Exploring Gaze Dynamics in Virtual Reality through Multiscale Entropy Analysis

Abstract

This study employs Multiscale Entropy (MSE) to analyze 5020 binocular eye movement recordings from 407 college-aged participants, as part of the GazeBaseVR dataset, across various virtual reality (VR) tasks to understand the complexity of user interactions. By evaluating the vertical and horizontal components of eye movements across tasks such as vergence, smooth pursuit, video viewing, reading, and random saccade, collected at 250 Hz using an ET-enabled VR headset, this research provides insights into the predictability and complexity of gaze patterns. Participants were recorded up to six times over a 26-month period, offering a longitudinal perspective on eye movement behavior in VR. MSE’s application in this context aims to offer a deeper understanding of user behavior in VR, highlighting potential avenues for interface optimization and user experience enhancement. The results suggest that MSE can be a valuable tool in creating more intuitive and immersive VR environments by adapting to users’ gaze behaviors. This paper discusses the implications of these findings for the future of VR technology development, emphasizing the need for intuitive design and the potential for MSE to contribute to more personalized and comfortable VR experiences.