Introduction & Purpose
As urban expansion continues, restricted access to natural environments has become a significant driver of stress among city populations. Immersive 360° panoramic videos offer a scalable and innovative medium for simulating natural settings, providing a potential tool for stress mitigation within controlled environments. This study investigated whether brief exposure to different virtual urban landscapes facilitates stress recovery and impacts subsequent risk-taking behavior.
Method
Using a within-subjects design, sixty-six participants underwent stress induction via the threat-of-shock paradigm (ToS). During each ToS block, a red square (stimulus duration = 8 s) appeared on the screen (23-inch LG monitor) six times, with interstimulus intervals ranging from 8 to 20 s. An electrical shock followed the stimulus in 50% of the trials. Participants then viewed 360° monoscopic videos featuring an urban park, a high-traffic highway, or a minimal-stimulation control condition. During the VR exposure, we recorded physiological markers, including heart rate and skin conductance responses. Additionally, self-reported stress levels were collected, and risk-taking was assessed using the Balloon Analogue Risk Task (BART).
Results
The results demonstrated that self-reported stress was significantly lower following the urban park exposure compared to both the highway and control conditions. Physiologically, heart rate and skin conductance responses were significantly reduced during the park condition relative to the highway. However, the virtual environments exerted no significant effect on heart rate variability or risk-taking behavior.
Conclusions
These findings suggest that while brief immersion in virtual green spaces effectively supports physiological and psychological recovery following acute stress, these benefits may not extend to shifting decision-making under risk. The effects were more robust for subjective stress than for physiological measures, highlighting potential differences in sensitivity across outcome domains.