Developing a Virtual Reality–Based Methodology forPeople with Dementia: A Feasibility Study

DARREN FLYNN, M.Sc.,1 PAUL VAN SCHAIK, Ph.D.,1 TIM BLACKMAN, Ph.D.,1CLIVE FEMCOTT, Ph.D.,2 BRIAN HOBBS, Ph.D.,3 and CARLOS CALDERON, Ph.D.2

ABSTRACT

The aim of this study was to examine the feasibility of virtual reality (VR) technology for useby persons with dementia (PWD). Data were obtained directly from six PWD regarding theirexperiences with a virtual environment (VE) of a large outdoor park. A user-centered methodwas developed to assess: (a) presence; (b) user inputs; (c) display quality; (d) simulation fi-delity; and (e) overall system usability. The extent to which PWD could perform four func-tional activities in the VE was also investigated (e.g., mailing a letter). In addition, physicaland psychological well-being of PWD while interacting with the VE was assessed objectivelyby recording heart rate during the VR sessions and subjectively with discrete questionnaireitems and real-time prompts. Symptom profiles associated with simulator sickness were as-sessed with an adapted version of the Simulator Sickness Questionnaire. The study foundthat PWD to some extent experienced presence; perceived that objects were realistic and movednaturally; generally felt in control of the interaction; and demonstrated little difficulty usinga joystick for navigation. The study also demonstrated that VR is an appropriate medium forassessing functional behavior within the context of an ecologically valid VE. PWD did not ex-perience any significant increase in symptoms associated with simulator sickness, or detri-ments to their psychological and physical well-being. These findings demonstrated that it isfeasible to work in VEs with PWD.

591

CYBERPSYCHOLOGY & BEHAVIOR

Volume 6, Number 6, 2003© Mary Ann Liebert, Inc.

INTRODUCTION

ADVANCES IN MEDICINE AND PUBLIC HEALTH haveachieved substantial increases in life expectancy,

which are likely to continue in the foreseeable fu-ture. However, this has also contributed to a con-comitant increase in the prevalence of dementia,which in developed countries is expected to doubleover the next 50 years.1 Internationally, the estimatedfigures are striking; for example, there are an esti-mated 700,000 and 2.3 million people with demen-tia (PWD) in the United Kingdom and the UnitedStates, respectively.2,3

Major symptoms of dementia that may be attrib-uted to both the underlying disease process and the

individual’s reaction to it include impairments ofmemory, language skills, attention, visual percep-tion and problem-solving skills, temporal and spa-tial disorientation, behavioral changes and losses ofsocial function, including the capacity to undertakeactivities of daily living.4 Dementia of the Alzhei-mer’s type accounts for approximately 60% of de-mentia cases diagnosed, characterized by a gradualonset and insidious decline over several years.5 Vas-cular dementia has a more stepwise progression andfrequently occurs after strokes. Dementia with Lewybodies is typically associated with more fluctuatingcognitive impairments and psychotic features suchas hallucinations. Other less prevalent causes of de-mentia include Korsakoff’s syndrome, Pick’s disease,

1School of Social Sciences and Law, 2School of Computing, 3School of Science and Technology, The University of Teesside, Mid-dlesbrough, United Kingdom.

Huntington’s disease, Parkinson’s disease, Creutz-feld Jakob Disease, AIDS, brain tumors, and headinjury. Comorbidity can include depression, agita-tion, sleep problems, aggression, inappropriate sex-ual behavior, and incontinence.

The development of a methodology for workingwith PWD in virtual environments (VEs) may enablethe development of virtual reality (VR)–based cog-nitive assessment techniques, cognitive rehabilita-tion strategies, and therapeutic activity for dementia.Design of both internal and external environmentsmay also benefit considerably from a VR-based ap-proach. How VR could make important contribu-tions in these areas is considered in the next section,followed by an introduction of the issues surround-ing the use of VR by PWD.

The potential of VR in dementia care

It is important that the presence of dementing ill-ness is discovered as early as possible, along withan identification of dementia sub-type and stage ofdisease, as this enables treatment and care to be ini-tiated that may have potential for minimizing theonset of neurodegeneration, optimizing cognitivefunctioning, and/or improving quality of life.6 Inorder to achieve this, cognitive, neuropsychologi-cal, and functional assessment tools that are sensi-tive, specific, reliable, and valid are needed.

However, existing “pencil and paper” cognitiveassessment tools have been heavily criticized forpossessing inadequate reliability and ecological va-lidity, as they are confounded by a respondent’sphysical capabilities, different testing environmentsand the quality of stimuli presented, and assess abil-ities in contexts detached from day-to-day function-ing.7,8 VR is a possible solution to these problemsbecause it has the potential to improve the reliabil-ity, sensitivity, specificity, and ecological validity ofcognitive assessment by enabling the precise con-trol and manipulation of stimuli presented to userswithin ecologically valid VEs that correspond toreal-life contexts.6,8–11

In addition, the neglected area of cognitive reha-bilitation strategies for dementia such as memoryskills training may be enhanced by utilizing ecolog-ically valid VEs.6,8,10 Cognitive rehabilitation strate-gies aimed at restricting neurodegeneration andmaintaining spared abilities have been given re-newed justification as a treatment option for earlystage dementia, as the new group of drugs called“acetylcholinesterase inhibitors” increase the pro-duction of acetylcholine needed for memory andlearning. The same VEs that are used to assess cog-nition could also be used as media for cognitive re-

habilitation, which could simulate a person’s homeor other familiar environment that is relevant to theindividual’s day-to-day functioning.

However, cognitive assessment and cognitive re-habilitation approaches may only be appropriate forpeople in the mild to moderate stages of dementia.People in the later stages of dementia may benefitmore from therapeutic activities such as multi-sensory therapy (MST) that are aimed at alleviatingthe more distressing psychological and behavioralsymptoms of dementia such as agitation. CAVE-based VEs could be used as an alternative to cur-rently available multi-sensory environments (MSEs)or Snoezelen, and this could improve the design,delivery, and outcome of MST for PWD in laterstages. MSEs utilizing VR technology have advan-tages over static environments as they are quicklyand easily modified, do not require the purchase ofadditional equipment, and have unlimited optionsin terms of the stimuli that may be presented tousers.

Design for dementia is another area where VRcould revolutionize dementia care and research. Thequality of design has a direct influence upon a per-son’s quality of life, and as people grow older, theybecome increasingly reliant upon their environmentto compensate for physical and cognitive decline.12

A substantial volume of research has investigatedindoor design features associated with improvementsin spatial orientation, vitality, wayfinding, and well-being.13–15 However, a major criticism of studies inthis area is that they often fail to justify why partic-ular environmental design features yield enhancedwell-being compared with others.16 This is primar-ily due to the potentially complex interactions be-tween elements in the environment and the costassociated with making real-world design changes.VR offers the possibility of elucidating these designfeatures by enabling the selective introduction andremoval of elements within the perceptual environ-ment to identify specific combinations of factors as-sociated with success in wayfinding and enhancedwell-being within settings such as residential care.

In the United Kingdom, an estimated 80% ofPWD are living at home in the community, almost aquarter on their own, and many are still active out-doors making trips for pleasure and visiting localamenities.17 Therefore, VR may have an importantcontribution to make in identifying barriers and fa-cilitators to wayfinding and well-being in outdoorenvironments as well as indoor care settings.18

VR has significant potential for improving thestate-of-the-art in the above areas; however, to dateonly two published studies investigating the use ofVR by PWD have appeared in the literature.19,20

592 FLYNN ET AL.

Furthermore, these studies do not address basicfeasibility issues necessary to enable the performanceof PWD in VEs to be deemed safe, reliable, and eco-logically valid.

Therefore, the aim of the current study was to ad-dress several of these feasibility issues in order toinform the development of VR applications. Sim-ply stated, it is unknown if PWD can navigatethrough a VE with standard input devices such asjoysticks or whether PWD are at risk of experienc-ing side effects such as simulator sickness or sufferdetriments to their psychological and physical well-being when interacting with VEs. In addition, it isunknown if PWD experience presence, which is astrong indicator of ecologically valid experiences inVEs. Indeed, one of the main advantages of VEs isthat they provide the opportunity to make an eco-logically valid assessment of behavior necessaryfor day-to-day functioning.8,10 However, to date, nostudy has demonstrated that PWD can perform func-tional tasks within the context of an ecologicallyvalid VE.

MATERIALS AND METHODS

Research design

The research made use of VR technology avail-able at the University of Teesside’s Innovation and

Virtual Reality Centre, a purpose-built resource forthe creation and development of real-time virtualexperiences in a wide range of application areas(http://vr.tees.ac.uk/). The project utilized the VRauditorium, where it is possible to run real-timemodels with a high degree of detail, powered by asilicon graphics infinite reality ONYX 1 computer.The VR Auditorium is comprised of a large 140-degree curved screen based on BARCO 1208 pro-jectors, which provides a semi-immersive view ofVEs with surround sound.

