonnais D, Bilodeau S, Lepage Y,et al. Effect of force-feedback treatmentsin patients with chronic motor deficits af-ter a stroke. Am J Phys Med Rehabil.2002;81:890–897.

16 Yoo E-Y, Chung B-I. The effect of visualfeedback plus mental practice on symmet-rical weight-bearing training in peoplewith hemiparesis. Clin Rehabil. 2006;20:388–397.

17 Jack D, Boian R, Merians AS, et al. Virtualreality-enhanced stroke rehabilitation.IEEE Trans Neural Syst Rehabil Eng.2001;9:308–318.



18 Todorov E, Shadmehr R, Bizzi E. Aug-mented feedback presented in a virtualenvironment accelerates learning of a dif-ficult motor task. J Mot Behav. 1997;29:147–158.

19 Schultheis MT, Rizzo AA. The applicationof virtual reality technology for rehabilita-tion. Rehabil Psychol. 2001;46:296–311.

20 You SH, Jang SH, Kim YH, et al. Virtualreality-induced cortical reorganization andassociated locomotor recovery in chronicstroke: an experimenter-blind randomizedstudy. Stroke. 2005;36:1166–1171.

21 Fung J, Richards CL, Malouin F, et al.Treadmill and motion coupled virtual real-ity system for gait training post-stroke. Cy-berpsychol Behav. 2006;9:157–162.

22 Bryanton C, Bosse J, Brien M, et al. Feasi-bility, motivation, and selective motorcontrol: virtual reality compared to con-ventional home exercise in children withcerebral palsy. Cyberpsychol Behav.2006;9:123–128.

23 Webster JS, McFarland PT, Rapport LJ,et al. Computer-assisted training for im-proving wheelchair mobility in unilateralneglect patients. Arch Phys Med Rehabil.2001;82:769–775.

24 O’Connor TJ, Cooper RA, Fitzgerald SG,et al. Evaluation of a manual wheelchairinterface to computer games. Neuroreha-bil Neural Repair. 2000;14:21–31.

25 Betker AL, Szturm T, Moussavi Z. Develop-ment of an interactive motivating tool forrehabilitation movements. Conf Proc IEEEEng Med Biol Soc. 2005;3:2341–2344.

26 Betker AL, Szturm T, Moussavi ZK, Nett C.Video game-based exercises for balance re-habilitation: a single-subject design. ArchPhys Med Rehabil. 2006;87:1141–1149.

27 Baratto L, Morasso PG, Re C, et al. A newlook at posturographic analysis in the clin-ical context: sway-density versus other pa-rameterization techniques. Motor Control.2002;6:248–273.

28 Szturm T, Fallang B. Effects of varying ac-celeration of platform translation andtoes-up rotations on the pattern and mag-nitude of balance reactions in humans.J Vestib Res. 1998;8:381–397.

29 Norris JA, Marsh AP, Smith IJ, et al. Abilityof static and statistical mechanics posturo-graphic measures to distinguish betweenage and fall risk. J Biomech. 2005;38:1263–1272.

30 Collins JJ, De Luca CJ, Burrows A, et al.Age-related changes in open-loop andclosed-loop postural control mechanisms.Exp Brain Res. 1995;104:480–492.

31 Shumway-Cook A, Horak FB. Assessing theinfluence of sensory interaction of bal-ance: suggestion from the field. Phys Ther.1986;66:1548–1550.

32 Allum JH, Zamani F, Adkin AL, Ernst A.Differences between trunk sway charac-teristics on a foam support surface and onthe Equitest ankle-sway-referenced sup-port surface. Gait Posture. 2002;16:264–270.

33 Betker AL, Moussavi Z, Szturm T. On mod-eling center of foot pressure distortionthrough a medium. IEEE Trans BiomedEng. 2005;52:345–352.

34 Edgerton VR, Kim SJ, Ichiyama RM, et al.Rehabilitative therapies after spinal cordinjury. J Neurotrauma. 2006;23:560–570.

35 Harkema SJ. Neural plasticity after humanspinal cord injury: application of locomo-tor training to the rehabilitation of walk-ing. Neuroscientist. 2001;7:455–468.



Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

computer games and spinal injuries

Documents