Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Ecological Validity of Virtual Reality Daily Living Activities Screening for Early Dementia: Longitudinal Study

Ecological Validity of Virtual Reality Daily Living Activities Screening for Early Dementia:... Background: Dementia is a multifaceted disorder that impairs cognitive functions, such as memory, language, and executive functions necessary to plan, organize, and prioritize tasks required for goal-directed behaviors. In most cases, individuals with dementia experience difficulties interacting with physical and social environments. The purpose of this study was to establish ecological validity and initial construct validity of a fire evacuation Virtual Reality Day-Out Task (VR-DOT) environment based on performance profiles as a screening tool for early dementia. Objective: The objectives were (1) to examine the relationships among the performances of 3 groups of participants in the VR-DOT and traditional neuropsychological tests employed to assess executive functions, and (2) to compare the performance of participants with mild Alzheimer’s-type dementia (AD) to those with amnestic single-domain mild cognitive impairment (MCI) and healthy controls in the VR-DOT and traditional neuropsychological tests used to assess executive functions. We hypothesized that the 2 cognitively impaired groups would have distinct performance profiles and show significantly impaired independent functioning in ADL compared to the healthy controls. Methods: The study population included 3 groups: 72 healthy control elderly participants, 65 amnestic MCI participants, and 68 mild AD participants. A natural user interface framework based on a fire evacuation VR-DOT environment was used for assessing physical and cognitive abilities of seniors over 3 years. VR-DOT focuses on the subtle errors and patterns in performing everyday activities and has the advantage of not depending on a subjective rating of an individual person. We further assessed functional capacity by both neuropsychological tests (including measures of attention, memory, working memory, executive functions, language, and depression). We also evaluated performance in finger tapping, grip strength, stride length, gait speed, and chair stands separately and while performing VR-DOTs in order to correlate performance in these measures with VR-DOTs because performance while navigating a virtual environment is a valid and reliable indicator of cognitive decline in elderly persons. Results: The mild AD group was more impaired than the amnestic MCI group, and both were more impaired than healthy controls. The novel VR-DOT functional index correlated strongly with standard cognitive and functional measurements, such as mini-mental state examination (MMSE; rho=0.26, P=.01) and Bristol Activities of Daily Living (ADL) scale scores (rho=0.32, P=.001). http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 1 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Conclusions: Functional impairment is a defining characteristic of predementia and is partly dependent on the degree of cognitive impairment. The novel virtual reality measures of functional ability seem more sensitive to functional impairment than qualitative measures in predementia, thus accurately differentiating from healthy controls. We conclude that VR-DOT is an effective tool for discriminating predementia and mild AD from controls by detecting differences in terms of errors, omissions, and perseverations while measuring ADL functional ability. (JMIR Serious Games 2013;1(1):e1) doi: 10.2196/games.2778 and rehabilitation, including virtual cities [9,10], school Introduction classrooms [11], and supermarkets [9,12]. As outlined previously, ecological validity can be seen as a key component A decade ago, Chaytor and Schmitter-Edgecombe [1] reviewed for assessing cognitive skills that are relevant for functional the ecological validity of neuropsychological tests by evaluating tasks in real-world contexts [13]. The results of such studies their efficacy in measuring everyday cognitive skills. They suggest that the use of VEs is valuable in enhancing our ability identified 6 studies that explored the issue of ecological validity to assess the functional behaviors of individuals with executive of executive functioning tests. The studies differed in terms of dysfunction in activities of daily living. the specific tests used, although both traditional (veridicality) and verisimilitude tests were employed. Veridicality refers to Activities of daily living (ADL) can be classified into basic the extent to which results of an assessment instrument are activities of daily living (BADL) and instrumental activities of related to scores on other tests that predict the performance of daily living (IADL) [14]. BADL is composed of more basic real-world tasks [2]. By contrast, verisimilitude refers to the self-care behaviors, such as ambulating, dressing, grooming, similarity between the task demands of the test and the demands bathing, feeding, and toileting. By contrast, IADL facilitates imposed in the everyday environment. independent living through behaviors such as transportation, telephone use, meal preparation, medication management, Their findings indicated that executive tests were not financial management, housekeeping, laundry, and shopping. significantly correlated with self-reported measures, but all IADL questionnaires play a vital role in assessing functional studies reviewed revealed significant associations between abilities and evaluating the impact of cognitive impairment on executive tests (traditional and verisimilitude) and everyday everyday activities in older adults [15]. abilities as measured by clinician ratings and informants’ (eg, relatives’) questionnaires. To date, commentaries on ecological IADL independence is one of the defining features that validity have primarily emphasized the increased consideration characterize normal aging from mild cognitive impairment of this concept in assessments of neurologically impaired (MCI) and dementia. As part of the diagnostic criteria for MCI, individuals, particularly in rehabilitative and forensic contexts. an individual must be classified as independent for BADL, but However, there are instances in which patients perform normally can have minimal disturbance in IADL [16,17]. Since the early on traditional executive tests, yet clearly have executive descriptions of MCI [16], there has been increasing interest in impairments in their daily lives [3]. its clinical characterization and prognosis [18,19]. In previous reports [20], people with MCI exhibited poorer cognitive Virtual environments (VEs) have numerous features that make functioning than healthy controls, but were not as impaired as them attractive for assessment and rehabilitation purposes. In patients with dementia were. contrast to traditional executive test measures, VEs actively engage participants by allowing them to be involved in a task Prognostic studies have stressed the necessity of this nosological while at the same time being less focused on the fact that they entity as a risk, or prodromal state for dementia, because of the are being tested [4,5]. More recently, researchers have used high rate of conversion of MCI to dementia (10%-15% of virtual reality (VR) systems for detailed response measurement patients who meet the criteria of amnestic MCI develop and analysis to examine specific behaviors characteristic of Alzheimer-type dementia per year, up to 80% at 5-year patients with executive dysfunction or people with intellectual follow-up) [21,22,18]. In Europe, approximately 17% of the disabilities [6]. Klinger and colleagues [7] examined planning senior population who have not been diagnosed with dementia deficits in patients with Parkinson’s disease compared to meet the current criteria for MCI [23] and MCI prevalence age-matched controls in a virtual supermarket. The researchers increases with age [24]. described the patients’ paths through the supermarket as Characterizing impairment using the IADL questionnaire has characterized by numerous stops, turns, and hesitancies as been criticized for several reasons. First, no objective standard compared to the paths of controls. Zhang and colleagues [8] exists as to the practical or theoretical definition of minimal used a virtual kitchen to assess selected cognitive functions of functional impairment in predementia [16,17]. For example, traumatic brain injury patients compared to normal volunteers. does functional disturbance entail perceptible impairment on a Task transparency and relevant functional tasks, such as finding few IADL tasks, such as shopping and meal preparation? Or is one’s way through a VE or remembering groceries for preparing it better understood as some problems across many commonly a breakfast in a virtual kitchen, are examples in which ecological assessed IADL tasks? Clinical judgment is called for by the validity can be described as enhanced when compared to abstract expert panel that created these standards [25,26], but the general traditional assessments of cognitive functions. A variety of VEs clinician or researcher is without much guidance regarding how have already been developed to enhance functional assessment to assess IADL impairment in predementia patients. Several http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 2 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al options exist, including performance-based tasks and participants [33], and (3) in mild AD and MCI patients, the rate questionnaires or interviews (with and without informant of change in individual performance in VR-DOT measures reports). However, different methods of assessing functional could predict the cognitive decline over 2 to 3 years. abilities require different estimates of IADL independence [27]; The main objective in developing the VR-DOT was to improve each method has advantages and disadvantages. the ecological validity of executive function measures by using Recently, it became clear that IADL, versus BADL, is a better a verisimilitude approach. We also proposed a framework to diagnostic instrument for predementia [28-31]. Although these objectively assess the functional impairment of elderly people studies were carried out in different countries and used different through an ecological and clinical longitudinal experiment using instruments to assess impairments in ADL, they all indicate that VR-DOT. Our motivation was to correlate this new instrument people who meet the criteria described for predementia show (VR-DOT) with normal cognitive neuropsychological measures some functional impairment in activities of daily living. In the and recent psychomotor discoveries regarding psychomotor clinical setting, rate of change in complex ADL performance velocity change and cognitive decline to see if the VR-DOT may be more useful than a cross-sectional measurement, which offers better sensitivity and specificity in assessing and could misclassify individuals into a nonimpaired category in predicting cognitive decline using only a virtual environment. activities of daily living [32]. Furthermore, the rate of change is a parameter that could be manipulated by designing Methods naturalistic VEs or serious games that can train the higher Virtual Reality Test Setup cognitive functions. We designed a fire evacuation Virtual Reality Day-Out Task Overview (VR-DOT) environment to (1) determine what kind of real-time The VR hardware consisted of a Pentium-based computer with cognitive and psychomotor performance and errors are 4 MB RAM, Intel Quad Core processor, and NVIDIA graphic associated with functional impairment in activities of daily cards with 512 MB memory. Other sensors used were a LEAP living, (2) identify the patterns and cutoff values of the these motion sensor (Leap Motion Inc, San Francisco, CA, USA) and cognitive and psychomotor profiles as independent predictors a Kinect camera (Microsoft Corp, Seattle, WA, USA). The of functional impairment in healthy elderly participants, LEAP motion sensor is still not commercially available at the single-domain amnestic MCI patients, and patients with mild time of this writing, but we were selected by their development Alzheimer-type dementia (AD), and (3) controlling for baseline team to use the hardware for our experiments (Figure 1). performance, objectively measure performance change over 2 Software Components to 3 years. Modeling was done using Maya software (Autodesk Inc, San We hypothesized that with VR-DOT (1) dementia and MCI Rafael, CA, USA) to create models and scenes. Then, the scenes patients will show significantly impaired independent were exported to Virtools, a 3D authoring tool (Dassault functioning in ADL and distinct performance profiles, (2) among Systèmes, Inc, Vélizy-Villacoublay Cedex, France) that handled patients with dementia or MCI, such impairment will be all programming including interactivity, setting, and associated with the degree of cognitive impairment and cognitive configuration. Microsoft Kinect software development kit (SDK) neurophysiological measures, whereas impaired functioning (Microsoft Corp, Seattle, WA, USA) was used to analyze will be only associated with sociodemographic and gestures and movements and a user interface (UI) system was anxiety/depression symptoms in healthy controls because developed using Microsoft Kinect SDK and precommercial subclinical levels of cognitive impairment and depression have Alterniity algorithms developed by Ioannis et al [34]. been associated with IADL impairment in mentally healthy http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 3 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 1. Virtual reality day-out task (VR-DOT) participant setup. allowing VR applications to read the skeletal joints as trackers The Naturalistic Setting of Executive Function in Virtual using any VRPN client. Additionally, the toolkit can also Reality Activities of Daily Living emulate keyboard input triggered by body posture and specific Virtual reality activities of daily living (VR-ADL) consists of gestures. This allows the user to add custom body-based control 2 modules: