Recognition of VR Motion Sickness Level Based on EEG and Functional Brain Network

Virtual reality motion sickness (VRMS) is a common problem for VR players. To study the causes and cures of VRMS, the recognition of VRMS level is a precondition. So, we aim to study how VRMS modulates EEG in functional brain networks and extract EEG features to measure VRMS. In this study, 23 subjects are recruited and their EEG are collected when they are in VMRS induced by “VRQ test”. We introduce a functional brain network approach based on empirical mode decompose and phase locking value to analysis the EEG, and recognize the VRMS level using graph-theoretic indexes and machine learning. The results suggest that the high frequency mode (about 30–40 Hz) of EEG has more significant features between low and high VRMS levels, and the classification accuracy in the task and the post-task rest states are up to 96.3% and 97.6% respectively. This work firstly reveals the effects of VRMS on the functional brain networks and provides a method to assess and guide the comfortable VR products design.