Autonomic Nervous System Factors Underlying Anxiety in Virtual Environments: A Regression Model for Cybersickness

The ability to predict whether people will experience anxiety is important for recruitment and selection in highly-stressful professions. Using a virtual reality environment (VRE) can provide a tool to predict whether a person will experience anxiety. This paper reports several regression models which suggest observed and self-reported measures of anxiety during and after immersion in a VRE can be used to predict an individual´s anxiety response to a simulated stressful environment. We found that respiration was a poor predictor of anxiety, but that cardiac activity accounted for around 39% of variance in self-reported anxiety responses using a four point scale. In contrast, responses from the simulator sickness questionnaire (SSQ) accounted for 98% of variance in anxiety responses. However, only four out of eighteen measures in the SSQ made a significant contribution to the model. The implication for predicting an individual´s anxiety responses using self-report or physiological measures is discussed.