The study utilized a VE of a large park surroundedby a fence with a backdrop of local industry and ge-ographical features (Fig. 1). The VE included mod-els of park benches, telephone boxes, post boxes,trees, refuse bins, picnic areas, examples of locallyrelevant sculpture, and other perceptual opportu-nities.21 To avoid the disorientation that “walkingoff the end of the world” may cause and to supportecological validity, view boundaries were pro-grammed into the VE to prevent PWD from navi-gating beyond the main area of the park. The speedof navigation was restricted to a maximum of “nor-mal walking speed” at an elevation of 1.65 m dur-ing the VR exercises.

PWD were seated next to their carer/keyworkersand a research assistant (RA) throughout the VRsession, and postural demands were reduced by seat-ing participants in comfortable chairs during the

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 593

FIG. 1. Screendump of the VE used for the VR exercises.

VR exercises. Participants were situated 2.1 m fromthe center of the cinema screen. A standardized levelof illumination was used throughout.

The input device used for navigation was a BGSystems Flybox© joystick situated on a table directlyin front of the PWD. A directional template wasplaced over the joystick, which served as a memorycue for the participant while they were navigatingthrough the VE. The joystick enabled movementwith eight horizontal degrees of freedom (i.e., left,right, forward, back, and four diagonals).

Interactions between PWD, their carers/keywork-ers, the RA, and the VE during the VR session wereaudio and video recorded for the purposes of archiv-ing and analysis. To monitor the physical well-beingof PWD, an Ohmeda 3800 pulse oximeter was usedto measure their heart rate during the VR exercises.The device was attached to the forefinger of PWD,and data were recorded manually at 10-sec inter-vals during the VR exercises.

The incidence and severity of adverse effects wereassessed with a modified version of the SimulatorSickness Questionnaire developed by Kennedy etal.22 The adapted version of the simulator sicknessquestionnaire for PWD (SSQPWD) used the three-factor solution, which includes 12 items to assessclusters of symptoms associated with oculomotordisturbances (eyestrain, difficulty focusing, blurredvision, headache, and fatigue), disorientation (dizzi-ness with eyes open, dizziness with eyes closed, andvertigo) and nausea (nausea, stomach awareness,increased salivation, and burping).22 The SSQPWDwas administered to PWD and carers/keyworkers(who acted as controls) immediately before and afterthe VR session (see the Appendix for a list of 12SSQPWD items; a complete version is available fromthe authors upon request). The study sought to ob-tain data directly from PWD themselves using spe-cially designed (discrete) self-rating scales. Basedon previous research and the principles underpin-ning the use of “memory wallets” in dementia care,

the response format used for the SSQPWD was afive-point Likert scale anchored with pictures. PWDindicated their desired response by pointing to therelevant graded category or marking their own re-sponse depending upon their preference. Figure 2is an example of a SSQPWD item.

Recruitment of PWD and consent procedure

Following ethical approval, permission was soughtfrom local health services and the Alzheimer’s So-ciety to approach PWD and their carers who metthe following inclusion criteria: (a) a diagnosis ofprobable DAT in the early stages; (b) the PWD wasambulant and an active user of outdoor publicspaces; (c) no evidence of susceptibility to motionsickness; and (d) no history of epilepsy (includingphotosensitive epilepsy) or vertigo. Experience ofusing computers was not a prerequisite for partici-pation. Six PWD were recruited according to theabove criteria: three males and three females withan age range of 52–91 (Table 1).

The procedure for obtaining informed consentfrom PWD was taken very seriously, as ethical con-cerns are a major issue facing VR applications re-garding neurological conditions.8 A consent process,proposed by the Stirling Dementia Services Devel-opment Centre and elaborated by the School of Ar-chitecture at Oxford Brookes University, was usedas a guide to obtain written informed consent(www.brookes.ac.uk/schools/social/dementia/architect.html).

Stage 1: Involvement of carers and/or relatives.Carers and relatives of PWD meeting the above in-clusion criteria were given an information leafletthat detailed the rationale, aims, procedures, poten-tial for experiencing symptoms associated withsimulator sickness, and contact number for the RA,including details and photographs of the researchteam. Potential participants and carers/relatives re-

594 FLYNN ET AL.

FIG. 2. Example of a SSQPWD item.

TAB

LE

1.P

RO

FIL

EO

FT

HE

PW

D

Age

Age

at

Livi

ngM

MSE

Pre

viou

sH

ighe

st le

vel

Pre

viou

s V

EG

ende

r(y

ears

)re

ferr

alci

rcum

stan

ces

Han

dedn

ess

Med

icat

ion

scor

eoc

cupa

tion

of e

duca

tion

expe

rien

ce

Per

son

1M

ale

6058

Res

iden

tial

car

eR

igh

ta

aM

anu

alSc

hoo

lN

oP

erso

n 2

Fem

ale

74a

Wit

h s

pou

seR

igh

ta

aa

aN

oP

erso

n 3

Fem

ale

5250

Alo

ne

Rig

ht

Non

e12

Unk

now

nSc

hoo

lN

oP

erso

n 4

Mal

e83

aW

ith

sp

ouse

Rig

ht

Ari

cep

ta

aD

egre

eN

oP

erso

n 5

Mal

e58

57W

ith

sp

ouse

Rig

ht

aa

Skill

eda

Yes

Per

son

6Fe

mal

e91

aA

lon

eR

igh

tA

rice

pt

aa

aN

o

a Unk

now

n.M

MSE

, Min

i-M

enta

l Sta

te E

xam

inat

ion

.

ceived leaflets directly or via the mail system fromstaff belonging to health services and the Alzhei-mer’s Society. The information leaflet invited inter-ested recipients to contact the RA to arrange asuitable time to visit them in their homes or othersuitable venue to gain more detailed informationabout the study. We did not allow carers or rela-tives to give consent on behalf of PWD, but carersand relatives were closely involved in the consentprocess.

Stage 2: Explaining the study to the participant.The RA visited potential participants who expressedan interest at their homes or other suitable venue tofurther explain the study. At the visit, a video of theVR Auditorium that included examples of a VEwas shown to PWD and their carers, followed byan opportunity to ask questions about any aspect ofthe study. If in the opinion of the RA the PWD un-derstood the study, he/she was asked to sign a con-sent form. Informed consent was only recognized ifPWD responded to all the items on the consentform in the affirmative. If a PWD expressed signs ofunwillingness or concern about taking part in thestudy, this was accepted as a wish to not take partand he/she was not considered for participationany further. If carers or relatives had no concernsand were satisfied that their questions have beenanswered, they were asked to read and sign the“carers and relatives form” to confirm that they wereaware of the research aims and requirements, anddid not object to the PWD taking part in the study.

Stage 3: Continuing consent. This applied to allPWD who initially agreed to take part and involvedexplaining the study and obtaining written consentfor a second time immediately before the VR ses-sion. This ensured continued understanding and in-formed consent by the PWD. Again there was anopportunity to ask questions and participants wereinformed of their right to withdraw at any stagefrom the study without giving a reason.

Rose et al. stated that “side-effects do not appearto be a serious barrier to the use of VR in neurologi-cal rehabilitation; however, it is important to remainvigilant.”9 Therefore, our safeguards to reduce therisk of distress and discomfort as a result of simula-tor sickness involved establishing the psychologi-cal well-being of the PWD before the VR sessionsby asking them how comfortable they were feeling,and consulting carers. This established if PWD werehaving a “good” or “bad” day, and in cases wherePWD and/or a carers indicated that they were notfeeling “well,” then the study session was post-poned. PWD and carers were also screened for sus-

ceptibility to motion sickness and a history of vertigoand epilepsy, including photo-sensitive epilepsy.

The time spent interacting with the VE at any onetime was restricted to #20 min, which protectedagainst the increased risk of simulator sickness as-sociated with lengthy exposure periods.23 To avoidthe adverse effects associated with fast navigationand passive control of VEs,24 PWD were allowed tonavigate themselves through the VE with the joy-stick and were restricted to a maximum of normalwalking speed at an elevation of 1.65 m during theVR exercises. PWD and carers/keyworkers wereclosely monitored for signs of discomfort and dis-tress before, during, and after interacting with theVE by asking them how they were feeling.

At any time during the VR session, should PWDor carers/keyworkers display signs of simulatorsickness, distress, or discomfort, the session wasimmediately stopped. Members of the research teampresent during all VR sessions were in a position toprovide psychological support in the event of PWDand/or carers/keyworkers becoming distressed. Inaddition, the university’s nurse was available if re-quired and key workers were invited to the VR ses-sions to offer additional emotional support. Afterthe VR session, participants were escorted to a com-fortable room for refreshments, where they werefurther monitored for any adverse effects. Finally,to ensure the safety of the participants and theircarers, a member of the research team drove theparticipants home.

VR exercises

Each VR session lasted for approximately 50 minand PWD were asked to undertake two exercises,which included a 15-min, break after the first one.In a sensitive manner, it was explained to carers/keyworkers that they should refrain from answer-ing for PWD and influencing their navigationthrough the VE. However, they were encouraged tooffer reassurance and support to PWD in the eventof them experiencing difficulty or misunderstand-ing the instructions from the RA.

