Utilization of Neurophysiological Data to Classify Player Immersion to Distract from Pain

摘要

Painful experiences during clinical procedures can have detrimental effects on the physical and mental health of a patient. Current pain reduction methods can be effective in reducing pain, however these methods are not without fault. Active distraction via computer games have been proven to effectively reduce the experience of pain. However, the potential of this distraction to effectively alleviate pain is dependent on players' engagement with the game, which is determined by the difficulty of the game and the skill of the player. This paper aims to model and classify immersion through increasingly difficult levels of game play, in the presence of pain, using functional Near Infrared Spectroscopy (fNIRS) and heart rate data. Twenty people participated in a study wherein fNIRS data (4 channels located at the prefronlal cortex, four channels located at the somatosensory cortex) and heart rate data were collected whilst participants were subjected to experimental pain, via the Cold Pressor Test (CPT). Participants played a computer game at varying difficultly levels as a distraction. Data were then pre-processed using an Acceleration Based Movement Artefact Reduction Algorithm (AMARA) and Correlation Based Signal Improvement (CBSI). Classification was subsequently undertaken using Linear Discriminant Analysis (LDA), Support Vector Machine (SVM) and Recursive Partitioning (rPart). The results demonstrate a maximum accuracy of 99.2% for the binary detection of immersion in the presence of pain.