the VR-DOT and VR basic instrumental activities mechanisms to existing off-the-shelf games that do not provide of daily living (VR-IADLs). The VR-DOT is a complex task official support for depth sensors. breakdown and then a rehearsal exercise of a fire evacuation More specifically, the VR-DOT module is a naturalistic task drill consisting of 6 different scenarios of increasing difficulty. that requires multitasking in a fire evacuation drill setting with We chose to examine the VR-DOT virtual fire evacuation drill 6 different simulated fire situations (from easy to more difficult) in this study (Figure 2), based on the literature indicating that taking place at a virtual apartment block with 3 levels and 5 activities of daily living requiring complex reasoning are apartments per level. It is used to examine prospective memory sensitive to cognitive and functional impairment [35]. User as well as reasoning in a complex emergency routine in which tracking was performed by a flexible action and articulated older adults prioritize, organize, initiate, and complete a number skeleton toolkit (FAAST; University of Southern California, of subroutines to evacuate safely from an apartment level CA, USA), a middleware to facilitate integration of full-body (second floor) to the ground area (eg, determine and gather control with games and VR applications, using either OpenNI information on the size of the fire, avoid smoke). Previous or the Microsoft Kinect for Windows skeleton tracking software. research shows that motion tracking while navigating a virtual FAAST includes a custom virtual reality peripheral network environment is a valid and reliable indicator of cognitive decline (VRPN) server to stream up to 4 user skeletons over a network, in elderly persons. (ie, [36]). http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 4 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 2. Sample sequential virtual reality day-out task (VR-DOT) screenshots, showing different tasks and viewpoints. Procedure Functional and Psychomotor Rate of Change Participants A measurement of the rate of change of functional impairment was computed from all information collected from the LEAP A total of 405 elderly people were screened during 2010 in 2 motion and the Microsoft Kinect camera sensor inside the Alzheimer Hellas, Non-Government Organization (NGO) day VR-DOT. Simple performance-based functional impairment clinics of the Papanikolaou University Hospital in Thessaloniki, measures have been used previously, but not with data collected Greece. Ethics approval was obtained from the Papanikolaou from motion sensors [37]. At baseline, a simple quantitative University Hospital Ethics Committee. Inclusion criteria were ratio of efficacy was computed by dividing the total time (in age older than 60 years, meeting the diagnostic criteria for MCI sec) spent by the participant performing the listed activities by as defined in Petersen et al [18], living in the community, and the total time spent in VR-DOT (efficacy ratio). Then, 4 activity providing informed consent approved by the Ethics Committee. parameters with a high likelihood of corresponding to functional Exclusion criteria were living in an assisted-living residence, decline were collected: (1) omission of 1 of the activities (k ), cognitive functioning suggesting a possible diagnosis of (2) repetition of the same activity (k ), (3) incorrect order in dementia (see subsequent description), previous diagnosis of dementia, other psychiatric disorder according to the Diagnostic performing the activities (k ), and (4) number of attempts before and Statistical Manual of Mental Disorders (Fourth Edition, completing a given activity (k ). The first quantitative ratio of Text Revision; DSM-IV-TR) at the time of recruitment, efficacy was then adjusted by these parameters. This led to a presenting a moderate or higher degree of fear or dislike of functional impairment score according to the formula presented computers (technophobia), presenting a moderate or higher in Multimedia Appendix 1. degree of disability because of other conditions than MCI, and To determine values of the model parameter set (k , k , k , k ), severe language impairments that would compromise active 1 2 3 4 participation. we ran a pilot with healthy participants (n=25; mean age 73.7 years, SD 4.0), amnestic single-domain MCI patients (n=26; The baseline psychomotor evaluation inside VR-DOT consisted mean age 74.2 years, SD 2.0), and patients with mild AD (n=24; of a number of simple and complex measures addressing the mean age 76.7 years, SD 3.0). Second, multiple-model parameter participant’s ability to understand and perform with accuracy sets (k , k , k , k ) to produce a good fit were selected if their 1 2 3 4 specific physical performance tasks. These tasks/metrics were: associated scores were both strongly and positively correlated Grip Strength with the mini-mental state examination (MMSE) scores, as well as being strongly and negatively correlated with IADL scores Forearm muscle strength was measured in kilograms by a using a nonparametric Spearman correlation coefficient as the hand-held Jamar A dynamometer. For this analysis, we used criterion distance of good fit. For our analyses, the final the best of 3 attempts in the dominant hand. functional impairment score (k , k , k , k ) was calibrated using 1 2 3 4 Timed Walk on the Treadmill the combination of the mean of the parameters, which was The time (to 0.1 s) required for a participant to walk a 4.6-m selected as the model parameter set during the second step of course at his or her usual pace after starting from a standstill the fitting procedure. was recorded by stopwatch. We converted the results to meters per second. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 5 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Number of Steps to Walk Course on the Treadmill Significant effects were further tested with post hoc tests that were corrected for multiple comparisons using Tukey’s Honestly A technician recorded the number of steps required to walk the Significant Difference (HSD) [42]. We used similar statistical 4.6-m course. Hereafter, we refer to stride length, which is models to estimate the RR of having significant VR-DOT derived by dividing the distance walked by the number of steps. difficulty or inability (relative to no or mild difficulty) and the Finger-Tapping Test RR of cognitive decline (relative to no or mild difficulty) at Using their dominant hand, participants tapped in midair, just year 3 for upper extremity function. For a given performance above the LEAP motion sensor, with the index finger as fast as measure, the first year 1 value and the third year 1-3 slope of they could for 15 seconds. change were treated as separate predictor variables. For each outcome, a separate regression model was run for each predictor After applying the inclusion/exclusion criteria, 232 participants performance variable, adjusting for age, gender, and the were included in this study. The participants were measured VR-DOT task of more difficulty with, or disability in, the each year of the 3-year study period. Dropouts after baseline outcome measure between years 1 and 3. Next, we were 27 (11.6%); hence, 205 participants completed all simultaneously entered all predictor performance variables into measurements and their data are included in this study (N=205; a second set of models, adjusting for the same covariates. The male=88, female=117; mean age 72.73 years, SD 6.89; mean component variables from each model were entered, in turn, education 12.53 years, SD 3.20; mean baseline MMSE 24.75, into a stepwise backward regression for the respective outcomes, SD 2.18). with a P value to enter the model set at <.10. This procedure yielded a set of simpler, more parsimonious final models. All Neuropsychological Assessment statistical analyses were run using SPSS 19.0 statistical software Cognitive assessment was performed by means of a (IBM Corp, Armonk, NY, USA). neuropsychological test battery designed to comprehensively evaluate attention, working memory, memory, executive Results functioning, and language. In addition to the cognitive assessment, all groups were also assessed for depression with Demographics and baseline scores for all groups are shown in the geriatric depression scale (GDS) [38]. We also chose the Table 1. Digit Symbol (DSym), Functional Activities Questionnaire (FAQ), Neuropsychiatric Inventory brief questionnaire form, After corrected with age, gender, and education status, our Apathy item (NPI-Q Apathy), Neuropsychiatric Inventory brief results showed that the VR-DOT functional index was correlated questionnaire form, Depression item (NPI-Q Depression), Rey strongly with standard cognitive and functional measurements, Auditory Verbal Learning Test T (RAVL), Trailmaking Test such as MMSE (rho=0.26, P=.01) and Bristol ADL scores A (TMT-A), Trailmaking Test B (TMT-B), Trailmaking Test (rho=0.32, P=.001), thus accurately differentiating from healthy B minus Trailmaking Test A (TMT-B-A), the Bristol ADL control participants (Table 2). scale, and the short form of the Blessed ADL scale for this study In the prediction models for individual performance measures because they were evaluated and validated for the Greek (not shown), the VR-DOT and upper extremity function population [39]. The original Bristol and short-form Blessed psychomotor performance (finger tapping, etc) at year 3 for the scales consist of 20 and 11 items, respectively. MCI and mild AD group, as well as the slopes of change, had a significance of P<.10. Compared with the control, weaker Statistical Analysis results of the MCI and mild AD independently predicted Performance results from the VR-DOT, gait velocity assessment, cognitive decline at year 3 in all 3 domains (VR-DOT, and neuropsychological tests were analyzed using multivariate neuropsychological, and gait velocity assessment). The change analyses of variance (MANOVA) in mixed designs with group slope for upper extremity function inside the VR-DOT was also as the between-subject factor using linear mixed-effects models associated with the outcome. with random intercept and slope to estimate the annual rate of change between study years 1 and 3 for each performance For functional independence, the healthy group showed better measure of each participant [40,41]. Before this approach, we functional adjustment than the MCI and mild AD group plotted numerous individual trajectories for the gait velocity according to VR-DOT total monitoring data. When the amnestic performance variables by using robust splines to smooth the MCI group was examined using the VR-DOT total score, curves. The consistent linearity of the trajectory patterns justified cognitive domain and gait velocity assessment showed a the use of linear models. Gait speed and stride length were similarly impaired profile as cognitive functioning, after adjusted to a 50-cm knee-heel length and this adjustment was controlling for age, education, and GDS score. The mild AD included in the models when it reached 10% significance. We patients showed a higher degree of functional impairment than also used multinomial Poisson log-linear models to estimate both healthy controls and amnestic MCI patients in life the relative risk (RR) of cognitive decline relative to efficacy, activities, and participation subscales, respectively, and in the gait velocity, and neuropsychological assessment at year 2 and VR-DOT mobility domain. The total VR-DOT functional ability year 3 (2010-2013) for VR-DOT and receiver-operating curve measures showed a consistent functional impairment of mild (ROC) analysis was conducted on VR-DOT, MMSE, the AD and amnestic MCI in comparison with healthy participants. RAVLT, and the Bristol and Blessed ADL scale scores. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 6 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Table 1. Participant demographics and scores on cognitive tests for all participants, healthy controls, patients with amnestic-type mild cognitive impairment (aMCI), and patients with mild Alzheimer-type dementia (AD). Group All participants Controls aMCI Mild AD N=205 n=72 n=65 n=68 Age, mean (SD) 72.73 (6.8) 72.63 (5.06) 72.78 (6.21) 72.58 (6.21) Sex, n (%) Male 88 (43%) 25 (38%) 30 (43%) 33 (46%) Female 117 (57%) 37 (62%) 40 (57%) 40 (54%) Education, mean (SD) 15.6 (3.0) 16.1 (2.9) 15.7 (3.0) 14.6 (3.2) Test, mean (SD) MMSE 24.75 (2.18) 29.1 (1.0) 26.1 (1.8) 23.4 (2.0) RAVLT delayed recall 3.7 (4.0) 7.4 (3.7) 2.9 (3.3) 0.7 (1.6) GDS 1.4 (1.4) 0.8 (1.1) 1.6 (1.4) 1.6 (1.4) NPI-Q depression 0.2 (0.5) 0.1 (0.3) 0.2 (0.5) 0.4 (0.6) NPI-Q apathy 0.2 (0.6) 0.01 (0.1) 0.2 (0.5) 0.5 (0.8) FAQ 4.8 (6.4) 0.1 (0.6) 3.8 (4.4) 12.7 (6.7) TMT-A 46.6 (25.5) 36.3 (13.0) 44.2 (21.7) 64.8 (34.5) TMT-B 134.5 (80.2) 89.3 (44.3) 130.8 (73.2) 200.5 (86.6) TMT-B–A 88.0 (66.9) 53.0 (38.8) 86.6 (63.1) 135.8 (74.3) Bristol ADL scores 6.88 (0.56) 4.46 (0.5) 5.59 (0.9) 10.59 (0.9) Blessed ADL impairment score 2.87 (0.26) 1.85 (0.27) 2.38 (0.56) 4.38 (0.56) Geriatric depression scale 5.19 (5.0) 4.59 (4.1) 5.49 (5.76) 5.29 (4.45) Digit Symbol 37.4 (12.9) 45.8 (10.2) 37.0 (11.1) 27.6 (12.5) Gait speed (m/s), mean (SD) Combined 0.91 (0.22) 0.96 (0.23) 0.91 (0.24) 0.86 (0.20) Women 0.85 (0.14) 0.94 (0.24) 0.84 (0.04) 0.77 (0.14) Men 0.98 (0.13) 1.00 (0.21) 1.01 (0.03) 0.95 (0.04) Tapping speed dominant (taps/second), mean (SD) Combined 3.79 (0.8) 3.87 (0.8) 3.77 (0.81) 3.74 (0.8) Women 3.48 (0.78) 3.53 (0.71) 3.49 (0.84) 3.43 (0.77) Men 4.23 (0.75) 4.29 (0.77) 4.21 (0.73) 4.19 (0.75) Tapping speed non-dominant (taps/second), mean (SD) Combined 3.60 (0.67) 3.63 (0.64) 3.61 (0.71) 3.58 (0.7) Women 3.73 (0.59) 3.41 (0.53) 3.38 (0.64) 3.33 (0.61) Men 3.92 (0.63) 3.91 (0.65) 3.96 (0.63) 3.90 (0.62) MMSE: mini-mental state examination, RAVLT: Rey Auditory Verbal Learning Test, GDS: Geriatric Depression Scale, NPI-Q Depression: Neuropsychiatric Inventory brief questionnaire form, Depression item, NPI-Q Apathy: Neuropsychiatric Inventory brief questionnaire form, Apathy item, FAQ: Functional Activities Questionnaire, TMT-A: Trailmaking Test ), TMT-B: Trailmaking Test B, TMT-B–A: Trailmaking Test B minus Trailmaking Test A, ADL: Activities of Daily Living. