A Gamified Approach to Cognitive Assessment with Machine Learning Based Predictions

Abstract

Cognitive Assessment is an important method for identifying cognitive impairment in individuals, and diagnosing diseases such as Alzheimer’s disease. However, it is usually performed using paper-based assessment, which can be frustrating and unengaging for patients. Low engagement can lead to inaccuracies and anomalous results. This paper aims to address this issue by taking a gamified approach to cognitive assessment. Using a physical prototype of a wack-a-mole inspired game, we accurately predicted cognitive ability of players using an SVR Machine Learning model. This model used inputs from participants playing the game including reaction times, in-game scores and heart rate, achieving an R-squared of 0.689 (3sf).