Predicting Presence using Environment-Activated Motion in Immersive Virtual Reality

Immersive virtual reality (IVR) has shown promise in being an effective method for a variety of interventions across a multitude of disciplines. Indeed, IVR allows researchers to place participants in realistic three-dimensional environments with a level of control that may not have been feasible in physical environments. Additionally, IVR elicits a higher sense of presence than computer-based interventions. Evidence suggests that presence, defined as the “sense of being there”, is a key factor in the success of these interventions. The relationship between presence and hardware and software factors have been extensively investigated. Behavioral factors have comparatively received little attention. While these system factors are important to investigate, behavioral differences across users may preclude them from engaging in meaningful IVR interventions, regardless of the hardware and software solutions applied. Therefore, this study investigated how an individual’s interaction with the virtual experience may influence their sense of presence. More specifically, we wanted to probe the relationship between environment-activated motion (EAM) and presence in IVR. EVM can be described as any motion within the experience initiated by the user or the environment. This is activated by user-activated motion (i.e., a user looks around) or environment-activated motion (i.e., a virtual dog moves within view). 80 younger adults (22 male, 1 non-binary) completed a 10-minute IVR experience in Vesper Peak using the HTC Vive IVR system. The experience consisted of playing with a virtual robot dog (ex., petting and playing fetch) and teleporting to different areas of the mountain. The session was screen captured using OBS. EVM data were collected by analyzing the videos via OpenCV’s implementation of Dense Optical Flow (DOF). Presence was collected via the IGroup Presence Questionnaire (IPQ, Cronbach’s ⍺ = 0.87). Five data were removed due to corruption, resulting in a final sample size of 75. A regression was conducted with EVM as the predictor and presence as the predictand. The overall model was non-significant (R2 = 0.04, RSE = 5.87, F(1,73) = 12.61, p = 0.054). These results suggest that environment-activated motion may not play a significant role in presence during IVR experiences. Future IVR-based interventions may not have to control for environment-activated motion and can focus more on user interaction and environment design. Further research is needed to look into how environment-activated motion and user-activated motion separately may affect presence in different ways.