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 7 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Table 2. The correlation matrix between Virtual Reality Day-Out Task (VR-DOT) functional index, mini-mental state examination (MMSE), and Bristol Activities of Daily Living (ADL) when controlling for age, gender, and education status. Test VR-DOT MMSE Bristol ADL rho P value rho P value rho VR-DOT 1 MMSE 0.26 .01 1 Bristol ADL 0.32 .01 0.43 .01 1 The AUC indicates that VR-DOT was the most powerful of all Predictors of Functional Status (Regression Analyses) tests in discriminating normal controls from the MCI groups, When the entire sample was analyzed together, attention, reaching optimal results with a cutoff point of 20 (97% psychomotor, and memory summary scores explained a total sensitivity, 100% specificity, 100% positive predictive values, variance of 8.2% and 0.8% of the VR-DOT. When depression and 96% negative predictive value). Figure 3 shows the ROC and age were entered in both former models, the VR-DOT score of the normal control and MCI groups for VR-DOT total score. was predicted by depression symptoms as measured by the GDS Exploratory Prediction of Conversion to Alzheimer (19.2%) only in the healthy group. By contrast, VR-DOT total Disease (VR-DOT Performance Rate of Change) score was only predicted by psychomotor and executive functions (8.1%) among mild AD and amnestic MCI patients. According to the results, the task that better differentiated among For the amnestic MCI group, VR-DOT was predicted by healthy controls, amnestic MCI, and mild AD participants at executive functions and psychomotor profile only, with a total baseline, year 2, and year 3 follow-up, was the VR-DOT variance explained of 17.3% for amnestic MCI. VR-DOT score performance score (efficacy ratio). The VR-DOT and Bristol was predicted only by executive function and psychomotor and Blessed ADL scale scores were included as predictor profiles in amnestic MCI patients and by executive function, variables in a series of exploratory independent regression psychomotor profiles, and GDS scores in mild AD. When a analyses. Figure 4 shows the individual predictive power of the ROC analysis was carried out on the Bristol and Blessed ADL 3 test variables of interest (VR-DOT, Bristol, and Blessed ADL scales, they explained 9.1% variance of VR-DOT total profiles. scale scores), ranked in ascending order according to the magnitude of their odds ratios. The VR fire evacuation ROC analysis was conducted on VR-DOT, MMSE, RAVLT, performance score rate of change (VR-DOT REff) emerged as and Bristol and Blessed ADL scale scores obtained from the the best predictor of conversion to AD in this sample (VR-DOT; amnestic MCI and mild AD groups and the sensitivity, P=.008; OR 2.8, 95% CI 1.3-6.0; Nagelkerke R =0.564), with specificity, and cutoff values of both the scales were determined the regression model correctly classifying 88% of participants. (Table 3). The optimal cutoff score of the Bristol scale was 20 This was followed by Bristol ADL (P=.03; OR 1.9, 95% CI in differentiating amnestic MCI from mild AD with a sensitivity of 100% and specificity of 74.2%, and area under the curve 1.1-3.5; Nagelkerke R = 0.563), and the Blessed ADL (P=.01; (AUC) of 0.883 (95% CI 0.781-0.975). The optimal cutoff score OR 1.4, 95% CI 1.1-1.9; Nagelkerke R =0.466). of the modified Blessed scale was 10.5 in differentiating The resulting regression model revealed that the VR-DOT amnestic MDI from mild AD with a sensitivity of 100%, performance score threshold variable was a significant predictor specificity of 71%, and AUC 0.872 (95% CI 0.791-0.977). Post of conversion to AD in the regression equation (beta=–1.092, hoc analysis revealed that among the 3 groups, the mild AD P=.01) with OR 3.0 (95% CI 1.3-7.0). Using a cutoff score of group had the lowest scores in ADL, episodic memory, and less than 20 on the VR-DOT subscale achieved a sensitivity of VR-DOT (P<.001). 100% and a specificity of 94%. Table 3. Area under the curve (AUC) for standard neuropsychological test scores and Virtual Reality Day-Out Task (VR-DOT) for healthy controls versus patients with amnestic-type mild cognitive impairment (aMCI) and patients with aMCI versus patients with mild Alzheimer-type dementia (AD). Test Healthy control vs aMCI aMCI vs mild AD AUC (95% CI) P value AUC (95% CI) P value MMSE 0.79 (0.68, 0.91) <.001 0.99 (0.97, 1.00) <.001 Bristol scores ADL 0.75 (0.62, 0.88) <.001 0.88 (0.78, 0.97) <.001 Blessed score ADL 0.77 (0.64, 0.89) .002 0.87 (0.79, 0.98) .02 RAVLT delayed recall 0.82 (0.77, 0.93) .001 0.88 (0.79, 0.98) <.001 DOT-VR 0.96 (0.88, 0.99) <.001 0.95 (0.88, 1.00) <.001 MMSE: Mini-Mental State Examination, ADL: Activities of Daily Living, RAVLT: Rey Auditory Verbal Learning Test. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 8 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 3. Receiver-operating curve (ROC) for the Virtual Reality Day-Out Task (VR-DOT) total score when discriminating among nondemented (healthy controls), amnestic mild cognitive impairment (aMCI), and patients with mild Alzheimer-type dementia (mild AD). Figure 4. Odds ratios from exploratory individual regression analyses using VR-DOT, Bristol, and Blessed ADL scale scores rate of change as predictors for conversion from mild cognitive impairment to Alzheimer disease (bars represent 95% CI). reality quantitative performance measure of functional ability Discussion (VR-DOT) showed adequate psychometric properties (ie, discriminant power) to contribute to a predementia diagnosis. There is still debate as to the utility of MCI as a diagnostic In addition, functional measures based on quantitative rates of category. Many older people report subjective cognitive the number and quality of ADL performed seem to be more complaints in the absence of objective impairment [43] and not sensitive to identifying functional impairment in predementia all such complaints are predictive of dementia [44]. MCI may than those based on a subjective judgment of disability. be viewed as being on a continuum from normal aging to dementia and the present data show a large overlap between As a result of this paradigm shift, and in light of previous and groups that coheres with this view [45]. In that context, only a the present results, it would be very helpful for clinicians, few studies have systematically examined the rate of change in caregivers, and health-system managers if MCI definitions complex ADL performance as a predictor of cognitive decline. included an objective measure of impairment of functional abilities as a clinical feature inherent to MCI. We found that Our results show that functional impairment is a defining feature VR-DOT has greater sensitivity and specificity, as well as of both amnestic MCI and mild AD, and that the impairment having both positive and negative predictive values compared showed by amnestic MCI patients is partially dependent on the to other screening tests in discriminating amnestic MCI and degree of their cognitive impairment. Furthermore, a virtual http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 9 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al mild AD from normal aging. In summary, these and previous was in the development and empirical study of a novel VR results emphasize the presence of qualitative and quantitative environment (VR-DOT) that was less structured and that more functional impairments of both basic and complex ADL in closely resembled actual everyday errands than existing predementia as a logical consequence of cognitive impairment. questionnaires. This research demonstrated that tests of Although dementia is characterized by a more severe degree of verisimilitude may be better predictors of real-world behaviors disability than predementia, the World Health Organization than many of the most commonly employed traditional executive (WHO) International Classification of Functioning, Disability function tests. and Health (ICF) conceptualization of disability would include Our approach with VR-DOT is part of a general effort to predementia as a disabling condition, although to a lesser degree. manifest marked impairment in cognitive performance, The need for a better definition of disability as a diagnostic particularly executive functions during everyday activities by criterion (putatively by shifting from a categorical notion of means of VR (VR-ADL). Studies directly investigating ADL able/disabled to a more spectrum/gradual approach) to have found mild and tardive impairment in MCI, and a relation discriminate predementia from dementia must not conceal the with certain executive functions, but the targeted ADL were fact that dementia patients have their own health/functional very simple tasks, such as memorizing a telephone number or assistance needs. walking a few meters, and have always been strictly limited to Moreover, given the moderately good psychometric properties the accuracy domain, excluding any performance or a demonstrated in our study of the VR-ADL in discriminating rate-of-change factor. The purpose of VR-DOT was (1) to healthy from predementia and mild dementia patients, assessing investigate performance, in an experimentally controlled real-time functional ability would improve the identification of manner, on a complex ADL (planning and evacuating a fire predementia patients, and the use of objective, VR qualitative under time pressure) that is more indicative of the true quality and/or quantitative impairment of functional abilities as a of life of senior citizens, and (2) to scrutinize its cognitive diagnostic criterion should be further explored. Goldberg et al structure as a diagnostic instrument, which can screen functional [26] found that a sensitive performance-based measure they impairments at a very early phase of AD. With regard to real-life developed (the University of California, San Diego ADL, this investigation presents the advantage and innovation Performance-Based Skills Assessment; UPSA) had a remarkably of a VR quantitative scoring grid of a very complex set of good discriminant power to distinguish healthy participants sequential activities under demanding time constraints. from amnestic MCI participants (AUC 0.84), and to distinguish This study found that VR-DOT is comparatively better in amnestic MCI patients from patients with AD (AUC 0.88). detecting amnestic MCI from normal aging individuals. From Hence, the inclusion of functional competence measures seems quantitative and qualitative data extracted from VR-DOT, a convenient for the screening and early identification of functional index was computed, validated, and compared with neurodegenerative processes characterized by cognitive current clinical rating scales. Results of this pilot study are impairment. promising and must be substantiated with a larger sample and The rates of change in complex everyday activities, easily in another assessment setting to evaluate its reproducibility. determined in longitudinal practice settings, provide important Verisimilitude instruments, such as VR-DOT, can potentially prognostic information for late-life disability and death that are play valuable roles in both executive function assessment and independent of the predictive value of a performance intervention and, consequently, may help place clinical measurement obtained at a single point in time, which could be neuropsychology on firmer scientific ground. Researchers and inaccurate because of recent injury or illness. By predicting clinicians have the responsibility and opportunity to design, decline in ADL and IADL, upper extremity functionality, and test, and implement effective therapeutic strategies to improve, more generalized daily activities, longitudinal views capture or at least preserve, functional and cognitive functioning in broader deteriorations in function within an individual, predementia. suggesting a shared causal pathway. For these purpose, it is assumed that the visual quality and Our study has limitations. Although we used a population-based realism of the VEs are of central importance for patients to cohort, the exclusion from the analyses of participants with recognize and acknowledge the relevance of the task and context technophobia [46] may have introduced bias and reduced the at hand. Essential characteristics of virtual scenarios and tasks generalizability of the results. Although we only observed linear (ie, transparency, believability, plausibility, and relevance) are patterns in the many performance trajectories that we plotted, summarized under the term “realism” to describe that the patient some individual trajectories could have been nonlinear causing can recognize the employed tasks and scenarios and refer to inaccurate estimates of annual performance change. Our them based on past experiences. VR-ADLs capture the patient’s statistical models contained a limited number of covariates. interest and improve long-term motivation to use the virtual Although the addition of comorbid conditions to the models did tasks at high frequencies. Transparency and realism in a broader not significantly alter the results, we may have omitted important sense can relate to plausibility and place illusions that are confounders. described by Slater [47]. Plausibility illusion refers to the fact that the user believes the virtual scenario is actually occurring. The present research described the ecological validity of It is caused by events and the scenario relating directly to the verisimilitude and traditional activities of daily living measures user (eg, the virtual character talking to the user). Place illusion and the characterization of various subcomponents of the refers to the sensation that the user is actually situated in the executive function system. The unique contribution of this study displayed location and is described in relation to sensorimotor http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 10 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al contingencies of the VR system (eg, user interaction, tracking, demands (cognitive, motor, visual), easily allowing the creation and multimodal user feedback). VR-ADL, task transparency, of an ecologically valid study and testing in a variety of clinical and relevant virtual scenarios are believed to contribute to the and research settings. described illusions that virtual events and locations are actually In conclusion, relative to age-matched controls, VR-ADL relevant to the user and engaging for cognitive rehabilitation. exercises outperform the clinical predictive validity of traditional For example, a cognitive task embedded in a user-relevant assessments as an indicator of real-world difficulties in IADLs. scenario directly relates to the therapy goal of the patient and This result is very promising, but we will need advanced represents a desired outcome of the patient’s rehabilitation (eg, imaging techniques, such as amyloid-positron emission testing a virtual kitchen with cooking tasks relates to the scenario that or functional magnetic resonance imaging, to study this the patient aims to engage in independently at home). relationship and perform a longitudinal study that would The VR system used here is portable and can be manipulated correlate our results with neuroimaging data as well. to simulate different environments and different navigation Conflicts of Interest None declared. Multimedia Appendix 1 Efficacy ratio and functional impairment score formula. [JPG File, 28KB-Multimedia Appendix 1] References 1. Chaytor N, Schmitter-Edgecombe M. The ecological validity of neuropsychological tests: a review of the literature on everyday cognitive skills. Neuropsychol Rev 2003 Dec;13(4):181-197. [Medline: 15000225] 2. Franzen MD, Wilhelm KL. Conceptual foundations of ecological validity in neuropsychological assessment. In: Sbordone RJ, Long CJ, editors. Ecological Validity of Neuropsychological Testing. Boca Raton, FL: St Lucie Press; 1996. 