Validation of Cognitive and Behavioral Biomarkers of Parkinsons Disease through Human Performance Data in VR Enviorments

Overview

Abstract- Currently, 11.7 million people worldwide are affected with Parkinson’s disease, a neurodegenerative movement disorder of the central nervous system, in which dopamine-producing cells degenerate. This study aims to explore the present clinical issues associated with Parkinson's, which leads to the development of a new methodology using VR-based AI eye-tracking technology. This was achieved by testing and assessing 56 participants, including healthy controls and individuals at different stages of Parkinson's disease. This aims to fill the existing gap in the availability of VR-based eye-tracking data, particularly on the behavioral eye movements associated with Parkinson's disease. This study was done using a three-phase framework, including six VR-based eye-tracking simulations, which focused on the behavioral eye movements associated with the different stages of Parkinson's disease. Phase 1 focused on single-task foundational eye movements, including Pro-Saccades, Anti-Saccades, and Smooth Pursuit. Phase 2 focused on dual-task assessments, including auditory and visual processing, and Phase 3 focused on Parkinson’s Disease intervention. The analysis of the data revealed 14 statistically significant biomarkers, including the fatigue slope. This study provides the first evidence of human data supporting the theory of the primary early indicator of Parkinson's disease in prodromal stages, which is associated with the theory of active fatigue. Other critical markers, such as Motor Speed, Coordination Error, and Smooth Pursuit Velocity Variability, were also found to be critical for Parkinson’s detection, validating the results from previous studies. The unsupervised and supervised machine learning (ML) models obtained from the analysis achieved an accuracy of 91% and 92%, respectively. Overall, these tasks provide an in-depth, VR-based method for studying Parkinson’s related eye movement through interactive testing, making VR more accessible, accurate, and providing further help in the medical field for early detection.