JMIR Publications

ABSTRACT

Background:

The gamification of digital health applications for older adults (e.g. for rehabilitation) is a growing trend, however many older adults are not familiar with digital games. This lack of experience could create stress that might impede participant’s motivations to adopt these assisted care technologies that use gamification strategies.

Objective:

This crossover longitudinal multifactorial study uses the Affective Game Planning for Health Applications (AGPHA) framework to examine the interactions between game difficulty, appraisal, cognitive ability and physiological and cognitive responses that indicate game stress.

Methods:

18 (71±4.5 years, 12 women) volunteers completed a 3-session study to evaluate different genres of games in increasing order of difficulty (S1-BrainGame, S2-CarRace, S3-Exergame). Each session included identical sequence of Activities (t1-Baseline, t2-Picture encode, t3-Play, t4-Stroop test, t5-Play, t6-picture_recall), and repeated sampling of salivary cortisol, and time-tagged ambulatory data from a wrist-worn device. Generalized estimating equations (GEEs) were used to investigate the effect of Session x Activity, or Session x Activity x Cognitive Ability on physiology and cognitive performance. Scores derived from the Montreal Cognitive Assessment test were used to define cognitive ability (MoCA-high, N=11/18, MoCA>27). Kruskal-Wallis tests were used to test Session or Session x Group effects on the scores of the post-game appraisal questionnaire.

Results:

Session x Activity effects were significant on all ambulatory measures (χ2(df=10)'s >20, p's<.001) other than cortisol (p=.37). Compared to S1 and S2, S3 was associated with approximately 10 bpm higher heart rate (p<.001), and approximately 5 muS higher EDA (p<.001), which was independent of movement caused by the exergame. Compared to S1, we measured a significant drop in the rate of immediate recall (12%), and rate of delayed recall (22%) in S3. Low-MoCA did not differ from the High-MoCA group in general characteristics (age, general self-efficacy, perceived stress), but was more likely to agree with the statements such as "digital games are too hard to learn". In addition, the Low-MoCA group was more likely to dislike the gaming experience, and find it useless, uninteresting and visually more intense (Chi-squared 's> 4, p's<.04). Group differences in ambulatory signals did not reach statistical significance, however the rate of cortisol decline with respect to the baseline was significantly larger in the Low-MoCA group.

Conclusions:

Our results show that to play digital games was not experienced as stressful. Comparatively, the neurophysiological effects of exergame were more pronounced, in the low-MoCA group, suggesting the greater potential of this genre of games for cognitive and physical stimulation by gamified interventions, however the need for enjoyment of this type of challenging game must be addressed.