Studying the Influence of Single Social Interactions on Approach and Avoidance Behavior – A Multimodal Investigation in Immersive Virtual Reality

Abstract

Adaptive social approach and avoidance behavior is of utmost importance for social functioning and an imbalance in these behavioral tendencies may constitute a risk factor for the etiology and maintenance of mental illness. To investigate these processes in a naturalistic environment with high ecological validity and experimental control, we implemented an immersive virtual environment that combines a social conditioning procedure with a social approach-avoidance test. In this environment, we conducted two experiments (both with N = 48 female participants; data collected in summer 2023 and winter 2023/24) to examine whether participants differing in trait social anxiety develop adaptive approach and avoidance tendencies towards differently conditioned virtual agents on a behavioral, subjective, and autonomic level. Overall, we observed significant effects of the social conditioning procedure on participants’ exploration behavior. After initial hypervigilance, participants exhibited avoidance of social threats as indicated by a higher interpersonal distance and decreased visual attention towards the negatively conditioned virtual agent. We found no association between hypervigilance and trait social anxiety, but observed higher fear ratings and enhanced avoidance of social threats in participants with elevated anxiety levels. This study demonstrates the potential of immersive virtual environments for examining social learning processes under conditions resembling real-life social encounters.

The influence of social anxiety and visual exploration on threat responding and generalization

“…Likewise, the current study offers a rather artificial (yet highly controlled) environment for the investigation of fear generalization processes. Hence, additional research with high ecological validity is needed to further corroborate the current findings (Gado & Gamer, 2024;Großekathöfer et al, 2021;Horn et al, 2022;Rösler et al, 2021;Rubo et al, 2020). Regardless of this limitation, the validity of a highly controlled and carefully edited set of facial photographs has been demonstrated: We observed no significant influence of the diagnostic region on threat ratings, suggesting a successful manipulation of our stimulus set resulting in similar difficulty for different pairs of faces.…”