Abstract
It is widely believed that cybersickness during head-mounted display based virtual reality (HMD-VR) is caused by illusory self-motion (known as vection). However, empirical support is mixed. This study examined the possibility that only unexpected vection is problematic (i.e., to be avoided). It also tested predictions of the Postural Instability Theory (PIT) of motion sickness. The 40 participants in this study played a commercially available VR game (“Aircar”) for up to 14 minutes (or until the first onset of sickness). During gameplay, participants were verbally guided along a pre-selected route (landing their flying car in a pre-determined order at up to 8 different checkpoints in a virtual futuristic city). Their standing postural activity was measured twice pre-exposure, and continuously throughout the game. Directly after leaving VR, we also assessed their experience of vection during the game (its average strength, and whether it was expected/unexpected) as well as their current experiences of sickness (did they feel sick? if so, how severe was it? and what were its symptoms?). Cybersickness was found to be significantly more likely to occur after unexpected vection. A positive exponential scaling relationship was found between vection strength and cybersickness severity, but only when unexpected vection was experienced. Vection type (expected/unexpected) was the most important predictor of cybersickness when our data was subjected to machine learning analyses. Combined with vection strength, it predicted cybersickness incidence with an accuracy of 85%. Partial support for PIT was also found. Pre-exposure Medio-lateral positional variability was found to be significantly different for sick and well participants. Compelling vection is often a necessary and desirable outcome for immersive HMD-VR. Thus, our findings suggest that when it is applied carefully, expected vection can be used to significantly enhance HMD-VR experiences without increasing the risk of cybersickness.