Exercise 1: Quality of the VE experience

The time allocated for this exercise was 20 min.The aims of this task were to determine if PWD: (a)experience presence in the VE; (b) can navigate inthe VE using a joystick; (c) perceive objects in theVE as realistic; (d) perceive objects in the VE asmoving in a ecologically valid fashion, and (e) feelin control of what they are doing. These aspectswere assessed with items adapted from the VRuse

596 FLYNN ET AL.

questionnaire.25 The VRuse in its original form as-sessed 10 factors (10 items for each factor) associ-ated with the usability of VR systems as a functionof the attitudes and perceptions of users regardingthe interface and method of interaction: functional-ity, user input, system output, user guidance andhelp, consistency, flexibility, simulation fidelity, errorcorrection/handling and robustness, presence, andoverall system usability. As PWD can become con-fused and distressed by excessive questioning26 anda requirement of the current study was that the in-teraction should be restricted to #20 min to avoidsimulator sickness, an abridged version of the VRusewas needed, referred to as the Dementia VRuse(DVRuse). Consequently, five of the 10 usabilityfactors were selected to construct the DVRuse: (a)presence, (b) user input (the extent they can movearound in the VE using the joystick), (c) system out-put (quality of the display), (d) simulation fidelity(whether the objects in the VE move realisticallyand naturally) and (e) overall system usability (feel-ing in control of their actions and enjoying their in-teractions with the VE). Two items—rather than thefull 10 from each of these five-factors—were usedto reduce the time required to administer the scale.Given that PWD have impairments of short-termmemory, the DVRuse was administered in real-timeat standardized intervals to overcome problems as-sociated with recall and “proxy assessments.”6

Each DVRuse item was printed on a separate A4card for presentation during the VR exercise andthe order of item presentation was identical foreach PWD. Text was printed using a font size of 48with black text on white background and with colorimages to compensate for the decline of visual acu-ity associated with ageing. The response formatused for the DVRuse was based on the same princi-ples as the SSQPWD, but a three-point pictorial re-sponse format was used as follows: (a) a personsmiling with thumbs up—representing “very muchso,” (b) an expression of puzzlement—representing“a little,” and (c) frowning with thumbs down—representing “not at all.”

At the beginning of exercise 2, the RA demon-strated to PWD how to navigate through the VEusing the joystick. The PWD was then given the op-portunity to practise navigating with the joystickand to ask questions about the nature of the exer-cise. All PWD started the interaction at the samepoint in the VE (the gate at the entrance to thepark), and were asked to walk through and explorethe park in any direction they wished. They werealso informed that they could stop to look at objectsin the VE and did not have to be constantly in mo-tion. However, if a participant remained stationary

for $30 sec, they were encouraged to explore theVE further with the prompt “where do you want togo you now?” Throughout the exercise, the RA andcarers/relatives or keyworkers engaged participantsin conversation about elements in the VE and offeredcontinued reassurance and encouragement as totheir performance.

Approximately every 90 sec during the interaction,the RA spoke aloud a DVRuse item. The RA thenheld the item printed on a card in the participant’sfield of view, minimizing information-processingdemands and reductions in presence that may becaused by divided attention between the card andthe VE. Participants were then asked to indicate theirdesired response to each item by pointing or sayingaloud one of the three graded categories. They werenot asked to stop interacting with the VE while re-sponding to these items. However, if they decidedto stop the interaction to respond to any of the items,their choice to do so was respected. PWD weregiven ample time to respond to each item and wereencouraged to carefully consider their response.However, if participants experienced difficulty bytaking $20 sec to respond, the RA further explainedthe item using appropriate frames of reference, andwith assistance from a carer and/or keyworkerwhere appropriate. In the event of participant stillhaving difficulty with an item, they were reassuredthat a response was not required and the item wasrecorded as missing data. This was done to preventPWD from becoming frustrated as a result of con-tinued failure and repetitive questioning. This pro-cedure was followed for each of the five-factors ofthe DVRuse questionnaire, but if the time spentinteracting with the VE exceeded 20 min and a par-ticipant had not responded to all 10 items the inter-action was stopped. During the 10-min break afterthis exercise, participants were monitored for signsof simulator sickness.

Exercise 2: Functional tasks

The time allocated for this exercise was 20 min.The aim of this exercise was to determine if PWDcould orientate and perform instrumental activitiesof daily living in the context of an ecologically validVE. The functional tasks were making a telephonecall, mailing a letter, disposing of litter, and findingan appropriate place to sit down and rest.

All participants began the interaction at the en-trance to the park. As in exercise 1, the RA initiatedthe exercise by demonstrating how to navigatethrough the VE using the joystick, and they weregiven the opportunity to practise using the joystickbefore the start of the exercise. Participants were

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 597

given complete freedom to explore, and they wereinformed that they could stop to look at objects inthe VE and did not have to be in motion constantlyalthough, as in exercise 1, if they remained station-ary for $20 sec, they were encouraged to explorethe VE further. At intervals of approximately every90 sec, the RA spoke aloud the instructions for thereal-time task and presented the participant with amemory reinforcer on an A4 card, which was heldwithin their field of view. The memory reinforcerreduced the need for additional prompting by theRA. The functional tasks were presented in the sameorder for all participants and each task was recordedeither as a success or a failure.

Participants were allocated ample time to explorethe VE to identify the target object associated withactions needed to perform each task. If a partici-pant gestured to, named, or was in close proximityto the target object in the VE, the task was recordedas a success. If after further explanation by the re-searcher and/or carer/keyworker a participant hadnot gestured to, or named, the target object alongwith the associated action, it was recorded as a fail-ure. Care was taken not to inadvertently reveal thenature of the target behavior associated with thefunctional task throughout this exercise. For exam-ple, for the task requiring a participant to identifyand navigate towards a telephone box in order tocontact their carer/keyworker, the words telephone(and telephone box) were avoided and the memoryreinforcer presented a person using a telephone ina home setting. Furthermore, instructions for eachtask were administered at locations in the VE wherethe target objects were outside their field of view.

RESULTS

Simulator sickness

The symptom profiles (as a function of symptomgroup and individual symptoms) obtained from theSSQPWD for PWD and their carers/keyworkers inboth sessions 1 and 2 (where applicable) are pre-sented in Tables 2 and 3, respectively. Mean scoresfor each individual symptom and symptom group(subscale) were calculated for each PWD and carer/keyworker in both sessions 1 and 2. Items withineach subscale of the SSQPWD were summed andcorrelated with each other. (Inferential statistics arepresented in each section using non-parametric tech-niques because of the small sample size, but thesemust be interpreted with caution, and the reader isinvited to make a personal judgement as to theirsignificance.) Cronbach’s alpha reliability coefficients

for each symptom group were not calculated asseveral items had zero variance.

An examination of the descriptive statistics inTable 2 revealed that overall mean scores at base-line for each symptom group appeared to have in-creased after the VR exercises in session 1. Persons2 and 3 reported an increase in oculomotor distur-bances, while persons 4 and 5 reported a decreasein levels of oculomotor disturbances. Persons 2 and4 reported increases in disorientation and nausea,respectively. Using Kendall’s tau, no significant as-sociations between symptom groups were reported.Wilcoxon signed ranks tests revealed no significantdifferences between overall levels of oculomotordisturbances, disorientation, and nausea assessedbefore and after the VR exercises for session 1.

Increases in overall mean scores for disorienta-tion and nausea appeared to occur after the VR ex-ercises in session 2, but oculomotor disturbancesdemonstrated a decrease. Similar to session 1, per-sons 1, 3, and 4 reported a decrease in oculomotordisturbances, whereas person 2 reported an increase.Persons 2 and 3 reported an increase in disorienta-tion. Only person 4 reported an increase in nausea.No significant associations between symptom groupswere found, there were no significant differencesbetween overall levels of oculomotor disturbances,disorientation, and nausea assessed before and afterthe VR exercises, and differences between sessions1 and 2 were not significant. The rank order of meanscores for each symptom group assessed after theVR exercises in both sessions were: oculomotor dis-turbances > nausea > disorientation. However, Fried-man tests revealed that these differences were notsignificant at alpha level 0.05. However, the in-crease and decrease in sessions 1 and 2, respectively,for oculomotor disturbances were approaching sig-nificance (z = 1.84, p = 0.07).

