Real-time anxiety prediction in Virtual Reality exposure therapy
Detection of anxiety patterns in real-time within a virtual reality environment has many uses for medicinal, psychological or entertainment purposes. Virtual reality exposure therapy (VRET) is a therapy method that is quickly rising in popularity, and a built-in way to monitor anxiety levels within VRET applications can contribute to the therapy by providing physiological feedback from the user. This feedback can be used to make meaningful adjustments to context such as increasing exposure levels as user anxiety decreases. For the measurement of physiological signals within Virtual Reality applications, on-body biosensors are generally preferred due to mobility concerns. These biosensors can, however, be susceptible to noise due to movement and it is hard to extract information from a single type of signal. As a countermeasure, this study uses multimodal data and machine learning. The goal of the study is to integrate these signals into a virtual reality experience and accurately assess anxiety levels in real-time by examining patterns across different types of measurements and using a neural network to process information and reduce the effect of noise
Biosensors , Biosignals , VRET , VR , Anxiety , Neural Networks , Machine learning
Mevlevioğlu, D., Murphy, D. and Tabirca, S. (2021) 'Real-time Anxiety Prediction in Virtual Reality Exposure Therapy', IMX ’21: ACM International Conference on Interactive Media Experiences, Adjunct Proceedings, New York City, US, 21-23 June.