3. Thompson CL, Henry JD, Withall A, Rendell PG, Brodaty H. A naturalistic study of prospective memory function in MCI and dementia. Br J Clin Psychol 2011 Nov;50(4):425-434. [doi: 10.1111/j.2044-8260.2010.02004.x] [Medline: 22003951] 4. Cesa GL, Manzoni GM, Bacchetta M, Castelnuovo G, Conti S, Gaggioli A, et al. Virtual Reality for Enhancing the Cognitive Behavioral Treatment of Obesity With Binge Eating Disorder: Randomized Controlled Study With One-Year Follow-up. J Med Internet Res 2013;15(6):e113 [FREE Full text] [doi: 10.2196/jmir.2441] [Medline: 23759286] 5. Rizzo AA, Buckwalter JG. Virtual reality and cognitive assessment and rehabilitation: the state of the art. Stud Health Technol Inform 1997;44:123-145. [Medline: 10175336] 6. Hall V, Conboy-Hill S, Taylor D. Using Virtual Reality to Provide Health Care Information to People With Intellectual Disabilities: Acceptability, Usability, and Potential Utility. J Med Internet Res 2011;13(4):e91 [FREE Full text] [doi: 10.2196/jmir.1917] 7. Klinger E, Chemin I, Lebreton S, Marié RM. Virtual action planning in Parkinson's disease: a control study. Cyberpsychol Behav 2006 Jun;9(3):342-347. [doi: 10.1089/cpb.2006.9.342] [Medline: 16780402] 8. Zhang L, Abreu BC, Masel B, Scheibel RS, Christiansen CH, Huddleston N, et al. Virtual reality in the assessment of selected cognitive function after brain injury. Am J Phys Med Rehabil 2001 Aug;80(8):597-604; quiz 605. [Medline: 11475481] 9. Campbell Z, Zakzanis KK, Jovanovski D, Joordens S, Mraz R, Graham SJ. Utilizing virtual reality to improve the ecological validity of clinical neuropsychology: an FMRI case study elucidating the neural basis of planning by comparing the Tower of London with a three-dimensional navigation task. Appl Neuropsychol 2009 Oct;16(4):295-306. [doi: 10.1080/09084280903297891] [Medline: 20183185] 10. Zakzanis KK, Quintin G, Graham SJ, Mraz R. Age and dementia related differences in spatial navigation within an immersive virtual environment. Med Sci Monit 2009 Apr;15(4):CR140-CR150. [Medline: 19333197] 11. Rizzo AA, Buckwalter JG, van der Zaag C. Virtual environment applications in clinical neuropsychology. In: Stanney K, editor. The Handbook of Virtual Environments. New York: Lawrence Erlbaum Associates; 2002:1027-1064. 12. Lo Priore C, Castelnuovo G, Liccione D, Liccione D. Experience with V-STORE: considerations on presence in virtual environments for effective neuropsychological rehabilitation of executive functions. Cyberpsychol Behav 2003 Jun;6(3):281-287. [doi: 10.1089/109493103322011579] [Medline: 12855084] 13. Koenig S, Crucian G, Dalrymple-Alford J, Dünser A. Assessing navigation in real and virtual environments: A validation study. 2010 Presented at: 8th International Conference on Disability, Virtual Reality and Associated Technologies; Aug 31-Sep 2, 2010; Viña del Mar/Valparaíso, Chile p. 7-16. 14. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist 1969;9(3):179-186. [Medline: 5349366] http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 11 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al 15. Souder E, Chastain JR, Williams RD. Dementia in the new millennium. Medsurg Nurs 2002 Apr;11(2):61-9; quiz 70. [Medline: 11962170] 16. Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med 2004 Sep;256(3):183-194. [doi: 10.1111/j.1365-2796.2004.01388.x] [Medline: 15324362] 17. Winblad B, Palmer K, Kivipelto M, Jelic V, Fratiglioni L, Wahlund LO, et al. Mild cognitive impairment--beyond controversies, towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Intern Med 2004 Sep;256(3):240-246. [doi: 10.1111/j.1365-2796.2004.01380.x] [Medline: 15324367] 18. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999 Mar;56(3):303-308. [Medline: 10190820] 19. Nelson AP, O'Connor MG. Mild cognitive impairment: a neuropsychological perspective. CNS Spectr 2008 Jan;13(1):56-64. [Medline: 18204415] 20. Binegar DL, Hynan LS, Lacritz LH, Weiner MF, Cullum CM. Can a direct IADL measure detect deficits in persons with MCI? Curr Alzheimer Res 2009 Feb;6(1):48-51 [FREE Full text] [Medline: 19199874] 21. Amieva H, Letenneur L, Dartigues JF, Rouch-Leroyer I, Sourgen C, D'Alchée-Birée F, et al. Annual rate and predictors of conversion to dementia in subjects presenting mild cognitive impairment criteria defined according to a population-based study. Dement Geriatr Cogn Disord 2004;18(1):87-93. [doi: 10.1159/000077815] [Medline: 15087583] 22. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, International Psychogeriatric Association Expert Conference on mild cognitive impairment. Mild cognitive impairment. Lancet 2006 Apr 15;367(9518):1262-1270. [doi: 10.1016/S0140-6736(06)68542-5] [Medline: 16631882] 23. Petersen RC, Roberts RO, Knopman DS, Geda YE, Cha RH, Pankratz VS, et al. Prevalence of mild cognitive impairment is higher in men: The Mayo Clinic Study of Aging. Neurology 2010 Sep 06;75(10):889-897. [doi: 10.1212/WNL.0b013e3181f11d85] 24. Yesavage JA, O'Hara R, Kraemer H, Noda A, Taylor JL, Ferris S, et al. Modeling the prevalence and incidence of Alzheimer's disease and mild cognitive impairment. J Psychiatr Res 2002 Oct;36(5):281-286. [Medline: 12127595] 25. Bangen KJ. Complex activities of daily living vary by mild cognitive impairment subtype. Journal of the International Neuropsychological Society 2010;16:630-639. [doi: 10.1017/S1355617710000330] 26. Goldberg TE. Performance-based measures of everyday function in mild cognitive impairment. American Journal of Psychiatry 2010;167:845-853. [doi: 10.1176/appi.ajp.2010.09050692] 27. Burton CL, Strauss E, Bunce D, Hunter MA, Hultsch DF. Functional abilities in older adults with mild cognitive impairment. Gerontology 2009;55(5):570-581. [doi: 10.1159/000228918] [Medline: 19602873] 28. Bangen KJ, Jak AJ, Schiehser DM, Delano-Wood L, Tuminello E, Han SD, et al. Complex activities of daily living vary by mild cognitive impairment subtype. Journal of the International Neuropsychological Society 2010;16:630-639. [doi: 10.1017/S1355617710000330] 29. Aretouli E, Brandt J. Everyday functioning in mild cognitive impairment and its relationship with executive cognition. Int J Geriatr Psychiatry 2010 Mar;25(3):224-233 [FREE Full text] [doi: 10.1002/gps.2325] [Medline: 19650160] 30. Teng E, Becker BW, Woo E, Cummings JL, Lu PH. Subtle deficits in instrumental activities of daily living in subtypes of mild cognitive impairment. Dement Geriatr Cogn Disord 2010;30(3):189-197 [FREE Full text] [doi: 10.1159/000313540] [Medline: 20798539] 31. Teng E, Becker BW, Woo E, Knopman DS, Cummings JL, Lu PH. Utility of the Functional Activities Questionnaire for Distinguishing Mild Cognitive Impairment From Very Mild Alzheimer Disease. Alzheimer Dis Assoc Disord 2010 Jun 29:348-353 [FREE Full text] [doi: 10.1097/WAD.0b013e3181e2fc84] [Medline: 20592580] 32. Perneczky R, Pohl C, Sorg C, Hartmann J, Komossa K, Alexopoulos P, et al. Complex activities of daily living in mild cognitive impairment: conceptual and diagnostic issues. Age Ageing 2006 May;35(3):240-245 [FREE Full text] [doi: 10.1093/ageing/afj054] [Medline: 16513677] 33. Park B, Jun JK, Park J. Cognitive impairment and depression in the early 60s: which is more problematic in terms of instrumental activities of daily living? Geriatr Gerontol Int 2013 Apr 7. [doi: 10.1111/ggi.12055] [Medline: 23560396] 34. Tarnanas I, Harper B, Majer A, Testerman R. World Intellectual Property Organization: Patentscope. 2009. Virtual Reality Interface System, EPO patent WO2010075481 URL: http://patentscope.wipo.int/search/en/WO2010075481 [accessed 2013-07-30] [WebCite Cache ID 6IUw2U6aI] 35. Perneczky R, Pohl C, Sorg C, Hartmann J, Tosic N, Grimmer T, et al. Impairment of activities of daily living requiring memory or complex reasoning as part of the MCI syndrome. Int J Geriatr Psychiatry 2006 Feb;21(2):158-162. [doi: 10.1002/gps.1444] [Medline: 16416470] 36. Rahimi F, Duval C, Jog M, Bee C, South A, Jog M, et al. Capturing whole-body mobility of patients with Parkinson disease using inertial motion sensors: expected challenges and rewards. Conf Proc IEEE Eng Med Biol Soc 2011;2011:5833-5838. [doi: 10.1109/IEMBS.2011.6091443] [Medline: 22255666] 37. Yamaguchi DA, Foloppe P, Richard E, Allain P. A dual-modal virtual reality kitchen for (re)learning of everyday cooking activities in Alzheimer’s disease. Presence: Teleoperators and Virtual Environments 2012:43-57. [doi: 10.1162/PRESa00080] http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 12 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al 38. Sheikh JI, Yesavage JA. Geriatric Depression Scale (GDS). Recent evidence and development of a shorter version. In: Brink TL, editor. Clinical Gerontology: A Guide to Assessment and Intervention. New York: The Haworth Press; 1986:165-173. 39. Visser PJ, Verhey FR, Boada M, Bullock R, De Deyn PP, Frisoni GB, et al. Development of screening guidelines and clinical criteria for predementia Alzheimer's disease. The DESCRIPA Study. Neuroepidemiology 2008;30(4):254-265 [FREE Full text] [doi: 10.1159/000135644] [Medline: 18515975] 40. Dodge H, Mattek N, Austin D, Hayes T. In-home walking speeds and variability trajectories associated with mild cognitive impairment. Neurology 2012;78(24):1946-1952. [doi: 10.1212/WNL.0b013e318259e1de] 41. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics 1982 Dec;38(4):963-974. [Medline: 7168798] 42. Tukey JW. The philosophy of multiple comparisons. Statist Sci 1991 Feb;6(1):100-116. [doi: 10.1214/ss/1177011945] 43. Akobeng AK. Understanding diagnostic tests 3: Receiver operating characteristic curves. Acta Paediatr 2007 May;96(5):644-647. [doi: 10.1111/j.1651-2227.2006.00178.x] [Medline: 17376185] 44. Slavin M, Brodaty H, Kochan N, Crawford JD, Trollor J, Draper B, et al. Prevalence and predictors of “subjective cognitive complaints” in the Sydney Memory and Ageing Study. Am J Geriatr Psychiatry 2010 Aug;18(8):701-710. [Medline: 21491631] 45. Simoes M, Castro L, Ribeiro O, Salgado T, Paz C. P03-94 Subjective Memory Complaints and Mild Cognitive Impairment. European Psychiatry 2009 Jan;24:S1093. [doi: 10.1016/S0924-9338(09)71326-4] 46. Leonardi C, Mennecozzi C, Not E, Pianesi R, Zancanaro M. Designing a familiar technology for elderly people. Gerontechnology 2008;7(2):151. [doi: 10.4017/gt.2008.07.02.088.00] 47. Slater M. Place illusion and plausibility can lead to realistic behaviour in immersive virtual environments. Philos Trans R Soc Lond B Biol Sci 2009 Dec 12;364(1535):3549-3557 [FREE Full text] [doi: 10.1098/rstb.2009.0138] [Medline: 19884149] Abbreviations AD: Alzheimer disease ADL: Activities of Daily Living AUC: area under the curve BADL: basic activities of daily living GDS: geriatric depression scale IADL: instrumental activities of daily living MANOVA: multivariate analyses of variance MCI: mild cognitive impairment MMSE: mini-mental state examination RAVLT: Rey Auditory Verbal Learning Test ROC: receiver-operating curve RR: relative risk UI: user interface VE: virtual environment VR: virtual reality VR-DOT: Virtual Reality Day-Out Task VRPN: virtual reality peripheral network Edited by G Eysenbach; submitted 18.06.13; peer-reviewed by W Greenleaf, G Russo; comments to author 08.07.13; revised version received 22.07.13; accepted 22.07.13; published 06.08.13 Please cite as: Tarnanas I, Schlee W, Tsolaki M, Müri R, Mosimann U, Nef T JMIR Serious Games 2013;1(1):e1 URL: http://games.jmir.org/2013/1/e1/ doi: 10.2196/games.2778 PMID: 25658491 ©Ioannis Tarnanas, Winfried Schlee, Magda Tsolaki, René Müri, Urs Mosimann, Tobias Nef. Originally published in JMIR Serious Games (http://games.jmir.org), 06.08.2013. This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Serious Games, is properly cited. The complete http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 13 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al bibliographic information, a link to the original publication on http://games.jmir.org/, as well as this copyright and license information must be included. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 14 (page number not for citation purposes) XSL FO RenderX http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png JMIR Serious Games JMIR Publications

Ecological Validity of Virtual Reality Daily Living Activities Screening for Early Dementia: Longitudinal Study

Download PDF
14 pages

Loading next page...