An examination of the descriptive statistics inTable 3 revealed that overall mean scores at base-line for each symptom group appeared to have in-creased after the VR exercises in session 1. Carers/keyworkers 2, 3, 4, and 5 reported an increase inoculomotor disturbances assessed after sessions 1and 2. In contrast to PWD, none of the carers/key-workers reported a decrease in levels of oculo-motor disturbances (with the exception of carer/keyworker 1 in session 2). Carers/keyworkers 2, 5,and 6 reported baseline increases in disorientation,and carers/keyworkers 2, 3, 4 (session 2 only), and5 reported a baseline increase in nausea. UsingKendall’s tau, no significant associations betweenoverall mean scores for each symptom group werereported in sessions 1 or 2. The overall mean scoresfor oculomotor disturbances increased significantly

598 FLYNN ET AL.

TAB

LE

2.SY

MP

TO

MP

RO

FIL

ES

OB

TAIN

ED

FRO

MT

HE

SSQ

PW

D F

OR

PW

D

Sym

ptom

gro

upIn

divi

dual

sym

ptom

Per

son

1P

erso

n 2

Per

son

3P

erso

n 4

Per

son

5aP

erso

n 6a

Mea

n

Ocu

lom

otor

Eye

stra

in1b

(1)c

2 (3

)1

(3)

2 (2

)1

(1)

1 (1

)1

(1)d

1 (2

)1

(1)

2 (1

)—

—D

iffi

cult

y fo

cusi

ng1

(1)

1 (1

)1

(1)

1 (1

)3

(2)

1 (1

)1

(1)

1 (1

)3

(1)

1 (1

)—

—B

lurr

ed v

isio

n2

(2)

1 (1

)1

(1)

2 (1

)2

(1)

1 (1

)3

(2)

1 (1

)2

(2)

1 (1

)—

—H

ead

ache

1 (1

)1

(2)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(2)

1 (1

)1

(1)

——

Fati

gue

1 (1

)1

(3)

3 (2

)1

(1)

1 (1

)1

(1)

2 (1

)1

(1)

2 (1

)2

(1)

——

Mea

n1.

2 (1

.2)

1.2

(2.0

)1.

4 (1

.6)

1.4

(1.2

)1.

6 (1

.2)

1.0

(1.0

)1.

3 (1

.4)

1.6

(1.2

)1.

0 (1

.4)

1.8

(1.2

)1.

4 (1

.0)

——

1.5

(1.2

)D

isor

ien

tati

onD

izzi

ness

(eye

s op

en)

1 (1

)1

(2)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(2)

1 (2

)1

(1)

——

Diz

zine

ss (e

yes

clos

ed)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

——

Ver

tigo

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(2)

1 (2

)1

(1)

——

Mea

n1.

0 (1

.0)

1.0

(1.3

)1.

0 (1

.0)

1.0

(1.0

)1.

0 (1

.0)

1.0

(1.0

)1.

0 (1

.1)

1.0

(1.0

)1.

0 (1

.6)

1.0

(1.7

)1.

0 (1

.0)

——

1.0

(1.3

)N

ause

aN

ause

a1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)—

—St

omac

h a

war

enes

s1

(2)

1 (1

)1

(1)

1 (2

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (2

)—

—In

crea

sed

sal

ivat

ion

2 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

——

Bu

rpin

g1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (3

)—

—M

ean

1.3

(1.3

)1.

0 (1

.0)

1.0

(1.0

)1.

0 (1

.3)

1.0

(1.0

)1.

0 (1

.0)

1.0

(1.1

)1.

0 (1

.0)

1.0

(1.0

)1.

0 (1

.0)

1.0

(1.8

)—

—1.

0 (1

.2)

a Att

end

ed f

or o

nly

on

e se

ssio

n.b B

efor

e.c A

fter

.dSe

ssio

n 2

.—

, mis

sin

g d

ata.

TAB

LE

3.SY

MP

TO

MP

RO

FIL

ES

OB

TAIN

ED

FRO

MT

HE

SSQ

PW

D F

OR

CA

RE

RS/

KE

YW

OR

KE

RS

Sym

ptom

gro

upIn

divi

dual

sym

ptom

Car

er 1

Car

er 2

aC

arer

3a

Car

er 4

bC

arer

5a

Car

er 6

aM

ean

Ocu

lom

otor

Eye

stra

in1c

(1)d

1 (1

)1

(2)

1 (2

)1

(1)

1 (1

)1

(1)c

——

1 (4

)—

—D

iffi

cult

y fo

cusi

ng

1 (1

)1

(1)

1 (1

)—

1 (1

)1

(1)

2 (1

)—

—1

(1)

——

Blu

rred

vis

ion

1 (1

)1

(2)

1 (1

)5

(5)

1 (1

)1

(1)

1 (1

)—

—1

(1)

——

Hea

dac

he1

(1)

1 (2

)1

(1)

1 (1

)1

(2)

1 (1

)1

(1)

——

1 (3

)—

—Fa

tigu

e1

(1)

1 (1

)2

(3)

1 (1

)1

(1)

1 (1

)1

(1)

——

1 (3

)—

—M

ean

1.0

(1.0

)1.

0 (1

.4)

1.2

(1.4

)1.

6 (1

.8)

1.0

(1.2

)1.

0 (1

.0)

1.2

(1.4

)1.

2 (1

.0)

——

1.0

(2.4

)—

—1.

1 (1

.7)

Dis

orie

nta

tion

Diz

zin

ess

(eye

s op

en)

1 (1

)1

(2)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)—

—1

(1)

——

Diz

zin

ess

(eye

s cl

osed

)1

(1)

1 (2

)1

(1)

1 (1

)1

(2)

1 (2

)1

(1)

——

1 (1

)—

—V

erti

go1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

——

1 (1

)—

—M

ean

1.0

(1.0

)1.

0 (1

.7)

1.0

(1.0

)1.

0 (1

.0)

1.0

(1.3

)1.

0 (1

.3)

1.0

(1.2

)1.

0 (1

.0)

——

1.0

(1.0

)—

—1.

0 (1

.0)

Nau

sea

Nau

sea

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)—

—1

(1)

——

Stom

ach

awar

enes

s1

(1)

1 (4

)1

(2)

1 (1

)1

(2)

1 (1

)1

(1)

——

1 (1

)—

—In

crea

sed

sal

ivat

ion

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)1

(1)

1 (1

)—

—1

(1)

——

Bu

rpin

g1

(1)

1 (1

)1

(3)

1 (1

)1

(1)

1 (1

)1

(1)

—1

(2)

——

Mea

n1.

0 (1

.0)

1.0

(1.8

)1.

0 (1

.8)

1.0

(1.0

)1.

0 (1

.3)

1.0

(1.0

)1.

0 (1

.3)

1.0

(1.0

)—

—1.

0 (1

.3)

——

1.0

(1.1

)

a Did

not

att

end

ses

sion

2.

b Th

is p

erso

n (s

essi

on 1

onl

y) h

ad s

erio

usl

y im

pai

red

vis

ion.

c Bef

ore.

dA

fter

.e S

essi

on 2

.

from baseline in session 1 (z = 2.01, p < 0.05). Therank order of mean scores for each symptom groupassessed after the VR exercises in both sessions were:oculomotor disturbances > nausea > disorientation.The differences between overall mean scores werenot significant at alpha level 0.05 and there were nosignificant differences between the changes frombaseline for each symptom group in sessions 1 or 2.

After the VR exercises in session 1, carers/key-workers reported a significantly higher level ofnausea than PWD (z = 2.27, p < 0.05). Furthermore,the higher level of disorientation for carers/key-workers than PWD after the VR exercises in session1 was approaching significance (z = 1.73, p = 0.08).Wilcoxon tests were conducted to examine the dif-ferences between PWD and carers/keyworkers as afunction of changes from baseline for each symp-tom group. No significant changes as a function ofsymptom group were reported in sessions 1 or 2;however, the greater change in disorientation forcarers/keyworkers than for PWD in session 1 wasapproaching significance (z = 1.73, p = 0.08).

Quality of the VE experience (PWD only)

Responses to the five-factors of the DVRuse inboth sessions 1 and 2 are presented in Table 4. Thesix PWD responded differently to the DVRuse items(Table 4). However, with the exception of person 6,who experienced difficulty with several of theDVRuse items (recorded as missing data), all thePWD had an overall score of $20 (i.e., 67% of maxi-mum) on the DVRuse.

The modal response to the DVRuse items wasmaximal (i.e., yes, very much), accounting for 60%of the 104 responses given in sessions 1 and 2. AllPWD to some extent reported feeling “present”within the VE (presence); reported being able tonavigate through the VE (user input); reported thatobjects in the VE appeared realistic and moved innatural fashion (system output and simulation fi-delity, respectively); and were in control and enjoy-ing their experience with the VE (usability). Onlyperson one stated a negative response to the VRuseitems, stating “no” in response to both of the simu-lation fidelity subscale items in session 1. Friedmantests revealed that differences in mean scores onDVRuse subscales within sessions 1 and 2 were notsignificant.

Functional tasks (PWD only)

The time taken to complete each of the four func-tional tasks (telephone call, mailing a letter, dispos-ing of litter, and finding an appropriate place to

rest) is presented in Table 5. In both sessions, all thePWD successfully completed all four of the func-tional tasks. The time taken to complete the tasksdemonstrated substantial variability between par-ticipants, ranging from 142 to 614 sec. In terms ofthe rank order of mean time in seconds for eachfunctional task, no consistent pattern emerged, al-though in both sessions, the highest mean time inseconds was reported for disposing of litter (task3). Another finding was that, with the exception oftask 2, mean times for task completion in session 2appeared to be higher than session 1.

Psychological well-being

Data on psychological well-being of PWD andcarers/keyworkers assessed before and after the VRexercises in both sessions 1 and 2, where applica-ble, are presented in Table 6. All PWD rated theirwell-being as maximal (i.e., very comfortable) priorto undertaking the VR exercises in session 1, withthe exception of person three who reported a me-dian response. Immediately after the exercises insession 1, five of the PWD reported no baselinechanges in well-being. Person 6 reported a decreasein well-being after the exercises in session 1, althoughthe decrease was negligible.

