Behavioral approach tests are a common means of assessing specific phobias. In these tests, participants move towards an anxiety-inducing stimulus as close as they are willing to, with the final distance indicating the severity of the anxiety. In this work, we aim to evaluate a virtual reality implementation of the BAT. For this purpose, four different BATs were designed, consisting of two approach methods, both replicated in vivo and in virtuo. Evaluation of these BATs is done by using a standardised presence questionnaire, application-specific questions, as well as the physiological reactions of the participants. The study focuses on the fear of spiders and uses a real and virtual spider as an anxiety-inducing stimulus. Our results show that the developed VR BAT perform within established presence norms, while the different modalities influenced participants' subjective impressions. Furthermore, the standardized structure of the VR environment ensured a consistent experience regarding the anxiety-inducing stimulus. This differs from the observation in the real-world setting, where the behavior of the spider might differ between individuals and also between sessions. This highlights one of the key advantages of virtual reality: complete control over the stimulus and environment. Correlations between presence and physiological signals were found. Particularly, tonic electrodermal activity levels are more stable with increased presence. However, more research into this is required, as the effects of anxiety on the physiological signals make the correlations difficult to interpret. The evaluation has revealed, which design choices are particularly promising for increasing presence in VR applications, and some which should be avoided. Overall, these results indicates that our VR-based implementation is a promising tool for assessing avoidance behavior for individuals with spider phobia.

翻译：暂无翻译