 
/lp/jmir-publications/ecological-validity-of-virtual-reality-daily-living-activities-pehgZsqMwh
Publisher
JMIR Publications
Copyright
Copyright © The Author(s). Licensed under Creative Commons Attribution cc-by 4.0
ISSN
2291-9279
DOI
10.2196/games.2778
Publisher site
See Article on Publisher Site

Abstract

Background: Dementia is a multifaceted disorder that impairs cognitive functions, such as memory, language, and executive functions necessary to plan, organize, and prioritize tasks required for goal-directed behaviors. In most cases, individuals with dementia experience difficulties interacting with physical and social environments. The purpose of this study was to establish ecological validity and initial construct validity of a fire evacuation Virtual Reality Day-Out Task (VR-DOT) environment based on performance profiles as a screening tool for early dementia. Objective: The objectives were (1) to examine the relationships among the performances of 3 groups of participants in the VR-DOT and traditional neuropsychological tests employed to assess executive functions, and (2) to compare the performance of participants with mild Alzheimer’s-type dementia (AD) to those with amnestic single-domain mild cognitive impairment (MCI) and healthy controls in the VR-DOT and traditional neuropsychological tests used to assess executive functions. We hypothesized that the 2 cognitively impaired groups would have distinct performance profiles and show significantly impaired independent functioning in ADL compared to the healthy controls. Methods: The study population included 3 groups: 72 healthy control elderly participants, 65 amnestic MCI participants, and 68 mild AD participants. A natural user interface framework based on a fire evacuation VR-DOT environment was used for assessing physical and cognitive abilities of seniors over 3 years. VR-DOT focuses on the subtle errors and patterns in performing everyday activities and has the advantage of not depending on a subjective rating of an individual person. We further assessed functional capacity by both neuropsychological tests (including measures of attention, memory, working memory, executive functions, language, and depression). We also evaluated performance in finger tapping, grip strength, stride length, gait speed, and chair stands separately and while performing VR-DOTs in order to correlate performance in these measures with VR-DOTs because performance while navigating a virtual environment is a valid and reliable indicator of cognitive decline in elderly persons. Results: The mild AD group was more impaired than the amnestic MCI group, and both were more impaired than healthy controls. The novel VR-DOT functional index correlated strongly with standard cognitive and functional measurements, such as mini-mental state examination (MMSE; rho=0.26, P=.01) and Bristol Activities of Daily Living (ADL) scale scores (rho=0.32, P=.001). http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 1 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Conclusions: Functional impairment is a defining characteristic of predementia and is partly dependent on the degree of cognitive impairment. The novel virtual reality measures of functional ability seem more sensitive to functional impairment than qualitative measures in predementia, thus accurately differentiating from healthy controls. We conclude that VR-DOT is an effective tool for discriminating predementia and mild AD from controls by detecting differences in terms of errors, omissions, and perseverations while measuring ADL functional ability. (JMIR Serious Games 2013;1(1):e1) doi: 10.2196/games.2778 and rehabilitation, including virtual cities [9,10], school Introduction classrooms [11], and supermarkets [9,12]. As outlined previously, ecological validity can be seen as a key component A decade ago, Chaytor and Schmitter-Edgecombe [1] reviewed for assessing cognitive skills that are relevant for functional the ecological validity of neuropsychological tests by evaluating tasks in real-world contexts [13]. The results of such studies their efficacy in measuring everyday cognitive skills. They suggest that the use of VEs is valuable in enhancing our ability identified 6 studies that explored the issue of ecological validity to assess the functional behaviors of individuals with executive of executive functioning tests. The studies differed in terms of dysfunction in activities of daily living. the specific tests used, although both traditional (veridicality) and verisimilitude tests were employed. Veridicality refers to Activities of daily living (ADL) can be classified into basic the extent to which results of an assessment instrument are activities of daily living (BADL) and instrumental activities of related to scores on other tests that predict the performance of daily living (IADL) [14]. BADL is composed of more basic real-world tasks [2]. By contrast, verisimilitude refers to the self-care behaviors, such as ambulating, dressing, grooming, similarity between the task demands of the test and the demands bathing, feeding, and toileting. By contrast, IADL facilitates imposed in the everyday environment. independent living through behaviors such as transportation, telephone use, meal preparation, medication management, Their findings indicated that executive tests were not financial management, housekeeping, laundry, and shopping. significantly correlated with self-reported measures, but all IADL questionnaires play a vital role in assessing functional studies reviewed revealed significant associations between abilities and evaluating the impact of cognitive impairment on executive tests (traditional and verisimilitude) and everyday everyday activities in older adults [15]. abilities as measured by clinician ratings and informants’ (eg, relatives’) questionnaires. To date, commentaries on ecological IADL independence is one of the defining features that validity have primarily emphasized the increased consideration characterize normal aging from mild cognitive impairment of this concept in assessments of neurologically impaired (MCI) and dementia. As part of the diagnostic criteria for MCI, individuals, particularly in rehabilitative and forensic contexts. an individual must be classified as independent for BADL, but However, there are instances in which patients perform normally can have minimal disturbance in IADL [16,17]. Since the early on traditional executive tests, yet clearly have executive descriptions of MCI [16], there has been increasing interest in impairments in their daily lives [3]. its clinical characterization and prognosis [18,19]. In previous reports [20], people with MCI exhibited poorer cognitive Virtual environments (VEs) have numerous features that make functioning than healthy controls, but were not as impaired as them attractive for assessment and rehabilitation purposes. In patients with dementia were. contrast to traditional executive test measures, VEs actively engage participants by allowing them to be involved in a task Prognostic studies have stressed the necessity of this nosological while at the same time being less focused on the fact that they entity as a risk, or prodromal state for dementia, because of the are being tested [4,5]. More recently, researchers have used high rate of conversion of MCI to dementia (10%-15% of virtual reality (VR) systems for detailed response measurement patients who meet the criteria of amnestic MCI develop and analysis to examine specific behaviors characteristic of Alzheimer-type dementia per year, up to 80% at 5-year patients with executive dysfunction or people with intellectual follow-up) [21,22,18]. In Europe, approximately 17% of the disabilities [6]. Klinger and colleagues [7] examined planning senior population who have not been diagnosed with dementia deficits in patients with Parkinson’s disease compared to meet the current criteria for MCI [23] and MCI prevalence age-matched controls in a virtual supermarket. The researchers increases with age [24]. described the patients’ paths through the supermarket as Characterizing impairment using the IADL questionnaire has characterized by numerous stops, turns, and hesitancies as been criticized for several reasons. First, no objective standard compared to the paths of controls. Zhang and colleagues [8] exists as to the practical or theoretical definition of minimal used a virtual kitchen to assess selected cognitive functions of functional impairment in predementia [16,17]. For example, traumatic brain injury patients compared to normal volunteers. does functional disturbance entail perceptible impairment on a Task transparency and relevant functional tasks, such as finding few IADL tasks, such as shopping and meal preparation? Or is one’s way through a VE or remembering groceries for preparing it better understood as some problems across many commonly a breakfast in a virtual kitchen, are examples in which ecological assessed IADL tasks? Clinical judgment is called for by the validity can be described as enhanced when compared to abstract expert panel that created these standards [25,26], but the general traditional assessments of cognitive functions. A variety of VEs clinician or researcher is without much guidance regarding how have already been developed to enhance functional assessment to assess IADL impairment in predementia patients. Several http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 2 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al options exist, including performance-based tasks and participants [33], and (3) in mild AD and MCI patients, the rate questionnaires or interviews (with and without informant of change in individual performance in VR-DOT measures reports). However, different methods of assessing functional could predict the cognitive decline over 2 to 3 years. abilities require different estimates of IADL independence [27]; The main objective in developing the VR-DOT was to improve each method has advantages and disadvantages. the ecological validity of executive function measures by using Recently, it became clear that IADL, versus BADL, is a better a verisimilitude approach. We also proposed a framework to diagnostic instrument for predementia [28-31]. Although these objectively assess the functional impairment of elderly people studies were carried out in different countries and used different through an ecological and clinical longitudinal experiment using instruments to assess impairments in ADL, they all indicate that VR-DOT. Our motivation was to correlate this new instrument people who meet the criteria described for predementia show (VR-DOT) with normal cognitive neuropsychological measures some functional impairment in activities of daily living. In the and recent psychomotor discoveries regarding psychomotor clinical setting, rate of change in complex ADL performance velocity change and cognitive decline to see if the VR-DOT may be more useful than a cross-sectional measurement, which offers better sensitivity and specificity in assessing and could misclassify individuals into a nonimpaired category in predicting cognitive decline using only a virtual environment. activities of daily living [32]. Furthermore, the rate of change is a parameter that could be manipulated by designing Methods naturalistic VEs or serious games that can train the higher Virtual Reality Test Setup cognitive functions. We designed a fire evacuation Virtual Reality Day-Out Task Overview (VR-DOT) environment to (1) determine what kind of real-time The VR hardware consisted of a Pentium-based computer with cognitive and psychomotor performance and errors are 4 MB RAM, Intel Quad Core processor, and NVIDIA graphic associated with functional impairment in activities of daily cards with 512 MB memory. Other sensors used were a LEAP living, (2) identify the patterns and cutoff values of the these motion sensor (Leap Motion Inc, San Francisco, CA, USA) and cognitive and psychomotor profiles as independent predictors a Kinect camera (Microsoft Corp, Seattle, WA, USA). The of functional impairment in healthy elderly participants, LEAP motion sensor is still not commercially available at the single-domain amnestic MCI patients, and patients with mild time of this writing, but we were selected by their development Alzheimer-type dementia (AD), and (3) controlling for baseline team to use the hardware for our experiments (Figure 1). performance, objectively measure performance change over 2 Software Components to 3 years. Modeling was done using Maya software (Autodesk Inc, San We hypothesized that with VR-DOT (1) dementia and MCI Rafael, CA, USA) to create models and scenes. Then, the scenes patients will show significantly impaired independent were exported to Virtools, a 3D authoring tool (Dassault functioning in ADL and distinct performance profiles, (2) among Systèmes, Inc, Vélizy-Villacoublay Cedex, France) that handled patients with dementia or MCI, such impairment will be all programming including interactivity, setting, and associated with the degree of cognitive impairment and cognitive configuration. Microsoft Kinect software development kit (SDK) neurophysiological measures, whereas impaired functioning (Microsoft Corp, Seattle, WA, USA) was used to analyze will be only associated with sociodemographic and gestures and movements and a user interface (UI) system was anxiety/depression symptoms in healthy controls because developed using Microsoft Kinect SDK and precommercial subclinical levels of cognitive impairment and depression have Alterniity algorithms developed by Ioannis et al [34]. been associated with IADL impairment in mentally healthy http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 3 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 1. Virtual reality day-out task (VR-DOT) participant setup. allowing VR applications to read the skeletal joints as trackers The Naturalistic Setting of Executive Function in Virtual using any VRPN client. Additionally, the toolkit can also Reality Activities of Daily Living emulate keyboard input triggered by body posture and specific Virtual reality activities of daily living (VR-ADL) consists of gestures. This allows the user to add custom