Four PWD attended for a second session, and tworeported no changes in well-being from baseline lev-els after the VR exercises; however, persons 1 and 2reported an increase and decrease, respectively, inbaseline levels of well-being. Kendall’s tau tests re-vealed that the association between well-being as-sessed before and after the exercises in session 1 wasapproaching significance (tau = 0.75, p = 0.08). Nosignificant association between well-being assessedbefore and after the exercises in session 2 was found.Wilcoxon signed ranks tests revealed no significantdifferences in self-rated well-being assessed beforeand after the VR exercises in sessions 1 or 2.

Four of the six carers/keyworkers reported theirwell-being as maximal immediately before the VRexercises in session 1, with carers/keyworkers 1and 2 reporting less than optimal levels of well-being. Immediately after the exercises in session 1,four carers reported no change in baseline levels ofwell-being; however, carers 2 and 5 reported a de-crease and increase respectively immediately afterthe exercises.

Only two carers/keyworkers attended for a sec-ond session, and they reported a negligible increaseand decrease respectively from baseline levels ofwell-being. Associations between ratings of well-being assessed before and after the VR exercises insessions 1 and 2 were not significant. Wilcoxon tests

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 601

TAB

LE

4.R

ESP

ON

SES

TOT

HE

DV

RU

SEQ

UE

STIO

NN

AIR

EU

SED

INE

XE

RC

ISE

1 FO

RE

AC

HSU

BSC

AL

EA

ND

IND

IVID

UA

LIT

EM

Subs

cale

Item

Per

son

1P

erso

n 2

Per

son

3P

erso

n 4

Per

son

5cP

erso

n 6c

Mea

n

Pre

sen

ceD

o yo

u fe

el th

at y

ou a

re r

eally

wal

king

thro

ugh

this

par

k?2a

32

32

32b

33

3—

—D

o yo

u fe

el a

s if

you

are

act

ual

ly th

ere

in th

e p

ark?

13

23

2—

33

23

——

Tota

l3

64

62

32.

55

65

6—

—2.

8U

ser

inp

ut

Are

you

find

ing

it e

asy

to m

ove

arou

nd in

the

par

k?3

22

32

—3

33

3—

—A

re y

ou fi

ndin

g th

e jo

ysti

ck e

asy

to u

se?

23

23

3—

32

33

——

Tota

l5

54

65

—2.

56

56

6—

—2.

9S

yste

m o

utp

ut

Do

the

tree

s in

the

park

look

rea

listi

c?3

32

33

32

33

3—

—D

o th

e ob

ject

s in

the

park

look

rea

l?3

33

32

33

23

3—

—To

tal

66

56

53

2.8

55

66

——

2.8

Sim

ula

tion

fid

elit

yA

re th

e tr

ees

and

oth

er o

bjec

ts in

the

par

k m

ovin

g1

32

32

—in

a n

atu

ral w

ay?

13

23

——

Are

the

thin

gs in

the

par

k m

ovin

g as

if y

ou w

ere

13

23

33

real

ly w

alki

ng p

ast t

hem

?3

33

3—

—To

tal

26

46

53

2.4

46

56

——

2.6

Usa

bil

ity

Do

you

feel

in c

ontr

ol o

f wha

t you

are

doi

ng?

33

23

32

2—

23

——

Are

you

enj

oyin

g w

alki

ng th

rou

gh th

e p

ark?

23

23

33

1—

33

——

Tota

l5

64

66

32.

73

—5

6—

—2.

0O

vera

ll Sc

ore

2029

2130

2312

22.5

2322

2730

——

25.5

2.6

2.6

a Ses

sion

1.

b Ses

sion

2.

c Att

end

ed fo

r on

ly o

ne s

essi

on.

—, m

issi

ng d

ata.

1 =

no.

2 =

a li

ttle

. 3 =

yes

ver

y m

uch

.A

dap

ted

from

Kal

awsk

y, R

.S. (

1999

). V

RU

SE—

a co

mpu

teri

sed

dia

gnos

tic

tool

for

usa

bilit

y ev

alu

atio

n of

vir

tual

/sy

nthe

tic

envi

ronm

ent s

yste

ms.

App

lied

Er-

gono

mic

s,30

:11–

25.

TAB

LE

5.T

IME

TAK

EN

TO

CO

MP

LE

TE

TH

EFU

NC

TIO

NA

LTA

SKS

INE

XE

RC

ISE

2

Task

1:

Task

2:

Task

3:

Task

4:

tele

phon

e ca

ll,m

ailin

g a

lett

er,

disp

osin

g of

litt

erpl

ace

to r

est,

Per

son

Sess

ion

tim

e ta

ken

(sec

)ti

me

take

n (s

ec)

tim

e ta

ken

(sec

)ti

me

take

n (s

ec)

Tota

l

11

1912

110

414

52

121

9325

185

550

21

2811

327

817

62

2220

4060

142

31

140

105

109

6141

52

233

6319

212

661

44

191

3724

541

414

287

3316

533

318

5a1

8115

033

5431

82

——

——

—M

ean

172

8310

534

492

116

5216

276

102

294

406

a Att

end

ed f

or o

nly

on

e se

ssio

n.A

ll P

WD

su

cces

sfu

lly

com

ple

ted

all

the

task

s by

nav

igat

ing

tow

ard

s ta

rget

s ob

ject

s in

bot

h s

essi

ons.

Dat

a is

not

sh

own

for

per

son

6, w

ho w

asn

avig

ated

thro

ugh

the

VE

by

the

RA

.

revealed no significant differences in well-being as-sessed before and after the VR exercises in sessions1 and 2.

An examination of the descriptive data in Table 6reveals that levels of well-being assessed immedi-ately before and after the VR exercises between PWDand carers were similar. No significant differencesin well-being assessed before and after the VR exer-cises were reported between PWD and carers/key-workers in both sessions.

Physical well-being (PWD only)

The heart rate of the PWD in beats per minute(bpm) measured at 10-sec intervals during exercises1 and 2 are presented in Tables 7 and 8, respectively.Heart rate data were not collected for person 6, as itwas felt that, due to the difficulties encountered inthe first session, a further session may have beendetrimental to the person’s well-being. An exami-nation of the descriptive statistics in Table 7 revealedthat heart rate appeared to increase and decreasefrom baseline levels during exercise 1, althoughmean heart rate for persons 1, 3, 4, and 5 did not de-viate substantially from baseline and mean levels.Indeed, the maximum increase in heart rate frombaseline was 10 bpm for person 5. However, heartrate for person 2 demonstrated a substantial andrapid increase from baseline reaching more thandouble the minimum value within a 20-sec interval,which precipitated exercise 1 being stopped for safetyreasons. Subsequent discussions with person tworevealed no physical symptoms (e.g., chest pain ornausea), and she intimated that she had becomefrustrated by searching for objects in the VE that shehad remembered from the previous session.

Heart rate data were recorded for each of thefour functional tasks separately to enable compar-isons to be made between individual tasks (Table8). Negligible changes in heart rate across tasks foreach of the PWD were reported, with mean heartrate remaining relatively stable.

DISCUSSION

The findings of the current study revealed thatPWD did not experience any significant increase insymptoms associated with simulator sickness ordecreases in psychological and physical well-beingas a result of exposure to the VE. Responses to theDVRuse indicated that to some extent PWD experi-enced presence; they perceived that objects in theVE were realistic and moved naturally; they reportedlittle difficulty using a joystick for navigating throughthe VE; they reported feeling in control of the inter-action; and they enjoyed their experience with theVE. All the functional tasks were successfully per-formed by PWD, which suggested that VEs are ap-propriate media for assessing behavior andcognition necessary for day-to-day functioning.

