Use of physiological signals to predict cybersickness

摘要

Cybersickness is a common and unpleasant side effect of virtual reality immersion. We measured physiological changes that were experienced by seated subjects who interacted with a virtual environment (VE) first while viewing a display monitor and second while using a head-mounted display (HMD). Comparing results for these two conditions let us identify physiological consequences of HMD use. In both viewing conditions, subjects rated the severity of their symptoms verbally and completed a post-immersion cybersickness assessment questionnaire. In the HMD viewing condition but not in the display monitor condition, verbal reports of cybersickness severity increased significantly relative to baseline. Half of the subjects chose to exit the VE after six minutes of HMD use and reported feeling some nausea at that time. We found that changes in stomach activity, blinking, and breathing can be used to estimate post-immersion symptom scores, with R2 values reaching as high as 0.75. These results suggest that HMD use by seated subjects is strongly correlated with the development of cybersickness. Finally, a linear discriminant analysis shows that physiological measures alone can be used to classify subject data as belonging to the HMD or monitor viewing condition with an accuracy of 78%.