body-based control 2 modules: the VR-DOT and VR basic instrumental activities mechanisms to existing off-the-shelf games that do not provide of daily living (VR-IADLs). The VR-DOT is a complex task official support for depth sensors. breakdown and then a rehearsal exercise of a fire evacuation More specifically, the VR-DOT module is a naturalistic task drill consisting of 6 different scenarios of increasing difficulty. that requires multitasking in a fire evacuation drill setting with We chose to examine the VR-DOT virtual fire evacuation drill 6 different simulated fire situations (from easy to more difficult) in this study (Figure 2), based on the literature indicating that taking place at a virtual apartment block with 3 levels and 5 activities of daily living requiring complex reasoning are apartments per level. It is used to examine prospective memory sensitive to cognitive and functional impairment [35]. User as well as reasoning in a complex emergency routine in which tracking was performed by a flexible action and articulated older adults prioritize, organize, initiate, and complete a number skeleton toolkit (FAAST; University of Southern California, of subroutines to evacuate safely from an apartment level CA, USA), a middleware to facilitate integration of full-body (second floor) to the ground area (eg, determine and gather control with games and VR applications, using either OpenNI information on the size of the fire, avoid smoke). Previous or the Microsoft Kinect for Windows skeleton tracking software. research shows that motion tracking while navigating a virtual FAAST includes a custom virtual reality peripheral network environment is a valid and reliable indicator of cognitive decline (VRPN) server to stream up to 4 user skeletons over a network, in elderly persons. (ie, [36]). http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 4 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 2. Sample sequential virtual reality day-out task (VR-DOT) screenshots, showing different tasks and viewpoints. Procedure Functional and Psychomotor Rate of Change Participants A measurement of the rate of change of functional impairment was computed from all information collected from the LEAP A total of 405 elderly people were screened during 2010 in 2 motion and the Microsoft Kinect camera sensor inside the Alzheimer Hellas, Non-Government Organization (NGO) day VR-DOT. Simple performance-based functional impairment clinics of the Papanikolaou University Hospital in Thessaloniki, measures have been used previously, but not with data collected Greece. Ethics approval was obtained from the Papanikolaou from motion sensors [37]. At baseline, a simple quantitative University Hospital Ethics Committee. Inclusion criteria were ratio of efficacy was computed by dividing the total time (in age older than 60 years, meeting the diagnostic criteria for MCI sec) spent by the participant performing the listed activities by as defined in Petersen et al [18], living in the community, and the total time spent in VR-DOT (efficacy ratio). Then, 4 activity providing informed consent approved by the Ethics Committee. parameters with a high likelihood of corresponding to functional Exclusion criteria were living in an assisted-living residence, decline were collected: (1) omission of 1 of the activities (k ), cognitive functioning suggesting a possible diagnosis of (2) repetition of the same activity (k ), (3) incorrect order in dementia (see subsequent description), previous diagnosis of dementia, other psychiatric disorder according to the Diagnostic performing the activities (k ), and (4) number of attempts before and Statistical Manual of Mental Disorders (Fourth Edition, completing a given activity (k ). The first quantitative ratio of Text Revision; DSM-IV-TR) at the time of recruitment, efficacy was then adjusted by these parameters. This led to a presenting a moderate or higher degree of fear or dislike of functional impairment score according to the formula presented computers (technophobia), presenting a moderate or higher in Multimedia Appendix 1. degree of disability because of other conditions than MCI, and To determine values of the model parameter set (k , k , k , k ), severe language impairments that would compromise active 1 2 3 4 participation. we ran a pilot with healthy participants (n=25; mean age 73.7 years, SD 4.0), amnestic single-domain MCI patients (n=26; The baseline psychomotor evaluation inside VR-DOT consisted mean age 74.2 years, SD 2.0), and patients with mild AD (n=24; of a number of simple and complex measures addressing the mean age 76.7 years, SD 3.0). Second, multiple-model parameter participant’s ability to understand and perform with accuracy sets (k , k , k , k ) to produce a good fit were selected if their 1 2 3 4 specific physical performance tasks. These tasks/metrics were: associated scores were both strongly and positively correlated Grip Strength with the mini-mental state examination (MMSE) scores, as well as being strongly and negatively correlated with IADL scores Forearm muscle strength was measured in kilograms by a using a nonparametric Spearman correlation coefficient as the hand-held Jamar A dynamometer. For this analysis, we used criterion distance of good fit. For our analyses, the final the best of 3 attempts in the dominant hand. functional impairment score (k , k , k , k ) was calibrated using 1 2 3 4 Timed Walk on the Treadmill the combination of the mean of the parameters, which was The time (to 0.1 s) required for a participant to walk a 4.6-m selected as the model parameter set during the second step of course at his or her usual pace after starting from a standstill the fitting procedure. was recorded by stopwatch. We converted the results to meters per second. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 5 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Number of Steps to Walk Course on the Treadmill Significant effects were further tested with post hoc tests that were corrected for multiple comparisons using Tukey’s Honestly A technician recorded the number of steps required to walk the Significant Difference (HSD) [42]. We used similar statistical 4.6-m course. Hereafter, we refer to stride length, which is models to estimate the RR of having significant VR-DOT derived by dividing the distance walked by the number of steps. difficulty or inability (relative to no or mild difficulty) and the Finger-Tapping Test RR of cognitive decline (relative to no or mild difficulty) at Using their dominant hand, participants tapped in midair, just year 3 for upper extremity function. For a given performance above the LEAP motion sensor, with the index finger as fast as measure, the first year 1 value and the third year 1-3 slope of they could for 15 seconds. change were treated as separate predictor variables. For each outcome, a separate regression model was run for each predictor After applying the inclusion/exclusion criteria, 232 participants performance variable, adjusting for age, gender, and the were included in this study. The participants were measured VR-DOT task of more difficulty with, or disability in, the each year of the 3-year study period. Dropouts after baseline outcome measure between years 1 and 3. Next, we were 27 (11.6%); hence, 205 participants completed all simultaneously entered all predictor performance variables into measurements and their data are included in this study (N=205; a second set of models, adjusting for the same covariates. The male=88, female=117; mean age 72.73 years, SD 6.89; mean component variables from each model were entered, in turn, education 12.53 years, SD 3.20; mean baseline MMSE 24.75, into a stepwise backward regression for the respective outcomes, SD 2.18). with a P value to enter the model set at <.10. This procedure yielded a set of simpler, more parsimonious final models. All Neuropsychological Assessment statistical analyses were run using SPSS 19.0 statistical software Cognitive assessment was performed by means of a (IBM Corp, Armonk, NY, USA). neuropsychological test battery designed to comprehensively evaluate attention, working memory, memory, executive Results functioning, and language. In addition to the cognitive assessment, all groups were also assessed for depression with Demographics and baseline scores for all groups are shown in the geriatric depression scale (GDS) [38]. We also chose the Table 1. Digit Symbol (DSym), Functional Activities Questionnaire (FAQ), Neuropsychiatric Inventory brief questionnaire form, After corrected with age, gender, and education status, our Apathy item (NPI-Q Apathy), Neuropsychiatric Inventory brief results showed that the VR-DOT functional index was correlated questionnaire form, Depression item (NPI-Q Depression), Rey strongly with standard cognitive and functional measurements, Auditory Verbal Learning Test T (RAVL), Trailmaking Test such as MMSE (rho=0.26, P=.01) and Bristol ADL scores A (TMT-A), Trailmaking Test B (TMT-B), Trailmaking Test (rho=0.32, P=.001), thus accurately differentiating from healthy B minus Trailmaking Test A (TMT-B-A), the Bristol ADL control participants (Table 2). scale, and the short form of the Blessed ADL scale for this study In the prediction models for individual performance measures because they were evaluated and validated for the Greek (not shown), the VR-DOT and upper extremity function population [39]. The original Bristol and short-form Blessed psychomotor performance (finger tapping, etc) at year 3 for the scales consist of 20 and 11 items, respectively. MCI and mild AD group, as well as the slopes of change, had a significance of P<.10. Compared with the control, weaker Statistical Analysis results of the MCI and mild AD independently predicted Performance results from the VR-DOT, gait velocity assessment, cognitive decline at year 3 in all 3 domains (VR-DOT, and neuropsychological tests were analyzed using multivariate neuropsychological, and gait velocity assessment). The change analyses of variance (MANOVA) in mixed designs with group slope for upper extremity function inside the VR-DOT was also as the between-subject factor using linear mixed-effects models associated with the outcome. with random intercept and slope to estimate the annual rate of change between study years 1 and 3 for each performance For functional independence, the healthy group showed better measure of each participant [40,41]. Before this approach, we functional adjustment than the MCI and mild AD group plotted numerous individual trajectories for the gait velocity according to VR-DOT total monitoring data. When the amnestic performance variables by using robust splines to smooth the MCI group was examined using the VR-DOT total score, curves. The consistent linearity of the trajectory patterns justified cognitive domain and gait velocity assessment showed a the use of linear models. Gait speed and stride length were similarly impaired profile as cognitive functioning, after adjusted to a 50-cm knee-heel length and this adjustment was controlling for age, education, and GDS score. The mild AD included in the models when it reached 10% significance. We patients showed a higher degree of functional impairment than also used multinomial Poisson log-linear models to estimate both healthy controls and amnestic MCI patients in life the relative risk (RR) of cognitive decline relative to efficacy, activities, and participation subscales, respectively, and in the gait velocity, and neuropsychological assessment at year 2 and VR-DOT mobility domain. The total VR-DOT functional ability year 3 (2010-2013) for VR-DOT and receiver-operating curve measures showed a consistent functional impairment of mild (ROC) analysis was conducted on VR-DOT, MMSE, the AD and amnestic MCI in comparison with healthy participants. RAVLT, and the Bristol and Blessed ADL scale scores. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 6 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Table 1. Participant demographics and scores on cognitive tests for all participants, healthy controls, patients with amnestic-type mild cognitive impairment (aMCI), and patients with mild Alzheimer-type dementia (AD). Group All participants Controls aMCI Mild AD N=205 n=72 n=65 n=68 Age, mean (SD) 72.73 (6.8) 72.63 (5.06) 72.78 (6.21) 72.58 (6.21) Sex, n (%) Male 88 (43%) 25 (38%) 30 (43%) 33 (46%) Female 117 (57%) 37 (62%) 40 (57%) 40 (54%) Education, mean (SD) 15.6 (3.0) 16.1 (2.9) 15.7 (3.0) 14.6 (3.2) Test, mean (SD) MMSE 24.75 (2.18) 29.1 (1.0) 26.1 (1.8) 23.4 (2.0) RAVLT delayed recall 3.7 (4.0) 7.4 (3.7) 2.9 (3.3) 0.7 (1.6) GDS 1.4 (1.4) 0.8 (1.1) 1.6 (1.4) 1.6 (1.4) NPI-Q depression 0.2 (0.5) 0.1 (0.3) 0.2 (0.5) 0.4 (0.6) NPI-Q apathy 0.2 (0.6) 0.01 (0.1) 0.2 (0.5) 0.5 (0.8) FAQ 4.8 (6.4) 0.1 (0.6) 3.8 (4.4) 12.7 (6.7) TMT-A 46.6 (25.5) 36.3 (13.0) 44.2 (21.7) 64.8 (34.5) TMT-B 134.5 (80.2) 89.3 (44.3) 130.8 (73.2) 200.5 (86.6) TMT-B–A 88.0 (66.9) 53.0 (38.8) 86.6 (63.1) 135.8 (74.3) Bristol ADL scores 6.88 (0.56) 4.46 (0.5) 5.59 (0.9) 10.59 (0.9) Blessed ADL impairment score 2.87 (0.26) 1.85 (0.27) 2.38 (0.56) 4.38 (0.56) Geriatric depression scale 5.19 (5.0) 4.59 (4.1) 5.49 (5.76) 5.29 (4.45) Digit Symbol 37.4 (12.9) 45.8 (10.2) 37.0 (11.1) 27.6 (12.5) Gait speed (m/s), mean (SD) Combined 0.91 (0.22) 0.96 (0.23) 0.91 (0.24) 0.86 (0.20) Women 0.85 (0.14) 0.94 (0.24) 0.84 (0.04) 0.77 (0.14) Men 0.98 (0.13) 1.00 (0.21) 1.01 (0.03) 0.95 (0.04) Tapping speed dominant (taps/second), mean (SD) Combined 3.79 (0.8) 3.87 (0.8) 3.77 (0.81) 3.74 (0.8) Women 3.48 (0.78) 3.53 (0.71) 3.49 (0.84) 3.43 (0.77) Men 4.23 (0.75) 4.29 (0.77) 4.21 (0.73) 4.19 (0.75) Tapping speed non-dominant (taps/second), mean (SD) Combined 3.60 (0.67) 3.63 (0.64) 3.61 (0.71) 3.58 (0.7) Women 3.73 (0.59) 3.41 (0.53) 3.38 (0.64) 3.33 (0.61) Men 3.92 (0.63) 3.91 (0.65) 3.96 (0.63) 3.90 (0.62) MMSE: mini-mental state examination, RAVLT: Rey Auditory Verbal Learning Test, GDS: Geriatric Depression Scale, NPI-Q Depression: Neuropsychiatric Inventory brief questionnaire form, Depression item, NPI-Q Apathy: Neuropsychiatric Inventory brief questionnaire form, Apathy item, FAQ: Functional Activities Questionnaire, TMT-A: Trailmaking Test ), TMT-B: Trailmaking Test B, TMT-B–A: Trailmaking Test B minus Trailmaking Test A, ADL: Activities of Daily Living. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 7 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Table 2. The correlation matrix between Virtual Reality Day-Out Task (VR-DOT) functional index, mini-mental state examination (MMSE), and Bristol Activities of Daily Living (ADL) when controlling for age, gender, and education status. Test VR-DOT MMSE Bristol ADL rho P value rho P value rho VR-DOT 1 MMSE 0.26 .01 1 Bristol ADL 0.32 .01 0.43 .01 1 The AUC indicates that VR-DOT was the most powerful of all Predictors of Functional Status (Regression Analyses) tests in discriminating normal controls from the MCI groups, When the entire sample was analyzed together, attention, reaching optimal results with a cutoff point of 20 (97% psychomotor, and memory summary scores explained a total sensitivity, 100% specificity, 100% positive predictive values, variance of 8.2% and 0.8% of the VR-DOT. When depression and 96% negative predictive value). Figure 3 shows the ROC and age were entered in both former models, the VR-DOT score of the normal control and MCI groups for VR-DOT total score. was predicted by depression symptoms as measured by the GDS Exploratory Prediction of Conversion to Alzheimer (19.2%) only in the healthy group. By contrast, VR-DOT total Disease (VR-DOT Performance Rate of Change) score was only predicted by psychomotor and executive functions (8.1%) among mild AD and amnestic MCI patients. According to the results, the task that better differentiated among For the amnestic MCI group, VR-DOT was predicted by healthy controls, amnestic MCI, and mild AD participants at executive functions and psychomotor profile only, with a total baseline, year 2, and year 3 follow-up, was the VR-DOT variance explained of 17.3% for amnestic MCI. VR-DOT score performance score (efficacy ratio). The VR-DOT and Bristol was predicted only by executive function and psychomotor and Blessed ADL scale scores were included as predictor profiles in amnestic MCI patients and by executive function, variables in a series of exploratory independent regression psychomotor profiles, and GDS scores in mild AD. When a analyses. Figure 4 shows the individual predictive power of the ROC analysis was carried out on the Bristol and Blessed ADL 3 test variables of interest (VR-DOT, Bristol, and Blessed ADL scales, they explained 9.1% variance of VR-DOT total profiles. scale scores), ranked in ascending order according to the magnitude of their odds ratios. The VR fire evacuation ROC analysis was conducted on VR-DOT, MMSE, RAVLT, performance score rate of change (VR-DOT REff) emerged as and Bristol and Blessed ADL scale scores obtained from the the best predictor of conversion to AD in this sample (VR-DOT; amnestic MCI and mild AD groups and the sensitivity, P=.008; OR 2.8, 95% CI 1.3-6.0; Nagelkerke R =0.564), with specificity, and cutoff values of both the scales were determined the regression model correctly classifying 88% of participants. (Table 3). The optimal cutoff score of the Bristol scale was 20 This was followed by Bristol ADL (P=.03; OR 1.9, 95% CI in differentiating amnestic MCI from mild AD with a sensitivity of 100% and specificity of 74.2%, and area under the curve 1.1-3.5; Nagelkerke R = 0.563), and the Blessed ADL (P=.01; (AUC) of 0.883 (95% CI 0.781-0.975). The optimal cutoff score OR 1.4, 95% CI 1.1-1.9; Nagelkerke R =0.466). of the modified Blessed scale was 10.5 in differentiating The resulting regression model revealed that the VR-DOT amnestic MDI from mild AD with a sensitivity of 100%, performance score threshold variable was a significant predictor specificity of 71%, and AUC 0.872 (95% CI 0.791-0.977). Post of conversion to AD in the regression equation (beta=–1.092, hoc analysis revealed that among the 3 groups, the mild AD P=.01) with OR 3.0 (95% CI 1.3-7.0). Using a cutoff score of group had the lowest scores in ADL, episodic memory, and less than 20 on the VR-DOT subscale achieved a sensitivity of VR-DOT (P<.001). 100% and a specificity of 94%. Table 3. Area under the curve (AUC) for standard neuropsychological test scores and Virtual Reality Day-Out Task (VR-DOT) for healthy controls versus patients with amnestic-type mild cognitive impairment (aMCI) and patients with aMCI versus patients with mild Alzheimer-type dementia (AD). Test Healthy control vs aMCI aMCI vs mild AD AUC (95% CI) P value AUC (95% CI) P value MMSE 0.79 (0.68, 0.91) <.001 0.99 (0.97, 1.00) <.001 Bristol scores ADL 0.75 (0.62, 0.88) <.001 0.88 (0.78, 0.97) <.001 Blessed score ADL 0.77 (0.64, 0.89) .002 0.87 (0.79, 0.98) .02 RAVLT delayed recall 0.82 (0.77, 0.93) .001 0.88 (0.79, 0.98) <.001 DOT-VR 0.96 (0.88, 0.99) <.001 0.95 (0.88, 1.00) <.001 MMSE: Mini-Mental State Examination, ADL: Activities of Daily Living, RAVLT: Rey Auditory Verbal Learning Test. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 8 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al Figure 3. Receiver-operating curve (ROC) for the Virtual Reality Day-Out Task (VR-DOT) total score when discriminating among nondemented (healthy controls), amnestic mild cognitive impairment (aMCI), and patients with mild Alzheimer-type dementia (mild AD). Figure 4. Odds ratios from exploratory individual regression analyses using VR-DOT, Bristol, and Blessed ADL scale scores rate of change as predictors for conversion from mild cognitive impairment to Alzheimer disease (bars represent 95% CI). reality quantitative performance measure of functional ability Discussion (VR-DOT) showed adequate psychometric properties (ie, discriminant power) to contribute to a predementia diagnosis. There is still debate as to the utility of MCI as a diagnostic In addition, functional measures based on quantitative rates of category. Many older people report subjective cognitive the number and quality of ADL performed seem to be more complaints in the absence of objective impairment [43] and not sensitive to identifying functional impairment in predementia all such complaints are predictive of dementia [44]. MCI may than those based on a subjective judgment of disability. be viewed as being on a continuum from normal aging to dementia and the present data show a large overlap between As a result of this paradigm shift, and in light of previous and groups that coheres with this view [45]. In that context, only a the present results, it would be very helpful for clinicians, few studies have systematically examined the rate of change in caregivers, and health-system managers if MCI definitions complex ADL performance as a predictor of cognitive decline. included an objective measure of impairment of functional abilities as a clinical feature inherent to MCI. We found that Our results show that functional impairment is a defining feature VR-DOT has greater sensitivity and specificity, as well as of both amnestic MCI and mild AD, and that the impairment having both positive and negative predictive values compared showed by amnestic MCI patients is partially dependent on the to other screening tests in discriminating amnestic MCI and degree of their cognitive impairment. Furthermore, a virtual http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 9 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al mild AD from normal aging. In summary, these and previous was in the development and empirical study of a novel VR results emphasize the presence of qualitative and quantitative environment (VR-DOT) that was less structured and that more functional impairments of both basic and complex ADL in closely resembled actual everyday errands than existing predementia as a logical consequence of cognitive impairment. questionnaires. This research demonstrated that tests of Although dementia is characterized by a more severe degree of verisimilitude may be better predictors of real-world behaviors disability than predementia, the World Health Organization than many of the most commonly employed traditional executive (WHO) International Classification of Functioning, Disability function tests. and Health (ICF) conceptualization of disability would include Our approach with VR-DOT is part of a general effort to predementia as a disabling condition, although to a lesser degree. manifest marked impairment in cognitive performance, The need for a better definition of disability as a diagnostic particularly executive functions during everyday activities by criterion (putatively by shifting from a categorical notion of means of VR (VR-ADL). Studies directly investigating ADL able/disabled to a more spectrum/gradual approach) to have found mild and tardive impairment in MCI, and a relation discriminate predementia from dementia must not conceal the with certain executive functions, but the targeted ADL were fact that dementia patients have their own health/functional very simple tasks, such as memorizing a telephone number or assistance needs. walking a few meters, and have always been strictly limited to Moreover, given the moderately good psychometric properties the accuracy domain, excluding any performance or a demonstrated in our study of the VR-ADL in discriminating rate-of-change factor. The purpose of VR-DOT was (1) to healthy from predementia and mild dementia patients, assessing investigate performance, in an experimentally controlled real-time functional ability would improve the identification of manner, on a complex ADL (planning and evacuating a fire predementia patients, and the use of objective, VR qualitative under time pressure) that is more indicative of the true quality and/or quantitative impairment of functional abilities as a of life of senior citizens, and (2) to scrutinize its cognitive diagnostic criterion should be further explored. Goldberg et al structure as a diagnostic instrument, which can screen functional [26] found that a sensitive performance-based measure they impairments at a very early phase of AD. With regard to real-life developed (the University of California, San Diego ADL, this investigation presents the advantage and innovation Performance-Based Skills Assessment; UPSA) had a remarkably of a VR quantitative scoring grid of a very complex set of good discriminant power to distinguish healthy participants sequential activities under demanding time constraints. from amnestic MCI participants (AUC 0.84), and to distinguish This study found that VR-DOT is comparatively better in amnestic MCI patients from patients with AD (AUC 0.88). detecting amnestic MCI from normal aging individuals. From Hence, the inclusion of functional competence measures seems quantitative and qualitative data extracted from VR-DOT, a convenient for the screening and early identification of functional index was computed, validated, and compared with neurodegenerative processes characterized by cognitive current clinical rating scales. Results of this pilot study are impairment. promising and must be substantiated with a larger sample and The rates of change in complex everyday activities, easily in another assessment setting to evaluate its reproducibility. determined in longitudinal practice settings, provide important Verisimilitude instruments, such as VR-DOT, can potentially prognostic information for late-life disability and death that are play valuable roles in both executive function assessment and independent of the predictive value of a performance intervention and, consequently, may help place clinical measurement obtained at a single point in time, which could be neuropsychology on firmer scientific ground. Researchers and inaccurate because of recent injury or illness. By predicting clinicians have the responsibility and opportunity to design, decline in ADL and IADL, upper extremity functionality, and test, and implement effective therapeutic strategies to improve, more generalized daily activities, longitudinal views capture or at least preserve, functional and cognitive functioning in broader deteriorations in function within an individual, predementia. suggesting a shared causal pathway. For these purpose, it is assumed that the visual quality and Our study has limitations. Although we used a population-based realism of the VEs are of central importance for patients to cohort, the exclusion from the analyses of participants with recognize and acknowledge the relevance of the task and context technophobia [46] may have introduced bias and reduced the at hand. Essential characteristics of virtual scenarios and tasks generalizability of the results. Although we only observed linear (ie, transparency, believability, plausibility, and relevance) are patterns in the many performance trajectories that we plotted, summarized under the term “realism” to describe that the patient some individual trajectories could have been nonlinear causing can recognize the employed tasks and scenarios and refer to inaccurate estimates of annual performance change. Our them based on past experiences. VR-ADLs capture the patient’s statistical models contained a limited number of covariates. interest and improve long-term motivation to use the virtual Although the addition of comorbid conditions to the models did tasks at high frequencies. Transparency and realism in a broader not significantly alter the results, we may have omitted important sense can relate to plausibility and place illusions that are confounders. described by Slater [47]. Plausibility illusion refers to the fact that the user believes the virtual scenario is actually occurring. The present research described the ecological validity of It is caused by events and the scenario relating directly to the verisimilitude and traditional activities of daily living measures user (eg, the virtual character talking to the user). Place illusion and the characterization of various subcomponents of the refers to the sensation that the user is actually situated in the executive function system. The unique contribution of this study displayed location and is described in relation to sensorimotor http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 10 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al contingencies of the VR system (eg, user interaction, tracking, demands (cognitive, motor, visual), easily allowing the creation and multimodal user feedback). VR-ADL, task transparency, of an ecologically valid study and testing in a variety of clinical and relevant virtual scenarios are believed to contribute to the and research settings. described illusions that virtual events and locations are actually In conclusion, relative to age-matched controls, VR-ADL relevant to the user and engaging for cognitive rehabilitation. exercises outperform the clinical predictive validity of traditional For example, a cognitive task embedded in a user-relevant assessments as an indicator of real-world difficulties in IADLs. scenario directly relates to the therapy goal of the patient and This result is very promising, but we will need advanced represents a desired outcome of the patient’s rehabilitation (eg, imaging techniques, such as amyloid-positron emission testing a virtual kitchen with cooking tasks relates to the scenario that or functional magnetic resonance imaging, to study this the patient aims to engage in independently at home). relationship and perform a longitudinal study that would The VR system used here is portable and can be manipulated correlate our results with neuroimaging data as well. to simulate different environments and different navigation Conflicts of Interest None declared. Multimedia Appendix 1 Efficacy ratio and functional impairment score formula. [JPG File, 28KB-Multimedia Appendix 1] References 1. Chaytor N, Schmitter-Edgecombe M. The ecological validity of neuropsychological tests: a review of the literature on everyday cognitive skills. Neuropsychol Rev 2003 Dec;13(4):181-197. [Medline: 15000225] 2. Franzen MD, Wilhelm KL. Conceptual foundations of ecological validity in neuropsychological assessment. In: Sbordone RJ, Long CJ, editors. Ecological Validity of Neuropsychological Testing. Boca Raton, FL: St Lucie Press; 1996. 