Simulator sickness was assessed with the SSQPWDto obtain self-ratings of three symptom groups: ocu-lomotor disturbances, disorientation, and nausea.The SSQPWD was well-received by both PWD and

604 FLYNN ET AL.

TABLE 6. PSYCHOLOGICAL WELL-BEING BEFORE AND

AFTER VR SESSIONS (5-POINT SCALE, 1 = VERY

COMFORTABLE, TO 5 VERY UNCOMFORTABLE)

Before After

Session 1Person 1 1 1Carer/keyworker 1 2 2Person 2 1 1Carer/keyworker 2 1 3Person 3 3 3Carer/keyworker 3 1 1Person 4 1 1Carer/keyworker 4 1 1Person 5 1 1Carer/keyworker 5 2 1Person 6 1 2Carer/keyworker 6 1 1

Session 2Person 1 2 1Carer/keyworker 1 2 1Person 2a 1 3Carer/keyworker 2 — —Person 3a 3 3Carer/keyworker 3 — —Person 4b 1 1Carer/keyworker 4 1 2Person 5c — —Carer/keyworker 5 — —Person 6c — —Carer/keyworker 6 — —

Overall Session 1Persons with Dementia 1.3 1.5Carer/keyworkers 1.3 1.5

Overall Session 2Persons with Dementia 2.0 2.0Carer/keyworkers 1.5 1.5

aAttended for only one session.bAttended with two different carers.cAttended session 2 alone.

carers/keyworkers. PWD demonstrated little diffi-culty understanding the items and the response for-mat that combined words, numbers, and pictures.They were given a choice of pointing to or markingtheir desired response to the SSQPWD items, al-though the majority opted to point at the relevantgraded category that was representative of theirstatus, along with a verbal confirmation such as“yes that one.” This is consistent with previous re-search that reported the successful use of self-ratingscales with word, pictorial, and number responseformats for use by older people with cognitive im-pairments.27–29

Responses to the SSQPWD demonstrated thatPWD did not experience any significant increase inoculomotor disturbances, disorientation, or nauseaafter exposure to the VE. Furthermore, they reporteda decrease in oculomotor disturbances between thefirst and second VR session that was approachingsignificance. In contrast to PWD, carers/keyworkers,who acted as controls, reported a significant in-crease in oculomotor disturbances after the VR ex-ercises and a greater increase in disorientation, whichwas approaching significance in the first VR session.These findings are consistent with previous researchon simulator sickness among healthy volunteers re-porting that active navigation of VEs, in our studyby PWD, decreases the likelihood of experiencingsimulator sickness.24 In addition, a decrease in sim-ulator sickness with “habituation” to the VE is alsoconsistent with previous research using non-neuro-logical populations.30 Higher levels of simulatorsickness in carers/keyworkers than in PWD mayindicate reduced presence. This is likely to have beencaused by a lack of interaction with the VE, whichis required for the development of a mental modelof the virtual space, which forms the basis of pres-ence.31 Thus, simulator sickness is not a significantbarrier to working with PWD in VEs; PWD are af-fected by exposure to VEs in comparable ways topeople without cognitive impairments.

The negligible influence of simulator sicknessupon PWD in the current study was arguably theresult of using a semi-immersive system and adher-ing to safeguards such as restricting the speed andduration of the interaction.23,32–34 However, despitethe encouraging findings for PWD, simulator sick-ness may constitute a barrier to the involvement ofcarers/keyworkers in VR-based approaches to de-mentia care and research. This has important impli-cations for working with PWD in VEs, in particularas involvement of carers will invariably be a strongpredictor of participation by PWD. As observed inthe current study, the involvement of carers pro-vided an invaluable source of social support thatserved to reduce anxiety and enhance motivationof PWD during the VR exercises. The problems arethat carers who experience simulator sickness maybecome distracted by feelings of discomfort that willimpact negatively upon their ability to offer socialsupport, and secondly, they are likely to avoid at-tending VR sessions and consenting to their depen-dents being exposed to VEs. Involvement of carersin non-VR treatments is reported to enhance psy-chosocial interventions such as memory retraining,4and it is likely that these benefits would be trans-ferred to VR-based psychosocial interventions forPWD. Therefore, as a precautionary measure, werecommend that both PWD and carers/keyworkersshould be monitored for signs of simulator sicknessbefore, during, and after interactions with VEs.

Quality of the VE experience

The subjective views of PWD regarding the inter-face and method of interaction were assessed inreal-time with the DVRuse. With the exception ofperson 6, the DVRuse was also well-received, withitems and response format presenting no difficul-ties in terms of understanding. However, several par-ticipants stopped navigating in the VE in order torespond to many of the items, which demonstrated

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 605

TABLE 7. PHYSICAL WELL-BEING (HEART RATE IN BEATS PER MINUTE) DURING EXERCISE 1

Person Baselinea Range Mean (SD) S2 Maximum/minimum

1 74 72–79 75.8 (1.6) 2.5 1.12 80 65–140 90.7 (18.6) 344.6 2.23 73 69–78 73.3 (2.1) 4.3 1.14 60 53–69 61.8 (3.3) 10.9 1.35 72 66–82 70.2 (2.9) 8.6 1.2

Overall 71.8 74.4 (5.7) 74.2 1.4

aInitial value.

TAB

LE

8.P

HY

SIC

AL

WE

LL-B

EIN

G(H

EA

RT

RA

TE

INB

EA

TS

PE

RM

INU

TE) D

UR

ING

EX

ER

CIS

E2

Task

1:

Task

2:

Task

3:

Task

4:

tele

phon

e ca

llm

ailin

g a

lett

erdi

spos

ing

of li

tter

plac

e to

res

t

Bas

elin

eaM

ean

S2B

asel

inea

Mea

nS2

Bas

elin

eaM

ean

S2B

asel

inea

Mea

nS2

Per

son

(ran

ge)

(SD

)(m

ax/m

in)

(ran

ge)

(SD

)(m

ax/m

in)

(ran

ge)

(SD

)(m

ax/m

in)

(ran

ge)

(SD

)(m

ax/m

in)

171

(69–

73)

70.8

(1.3

)1.

8 (1

.1)

73 (7

1–75

)73

.3 (1

.3)

1.8

(1.1

)73

(72–

76)

74.1

(1.2

)1.

5 (1

.1)

74 (7

2–75

)73

.9 (1

.2)

1.6

(1.0

)2

60 (6

0–61

)60

.5 (0

.7)

0.5

(1.0

)66

(66–

67)

66.5

(0.7

)0.

5 (1

.0)

65 (6

5–68

)67

.0 (1

.4)

2.0

(1.1

)66

(63–

69)

66.5

(2.3

)5.

5 (1

.1)

372

(67–

75)

71.3

(2.5

)6.

5 (1

.1)

74 (7

1–74

)72

.3 (1

.0)

1.1

(1.0

)74

(70–

76)

72.8

(1.4

)2.

0 (1

.1)

75 (7

0–75

)71

.8 (1

.4)

2.0

(1.1

)4

58 (5

6–60

)58

.5 (1

.4)

2.0

(1.1

)58

(57–

58)

57.7

(0.6

)0.

3 (1

.0)

60 (5

4–61

)58

.0 (2

.3)

5.4

(1.1

)58

(57–

59)

58.0

(1.0

)1.

0 (1

.0)

566

(66–

70)

67.9

(1.2

)1.

6 (1

.1)

70 (6

9–72

)70

.7 (1

.3)

1.6

(1.0

)70

(70)

70.0

(0.0

)0.

0 (1

.0)

69 (6

9–73

)71

.6 (1

.9)

3.8

(1.1

)O

vera

ll65

.465

.8 (1

.4)

2.5

(1.1

)68

.268

.1 (1

.0)

1.1

(1.0

)68

.468

.4 (1

.3)

2.18

(1.1

)68

.468

.4 (1

.6)

2.78

(1.

1)

a In

itia

l val

ue.

that presenting items during the interaction causeddivided attention between the VE and DVRuse items.

Reponses given by PWD to the DVRuse revealedthat the VE interface was of a sufficient quality tocreate a sense of involvement or “being there” in thepark (presence) and a perception of realism (systemoutput and simulation fidelity). This is important asa sense of presence is an important predictor forecological validity of user experiences within VEs.35

Although presence is not an essential criterionfor producing ecologically valid user experienceswithin VEs, arguably a user’s perception of realismis important for ecological validity.36 In addition, anatural method of interaction that is under the user’scontrol is also important for ecological validity and,with the exception of person 6, participants reportedthey were in control of the interaction with the VEand that it was an enjoyable experience. None ofthe participants used the template marked with di-rectional arrows as a memory cue to facilitate theuse of the joystick, even though they had no previ-ous experience of joysticks, with exception of per-son 5, who had limited experience playing videogames. The high usability of the joystick is an im-portant result as without appropriate methods ofnavigation in VEs, PWD will not be able to experi-ence direct interaction, a core characteristic of VRthat enables users to undertake a self-directed jour-ney or activity in a VE.37 Without appropriate meth-ods of interaction, only the development of VR-basedtherapeutic immersion for people is likely to havepotential.

With the exception of person 6, who did not takepart in exercise 2, all PWD successfully completedall of the functional tasks. This provided objectiveevidence that, with an appropriate interface andmethod of interaction, it is possible for PWD to suc-cessfully wayfind in the context of an ecologicallyvalid VE. However, given that the VE was medium-scale, it is possible that success at wayfinding wasdue to exhaustive searching of the VE until an ob-ject was found, which may occur in VEs that areunfamiliar to users.38

Psychological and physical well-being

No significant decreases in self-reported psycho-logical well-being occurred after exposure to theVE. Physical well-being, assessed objectively by re-cording heart rate during the VR exercises, revealedthat PWD experienced no decline in physical well-being. The pulse oximeter did not cause discomfortor intrude on the VE experience. Physiological mea-sures such as heart rate have been reported to bepositively correlated with feelings of presence and

perceived realism.39 Overall, the heart rate of PWDin the current study increased slightly during theVR exercises, but it was unclear if this was associ-ated with a sense of presence or a perception of re-alism. It is possible that these negligible increasesmay have been due to the effects of situational anx-iety and/or the physical exertion required to oper-ate the joystick rather than exposure to the VE.Person 2 experienced a dramatic increase in heartrate during the first exercise in the second session,although this was an emotional reaction to her in-ability to find objects recalled from the previoussession. Despite person 2 reporting no physical in-dications of discomfort, this incident clearly dem-onstrated the value of physiological monitoring ofPWD in VEs. Physiological monitoring can serve asan early warning system for the detection of psy-chological and physical discomfort or distress ofPWD in VEs.

