Investigating the feasibility of neuro-cognitive games for detecting the onset of dementia using a phantom arm compared to touchscreen version

Abstract

In this preliminary study, a virtual reality game was developed to detect the onset of dementia. The game takes place in a 3D virtual kitchen, and the player is tasked to identify displaced objects from memory and to recall the order of displacement. Two different hardware platforms were used to play the game; a touchscreen tablet and a phantom robotic arm. Cognitive abilities such as object recognition, spatial memory and memory retention were assessed. Study participants were 45 seniors, out of which four were diagnosed with Mild Cognitive Impairment (MCI) and 3 with Alzheimer’s disease (AD). Their performances were evaluated against the Montreal Cognitive Assessment (MoCA) test. They performed the experiments both with a phantom arm mimicking humans’ arm and with a touchscreen version. Healthy older adults performed significantly better than MCI participants, who in turn performed better than AD participants. MoCA significantly correlated with the game score on both hardware interfaces. There was also a significant difference between the performance score while using phantom robotic arm compared to that when using the touchscreen, pointing towards a deficit of visuomotor ability in ageing. The scores of performances using touchscreen version of the games was a significant predictor of MoCA, while the scores of using phantom arm was a significant predictor of age. MCI participants performed much worse on order recall tasks compared to object identification tasks, suggesting a more pronounced deficit in memory retention. More MCI and AD participants should be investigated to determine the designed experiments’ sensitivity and specificity in detecting dementia.