3. Thompson CL, Henry JD, Withall A, Rendell PG, Brodaty H. A naturalistic study of prospective memory function in MCI and dementia. Br J Clin Psychol 2011 Nov;50(4):425-434. [doi: 10.1111/j.2044-8260.2010.02004.x] [Medline: 22003951] 4. Cesa GL, Manzoni GM, Bacchetta M, Castelnuovo G, Conti S, Gaggioli A, et al. Virtual Reality for Enhancing the Cognitive Behavioral Treatment of Obesity With Binge Eating Disorder: Randomized Controlled Study With One-Year Follow-up. J Med Internet Res 2013;15(6):e113 [FREE Full text] [doi: 10.2196/jmir.2441] [Medline: 23759286] 5. Rizzo AA, Buckwalter JG. Virtual reality and cognitive assessment and rehabilitation: the state of the art. Stud Health Technol Inform 1997;44:123-145. [Medline: 10175336] 6. Hall V, Conboy-Hill S, Taylor D. Using Virtual Reality to Provide Health Care Information to People With Intellectual Disabilities: Acceptability, Usability, and Potential Utility. J Med Internet Res 2011;13(4):e91 [FREE Full text] [doi: 10.2196/jmir.1917] 7. Klinger E, Chemin I, Lebreton S, Marié RM. Virtual action planning in Parkinson's disease: a control study. Cyberpsychol Behav 2006 Jun;9(3):342-347. [doi: 10.1089/cpb.2006.9.342] [Medline: 16780402] 8. Zhang L, Abreu BC, Masel B, Scheibel RS, Christiansen CH, Huddleston N, et al. Virtual reality in the assessment of selected cognitive function after brain injury. Am J Phys Med Rehabil 2001 Aug;80(8):597-604; quiz 605. [Medline: 11475481] 9. Campbell Z, Zakzanis KK, Jovanovski D, Joordens S, Mraz R, Graham SJ. Utilizing virtual reality to improve the ecological validity of clinical neuropsychology: an FMRI case study elucidating the neural basis of planning by comparing the Tower of London with a three-dimensional navigation task. Appl Neuropsychol 2009 Oct;16(4):295-306. [doi: 10.1080/09084280903297891] [Medline: 20183185] 10. Zakzanis KK, Quintin G, Graham SJ, Mraz R. Age and dementia related differences in spatial navigation within an immersive virtual environment. Med Sci Monit 2009 Apr;15(4):CR140-CR150. [Medline: 19333197] 11. Rizzo AA, Buckwalter JG, van der Zaag C. Virtual environment applications in clinical neuropsychology. In: Stanney K, editor. The Handbook of Virtual Environments. New York: Lawrence Erlbaum Associates; 2002:1027-1064. 12. Lo Priore C, Castelnuovo G, Liccione D, Liccione D. Experience with V-STORE: considerations on presence in virtual environments for effective neuropsychological rehabilitation of executive functions. Cyberpsychol Behav 2003 Jun;6(3):281-287. [doi: 10.1089/109493103322011579] [Medline: 12855084] 13. Koenig S, Crucian G, Dalrymple-Alford J, Dünser A. Assessing navigation in real and virtual environments: A validation study. 2010 Presented at: 8th International Conference on Disability, Virtual Reality and Associated Technologies; Aug 31-Sep 2, 2010; Viña del Mar/Valparaíso, Chile p. 7-16. 14. Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. Gerontologist 1969;9(3):179-186. [Medline: 5349366] http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 11 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al 15. Souder E, Chastain JR, Williams RD. Dementia in the new millennium. Medsurg Nurs 2002 Apr;11(2):61-9; quiz 70. [Medline: 11962170] 16. Petersen RC. Mild cognitive impairment as a diagnostic entity. J Intern Med 2004 Sep;256(3):183-194. [doi: 10.1111/j.1365-2796.2004.01388.x] [Medline: 15324362] 17. Winblad B, Palmer K, Kivipelto M, Jelic V, Fratiglioni L, Wahlund LO, et al. Mild cognitive impairment--beyond controversies, towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Intern Med 2004 Sep;256(3):240-246. [doi: 10.1111/j.1365-2796.2004.01380.x] [Medline: 15324367] 18. Petersen RC, Smith GE, Waring SC, Ivnik RJ, Tangalos EG, Kokmen E. Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999 Mar;56(3):303-308. [Medline: 10190820] 19. Nelson AP, O'Connor MG. Mild cognitive impairment: a neuropsychological perspective. CNS Spectr 2008 Jan;13(1):56-64. [Medline: 18204415] 20. Binegar DL, Hynan LS, Lacritz LH, Weiner MF, Cullum CM. Can a direct IADL measure detect deficits in persons with MCI? Curr Alzheimer Res 2009 Feb;6(1):48-51 [FREE Full text] [Medline: 19199874] 21. Amieva H, Letenneur L, Dartigues JF, Rouch-Leroyer I, Sourgen C, D'Alchée-Birée F, et al. Annual rate and predictors of conversion to dementia in subjects presenting mild cognitive impairment criteria defined according to a population-based study. Dement Geriatr Cogn Disord 2004;18(1):87-93. [doi: 10.1159/000077815] [Medline: 15087583] 22. Gauthier S, Reisberg B, Zaudig M, Petersen RC, Ritchie K, Broich K, International Psychogeriatric Association Expert Conference on mild cognitive impairment. Mild cognitive impairment. Lancet 2006 Apr 15;367(9518):1262-1270. [doi: 10.1016/S0140-6736(06)68542-5] [Medline: 16631882] 23. Petersen RC, Roberts RO, Knopman DS, Geda YE, Cha RH, Pankratz VS, et al. Prevalence of mild cognitive impairment is higher in men: The Mayo Clinic Study of Aging. Neurology 2010 Sep 06;75(10):889-897. [doi: 10.1212/WNL.0b013e3181f11d85] 24. Yesavage JA, O'Hara R, Kraemer H, Noda A, Taylor JL, Ferris S, et al. Modeling the prevalence and incidence of Alzheimer's disease and mild cognitive impairment. J Psychiatr Res 2002 Oct;36(5):281-286. [Medline: 12127595] 25. Bangen KJ. Complex activities of daily living vary by mild cognitive impairment subtype. Journal of the International Neuropsychological Society 2010;16:630-639. [doi: 10.1017/S1355617710000330] 26. Goldberg TE. Performance-based measures of everyday function in mild cognitive impairment. American Journal of Psychiatry 2010;167:845-853. [doi: 10.1176/appi.ajp.2010.09050692] 27. Burton CL, Strauss E, Bunce D, Hunter MA, Hultsch DF. Functional abilities in older adults with mild cognitive impairment. Gerontology 2009;55(5):570-581. [doi: 10.1159/000228918] [Medline: 19602873] 28. Bangen KJ, Jak AJ, Schiehser DM, Delano-Wood L, Tuminello E, Han SD, et al. Complex activities of daily living vary by mild cognitive impairment subtype. Journal of the International Neuropsychological Society 2010;16:630-639. [doi: 10.1017/S1355617710000330] 29. Aretouli E, Brandt J. Everyday functioning in mild cognitive impairment and its relationship with executive cognition. Int J Geriatr Psychiatry 2010 Mar;25(3):224-233 [FREE Full text] [doi: 10.1002/gps.2325] [Medline: 19650160] 30. Teng E, Becker BW, Woo E, Cummings JL, Lu PH. Subtle deficits in instrumental activities of daily living in subtypes of mild cognitive impairment. Dement Geriatr Cogn Disord 2010;30(3):189-197 [FREE Full text] [doi: 10.1159/000313540] [Medline: 20798539] 31. Teng E, Becker BW, Woo E, Knopman DS, Cummings JL, Lu PH. Utility of the Functional Activities Questionnaire for Distinguishing Mild Cognitive Impairment From Very Mild Alzheimer Disease. Alzheimer Dis Assoc Disord 2010 Jun 29:348-353 [FREE Full text] [doi: 10.1097/WAD.0b013e3181e2fc84] [Medline: 20592580] 32. Perneczky R, Pohl C, Sorg C, Hartmann J, Komossa K, Alexopoulos P, et al. Complex activities of daily living in mild cognitive impairment: conceptual and diagnostic issues. Age Ageing 2006 May;35(3):240-245 [FREE Full text] [doi: 10.1093/ageing/afj054] [Medline: 16513677] 33. Park B, Jun JK, Park J. Cognitive impairment and depression in the early 60s: which is more problematic in terms of instrumental activities of daily living? Geriatr Gerontol Int 2013 Apr 7. [doi: 10.1111/ggi.12055] [Medline: 23560396] 34. Tarnanas I, Harper B, Majer A, Testerman R. World Intellectual Property Organization: Patentscope. 2009. Virtual Reality Interface System, EPO patent WO2010075481 URL: http://patentscope.wipo.int/search/en/WO2010075481 [accessed 2013-07-30] [WebCite Cache ID 6IUw2U6aI] 35. Perneczky R, Pohl C, Sorg C, Hartmann J, Tosic N, Grimmer T, et al. Impairment of activities of daily living requiring memory or complex reasoning as part of the MCI syndrome. Int J Geriatr Psychiatry 2006 Feb;21(2):158-162. [doi: 10.1002/gps.1444] [Medline: 16416470] 36. Rahimi F, Duval C, Jog M, Bee C, South A, Jog M, et al. Capturing whole-body mobility of patients with Parkinson disease using inertial motion sensors: expected challenges and rewards. Conf Proc IEEE Eng Med Biol Soc 2011;2011:5833-5838. [doi: 10.1109/IEMBS.2011.6091443] [Medline: 22255666] 37. Yamaguchi DA, Foloppe P, Richard E, Allain P. A dual-modal virtual reality kitchen for (re)learning of everyday cooking activities in Alzheimer’s disease. Presence: Teleoperators and Virtual Environments 2012:43-57. [doi: 10.1162/PRESa00080] http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 12 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al 38. Sheikh JI, Yesavage JA. Geriatric Depression Scale (GDS). Recent evidence and development of a shorter version. In: Brink TL, editor. Clinical Gerontology: A Guide to Assessment and Intervention. New York: The Haworth Press; 1986:165-173. 39. Visser PJ, Verhey FR, Boada M, Bullock R, De Deyn PP, Frisoni GB, et al. Development of screening guidelines and clinical criteria for predementia Alzheimer's disease. The DESCRIPA Study. Neuroepidemiology 2008;30(4):254-265 [FREE Full text] [doi: 10.1159/000135644] [Medline: 18515975] 40. Dodge H, Mattek N, Austin D, Hayes T. In-home walking speeds and variability trajectories associated with mild cognitive impairment. Neurology 2012;78(24):1946-1952. [doi: 10.1212/WNL.0b013e318259e1de] 41. Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics 1982 Dec;38(4):963-974. [Medline: 7168798] 42. Tukey JW. The philosophy of multiple comparisons. Statist Sci 1991 Feb;6(1):100-116. [doi: 10.1214/ss/1177011945] 43. Akobeng AK. Understanding diagnostic tests 3: Receiver operating characteristic curves. Acta Paediatr 2007 May;96(5):644-647. [doi: 10.1111/j.1651-2227.2006.00178.x] [Medline: 17376185] 44. Slavin M, Brodaty H, Kochan N, Crawford JD, Trollor J, Draper B, et al. Prevalence and predictors of “subjective cognitive complaints” in the Sydney Memory and Ageing Study. Am J Geriatr Psychiatry 2010 Aug;18(8):701-710. [Medline: 21491631] 45. Simoes M, Castro L, Ribeiro O, Salgado T, Paz C. P03-94 Subjective Memory Complaints and Mild Cognitive Impairment. European Psychiatry 2009 Jan;24:S1093. [doi: 10.1016/S0924-9338(09)71326-4] 46. Leonardi C, Mennecozzi C, Not E, Pianesi R, Zancanaro M. Designing a familiar technology for elderly people. Gerontechnology 2008;7(2):151. [doi: 10.4017/gt.2008.07.02.088.00] 47. Slater M. Place illusion and plausibility can lead to realistic behaviour in immersive virtual environments. Philos Trans R Soc Lond B Biol Sci 2009 Dec 12;364(1535):3549-3557 [FREE Full text] [doi: 10.1098/rstb.2009.0138] [Medline: 19884149] Abbreviations AD: Alzheimer disease ADL: Activities of Daily Living AUC: area under the curve BADL: basic activities of daily living GDS: geriatric depression scale IADL: instrumental activities of daily living MANOVA: multivariate analyses of variance MCI: mild cognitive impairment MMSE: mini-mental state examination RAVLT: Rey Auditory Verbal Learning Test ROC: receiver-operating curve RR: relative risk UI: user interface VE: virtual environment VR: virtual reality VR-DOT: Virtual Reality Day-Out Task VRPN: virtual reality peripheral network Edited by G Eysenbach; submitted 18.06.13; peer-reviewed by W Greenleaf, G Russo; comments to author 08.07.13; revised version received 22.07.13; accepted 22.07.13; published 06.08.13 Please cite as: Tarnanas I, Schlee W, Tsolaki M, Müri R, Mosimann U, Nef T JMIR Serious Games 2013;1(1):e1 URL: http://games.jmir.org/2013/1/e1/ doi: 10.2196/games.2778 PMID: 25658491 ©Ioannis Tarnanas, Winfried Schlee, Magda Tsolaki, René Müri, Urs Mosimann, Tobias Nef. Originally published in JMIR Serious Games (http://games.jmir.org), 06.08.2013. This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Serious Games, is properly cited. The complete http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 13 (page number not for citation purposes) XSL FO RenderX JMIR SERIOUS GAMES Tarnanas et al bibliographic information, a link to the original publication on http://games.jmir.org/, as well as this copyright and license information must be included. http://games.jmir.org/2013/1/e1/ JMIR Serious Games 2013 | vol. 1 | iss. 1 | e1 | p. 14 (page number not for citation purposes) XSL FO RenderX

Journal

JMIR Serious Games – JMIR Publications

Published: Aug 6, 2013

There are no references for this article.