AREAS FOR IMPROVEMENT AND FURTHER RESEARCH

The small sample size seriously restricted the sta-tistical power of the inferential tests and we ac-knowledge that the results in this paper have a highlikelihood of Type II error. We also accept that thecurrent study has a degree of sampling error as,firstly, our sample of PWD was self-selected andhighly motivated and, secondly, PWD in the currentstudy may have lacked homogeneity in terms ofsymptom presentation and dementia sub-type ascognitive impairments between PWD are not uni-form and a definite diagnosis of sub-type is possi-ble only after a post-mortem examination.

A major barrier to the recruitment of sufficientnumbers of PWD in the early stages with heteroge-neous symptomology to achieve an acceptable levelof statistical power is that despite the prevalence ofdementia in the population, and established proto-cols for identifying the presence of dementing ill-ness, it is frequently overlooked or misdiagnosedin its early stages.40 Furthermore, PWD and carersmay avoid seeking help until symptoms seriouslyinhibit quality of life or a crisis point is reached.Consequently, the majority of individuals enteringthe health and social care system are in the laterstages of the disease. Commonly used cognitive as-sessment tools such as the Mini-Mental State Exam-ination lack sensitivity,41 or in many cases may havenot been conducted at all, resulting in many peoplefailing to receive a formal diagnosis of dementia.Indeed, even if cognitive assessments have beenundertaken they may have been conducted some

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 607

time ago and no longer be representative of currentcognitive functioning. These difficulties were en-countered in the current study, with only six PWDand their carers/keyworkers agreeing to take partafter a considerable recruitment effort. This involvedcontacting primary and secondary care profession-als, local health and social care services relevant todementia care and voluntary agencies such as theAlzheimer’s Society. Using the example of correla-tions between subscales on the DVRuse, a prospec-tive power analysis revealed that in order to achievea statistical power of 0.80 to detect a medium effectsize (r = 0.3) a total of 85 participants would need tobe recruited to establish the relationship (using atwo-tailed test) between usability factors; for a largeeffect size (r = 0.5) 29 participants would be required.Given the likely difficulties recruiting this numberof PWD, researchers in this area must be preparedto devote considerable time and effort, including fi-nancial resources, to recruitment.

Carers and keyworkers were asked to refrain fromhelping PWD to navigate and to offer encourage-ment and reassurance during the VR sessions. How-ever, a specific issue arose with person 6. This personexperienced significantly difficulty when using thejoystick and her carer was very critical of her per-formance during the latter part of first VR exercisethat, in the opinion of the research team, impactednegatively upon her well-being and seriously di-minished her performance in the VE. However, thecarer’s motivation for doing this was not maliciousand was probably due to a desire for person 6 to“do well.” This incident clearly demonstrated thatmore time should be spent with carers/keyworkersduring the consent process to ensure they are ab-solutely clear about their role during the VR sessions.

It was evident that reverse movements causedconfusion for several PWD, in particular for person6, as they were not associated with a change in view-point (i.e., walking backwards without turningaround). Clearly, this movement is not ecologicallyvalid, especially for extended periods. Therefore,navigation devices used by PWD should be cali-brated to enable 180 degree turns in an ecologicallyvalid fashion. It was also observed that using thejoystick for extended periods may have caused mus-cle fatigue in the arm and shoulder muscles. Thiswas due to participants having to reach forward tograsp the joystick as they were unable to place theirlegs underneath the table used to support it. Thismay have been distracting and caused reductions inpresence and perceptions of realism. Therefore, fu-ture work with PWD in VEs should develop strate-gies for minimising reaching demands associated

with the use of an input device by ensuring an er-gonomically sound arrangement of furniture andinput device.

Despite only one PWD experiencing significantdifficulty using the joystick to navigate through theVE, its relative utility compared to other navigationdevices, including other models of joysticks, is un-known. It is plausible that alternative navigationdevices such as spaceballs, mice and voice recogni-tion may have yielded a more natural method ofnavigation in the VE for some PWD, especially forpeople who have limited dexterity or muscle strengthas a result of arthritis or other physical impairments.Therefore, different navigation devices should beevaluated for their relative usability by PWD.

Future research should be conducted to deter-mine the psychometric properties of these measures.In particular, factor analyses are needed to confirmthe three- and five-factor structures of the SSQPWDand DVRuse, respectively. In addition, other psy-chometric properties such as repeatability (test re-testreliability) and responsiveness should be examined.

In the current study, we utilized a real-time mea-sure, the DVRuse, to investigate the views of PWDregarding the VE. However, it would have beenpreferable to have incorporated this measure intothe VE interface itself to maintain presence, guar-antee standardization of delivery, and remove theconfounding influence of divided attention betweenitems and the VE. Using an appropriate typefaceand font size for older people would also compen-sate for the decline in vision and hearing associatedwith ageing. PWD could then respond to items dis-played on the projection screen using an appropri-ate navigation device to select a graded categoryrepresentative of their status without having to at-tend to stimuli in the real-world. This approachcould also be used for administering instructions toPWD and to support error-free navigation.

Interactions with VEs should be compared withcorresponding abilities assessed in the real-worldto determine the ecological validity of behavior ex-hibited in VEs and to capitalize on the precise con-trol of stimuli afforded by VR-based approaches.11

In the current study, it is unclear if PWD wouldhave behaved the same way in real-world environ-ments. Therefore, future research should undertakea series of “validation exercises” in real-world envi-ronments that share the same environmental char-acteristics as the VEs. If performance in VEs andthe corresponding real-world environments are cor-related, then it can be assumed with confidencethat VEs are appropriate media for assessing be-havior that is relevant to daily functioning.

608 FLYNN ET AL.

A multitude of factors (and complex interactionsbetween factors) may impact negatively upon eco-logical validity of users’ experiences in VEs eitherby weakening presence or perceived realism. Severalof these factors were missing from the VE used inthe current study, such as commonly experiencedauditory and visual stimuli in park environments(e.g., bird song, meteorological features, and chil-dren playing). Furthermore, the VE interface al-lowed participants to walk through solid objects thatoccasionally created confusion.

Factors external to the VE such as optimal seat-ing position and design, distance from projectionscreen, field of view, and luminescence need to beclarified as they have the potential to impact uponpresence and perceived realism.33 Furthermore,person-related factors may also impact upon per-ceptions of the VE interface and method of interactionsuch as gender, dementia sub-type and symptoms,medication, and level of education.

Dementia presents significant challenges for in-terface design, and given that it is most prevalent inpeople aged over 65, the effects of the ageing pro-cess should be taken into account when designingthe VEs. Unfortunately, there is a dearth of researchon VR interface design for older people,8 and researchon ageing indicates that VE design should addressdeclines in vision, hearing, and psychomotor skills.42

CONCLUSION

This study has demonstrated that it is feasible towork in VEs with PWD, although, due to the smallsample size, generalizability of the results to otherPWD is not possible. However, the results are en-couraging, as they demonstrate VR can be an effec-tive user-centered methodology for eliciting andevaluating the perceptions, reactions, and perfor-mance of PWD in simulated environments. The newculture of dementia care encourages such user-centered methodologies for working with PWD, asthey respect personhood by providing a mediumfor PWD to communicate meaningfully with re-searchers, health care professionals, and carers.43,44

Despite much enthusiasm and significant biomed-ical efforts, a cure for dementia remains elusive. VRhas significant potential to empower people withdementia, and reductions in start-up costs for VRwill play a pivotal role in increasing the attractive-ness of VR-based approaches.45 However, if PWDand their carers are to benefit from these develop-ments, significant effort is needed to develop andevaluate VR applications to cognitive assessment,cognitive rehabilitation, therapeutic activity, and im-mersion and design, where there are potentiallysignificant advances that could be made.

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 609

APPENDIX

Items constituting the simulator sickness questionnaire for people with dementia

Symptom complex Symptom Item

Oculomotor disturbances Eyestrain Are your eyes feeling tired and sore?Difficulty Focusing Can you focus on this object?a

Blurred Vision Is your vision blurred at the moment?Headache Do you have any pain in your head at the moment?Fatigue Are you feeling tired or sleepy?

Disorientation Dizziness (eyes open) Are you feeling dizzy at the moment?Dizziness (eyes closed) Are you feeling dizzy at the moment?Vertigo Do feel off-balance?b

Nausea Nausea Do you feel like you may want to vomit?Stomach Awareness Do you feel queasy in your stomach?Increased Salivation Is your mouth watering more than usual?Burping Do you feel as if you want to burp at the moment?

aPerson was asked to look at the research assistant’s watch.bPerson was asked to stand up.Adapted from a scale developed by Kennedy et al. (1993). Simulator Sickness Questionnaire: An en-

hanced method for quantifying simulator sickness. International Journal of Aviation Psychology, 3:203–220.For response format used, see Figure 2.

ACKNOWLEDGMENTS

We would like to express our gratitude to all thestaff at the Virtual Reality and Innovation Centre,University of Teesside, especially Gary Quinn for histechnical advice and programming the view boundsinto the VE and Billy Hearne (the developer of theVE used in the research) for customizing the VE byincorporating additional objects. The research teamwould also like to thank the Alzheimer’s Society(Teesside Branch), the Young Person with Demen-tia Team (Eastbourne Road, Middlesbrough), andEasterside Day Centre for their assistance in recruit-ing participants. We are grateful to all the partici-pants for giving up their time to take part in theresearch, Steve Fenby for the loan of a pulse oxime-ter, George Reid for granting us permission to usehis photograph in the SSQPWD and the DVRusequestionnaire, and Christopher Leggett for assistingwith the running of the VR sessions.

REFERENCES

1. Cox, S. (1999). Home solutions: housing and support forpeople with dementia. London: HACT.

2. DoH. (2001). A National Service framework for older peo-ple. London: DoH.

3. Brookmeyer, R., Gray, S., & Kawas, C. (1998). Projec-tions of Alzheimer’s disease in the United States andthe public health impact of delaying disease onset.American Journal of Public Health 88:1337–1342.

4. Kasl-Godley, J., & Gatz, M. (2000). Psychosocial inter-ventions for individuals with dementia: an integra-tion of theory, therapy, and a clinical understandingof dementia. Clinical Psychology Review 20:755–782.

5. Zeman, S. (1997). All about dementia: a booklet for thosewanting to know more about dementia: The MentalHealth Foundation [On-line]. Available: www.mental-health.org.uk/html/content/bkdem.cfm.

6. Flynn, D., van Schaik, P., & Blackman, T. (2002). Thepotential of virtual reality technology for the cog-nitive assessment and rehabilitation of dementia.[Internal publication, The University of Teeside,Middleburgh, UK.]

7. Neisser, U. (1978). Memory: what are the importantquestions? In: Gruneberg, M.M., Morris, P.E., & Sykes,R.N. (eds.), Practical aspects of memory. London: Aca-demic Press, pp. 3–24.

8. Rizzo, A. (2001). Advances in the application of vir-tual environments for mental healthcare. Presentedat the IEEE Virtual Reality 2001, Japan.

9. Rose, F.D., Attree, E.A., & Johnson, D.A. (1996). Vir-tual reality: An assistive technology in neurologicalrehabilitation. Current Opinion in Neurology 9:461–467.

10. Rizzo, A., & Buckwalter, J.G. (1997). Virtual realityand cognitive assessment and rehabilitation: the

state of the art. In: Riva, G. (ed.), Virtual reality inneuro-psycho-physiology. Amsterdam: IOS Press, pp.123–145.

11. van Veen, H.A.H.C., Distler, H.K., Braun, S.J., et al.(1998). Navigating through a virtual city: Using vir-tual reality technology to study human action andperception. Future Generation Computer Systems 14:231–242.

12. Brawley, E.C. (2001). Environmental design for Alz-heimer ’s disease: a quality of life issue. Aging & Men-tal Health 5(Suppl. 1):79–83.

13. Cohen, U., & Day, K. (1991). Contemporary environ-ments for people with dementia. Baltimore: Johns Hop-kins University Press.

14. Elmstahl, S., Annerstedt, L., & Ahlund, O. (1997).How should a group living unit for demented elderlybe designed to decrease psychiatric symptoms? Alz-heimer Disease & Associated Disorders 11:47–52.

15. Passini, R., Rainville, C., Marchand, N., et al. (1998).Wayfinding and dementia: some research findingsand a new look at design. Journal of Architectural andPlanning Research 15:133–151.

16. Day, K., Carreon, D., & Stump, C. (2000). The therapeu-tic design of environments for people with dementia:a review of the empirical research. Gerontologist 40:397–416.

17. Commission, A. (2000). Forget me not: mental healthservices for older people. London: Audit Commission.

18. Blackman, T., Mitchell, L., Burton, E., et al. (2003).The accessibility of public spaces for people with de-mentia: a new priority for the “open city.” Disabilityand Society 18:357–437.

19. Rizzo, M., Reinach, S., McGehee, D. (1997). Simulatedcar crashes and crash predictors in drivers with Alz-heimer ’s disease. Archives of Neurology 54:545–551.

20. Tarnanas, I. (2000). A virtual environment for theassessment and the rehabilitation of the visuo-constructional ability in dementia patients. Presentedat the 8th Annual Medicine Meets Virtual Reality In-ternational Conference, Newport Beach, California.

21. Fencott, C., van Schaik, P., Ling, J., et al. (2003). Theeffects of movement of attractors and pictorial con-tent of rewards on users’ behaviour in virtual envi-ronments: an empirical study in the framework ofperceptual opportunities. Interacting with Computers15:121–145.

22. Kennedy, R.S., Lane, N.E., & Berbaum, K.S. (1993).Simulator sickness questionnaire: An enhancedmethod for quantifying simulator sickness. Interna-tional Journal of Aviation Psychology 3:203–220.

23. Kennedy, R.S., Lanham, D.S., & Drexler, J.M. (1997).A comparison of cybersickness incidences, symptomprofiles, measurement techniques and suggestionsfor further research. Presence 6:638–644.

24. Stanney, K.M., & Hash, P. (1998). Locus of user-initi-ated control in virtual environments: influences oncybersickness. Presence 7:447–459.

25. Kalawsky, R.S. (1999). VRUSE—a computerised di-agnostic tool for usability evaluation of virtual/

610 FLYNN ET AL.

synthetic environment systems. Applied Ergonomics30: 11–25.

26. Goldsmith, M. (1996). Hearing the voice of people withdementia. Opportunities and obstacles. London: JessicaKingsley.

27. Acello, B. (2001). Pain assessment in patients withdementia and how to use pain scales. Home Health-care Nurse 19:473–476.

28. Ferrell, B.A., Ferrell, B.R., & Rivera, L. (1995). Pain incognitively impaired nursing home patients. Journalof Pain Symptom Management 10:591–598.

29. Ferrell, B.A. (1996). An overview of aging and pain.In: Ferrell, B.A., & Ferrell, B.R. (eds.), Pain in the el-derly: a report of the task force on pain in the elderly of theinternational association for the study of pain. Seattle:IASP Press.

30. Regan, E.C. (1995). Some evidence of adaptation toimmersion in virtual reality. Displays 16:135–139.

31. Schubert, T., Friedmann, F., & Regenbrecht, H. (2001).The experience of presence: factor analytic insights.Presence 10:266–281.

32. Lewis, C.H., & Griffin, M.J. (1997). Human factorsconsideration in clinical applications of virtual reality.Studies in Health Technology & Informatics 44:35–56.

33. Nichols, S., & Patel, H. (2002). Health and safety im-plications of virtual reality: a review of empirical evi-dence. Applied Ergonomics 33:251–271.

34. Kesztyues, T.I., Mehlitz, M., Schilken, E., et al. (2000).Preclinical evaluation of a virtual reality neuropsy-chological test system: occurrence of side effects. Cy-berPsychology & Behavior 3:343–351.

35. Regenbrecht, H., & Schubert, T. (2002). Real and illu-sory interactions enhance presence in virtual envi-ronments. Presence 11:425–434.

36. Davies, R. (2002). Applications of system designusing virtual environments. In: Stanney, K. (ed.),Handbook of virtual environments. Mahwah, NJ: Erl-baum, pp. 1079–1100.

37. Gamberini, L. (2000). Virtual reality as a new re-search tool for the study of human memory. Cyber-Psychology & Behavior 3:337–343.

38. Ruddle, R.A., Payne, S.J., & Jones, D.M. (1999). Navi-gating large-scale virtual environments: what differ-ences occur between helmet-mounted and desk-topdisplays? Presence 8:157–168.

39. Jang, D.P., Kim, I.Y., Nam, S.W., et al. (2002). Analysisof physiological response to two virtual environ-ments: driving and flying simulation. CyberPsychol-ogy & Behaviour 5:11–18.

40. Norby, E.H., Van Heertum, R., Murphy, E., et al.(1999). PET in clinical practice. Lippincott’s PrimaryCare Practice 3:559–573.

41. Cummings, J.L., & Cole, G. (2002). Alzheimer ’s dis-ease. Journal of the American Medical Association 287:2335–2338.

42. Hawthorn, D. (2000). Possible implications of agingfor interface designers. Interacting with Computers12:507–528.

43. Kitwood, T. (1993). Towards a theory of dementiacare: The interpersonal process. Aging and Society 12:269–287.

44. Kitwood, T. (1997). Dementia reconsidered. Bucking-ham: Open University Press.

45. Ayers, A. (2002). Applied VR [On-line]. Available:www.vrnews.com/forums/virgic/virgichome.html.

Address reprint requests to:Paul van Schaik, Ph.D.

School of Social Sciences and LawThe University of Teesside

Middlesbrough, TS1 3BA, U.K.

E-mail: P.Van-Schaik@tees.ac.uk

A VIRTUAL REALITY–BASED METHODOLOGY FOR DEMENTIA 611