Implementation of a virtual reality environment to elicit acute mental stress.

Acute stress is a condition that arises in response to situations that are difficult to manage, manifesting in both physiological and psychological changes. Furthermore, the occurrence of stressors for a long period can result in a chronic stress condition, leading to serious health issues. To study stress in controlled conditions, research employs cognitive tests to induce acute mental stress and subsequently investigate people response and methods for stress mitigation. However, the results of cognitive tasks performed in a traditional way can be distorted by external events affecting participants, raising doubts about whether the observed effects can be attributed exclusively to the test itself. Recent studies have started implementing stress-inducing protocols in virtual reality (VR). Notably, VR-based approaches have shown better results than traditional cognitive tasks, making VR a promising tool for stress research. In this thesis, a game in VR was developed to elicit a stressful response. Notably, it does not require specific technical or social skills to be performed, making it suitable for a huge variety of people. The game is designed to stimulate universally present traits, such as reactivity to intense or unexpected stimuli and competitiveness (either against others or oneself), with the aim of generating a more uniform response. One of the purposes of this work is to assess whether stress levels measured during the VR game are as good as a mental arithmetic task (MAT) or higher. The experimental protocol, conducted on 20 subjects, included relaxation phases with calming videos and music to establish a baseline, the VR game and the MAT as a comparison. Stress during the protocol was quantified using self-report questionnaires and features of biological signals like the electrocardiogram (ECG) and the electrodermal activity (EDA). Results show significant differences in heart rate, breath rate and in the mean, median, maximum value and cumulative area of the EDA phasic component between the relaxation phase and both stress-inducing tasks, suggesting that the proposed game elicit a consistent stressful response. These findings are supported by self-assessment questionnaires, which show no significant difference in perceived stress between the two stress-inducing tasks, but highlight a clear difference between the relaxation phase and both tasks. Questionnaire data also indicate that 95% of participants felt fully immersed in VR and disconnected from the external environment, confirming the effectiveness of the experiment. Future work will focus on techniques to reduce stress. Although this aspect was not considered in the current study, a preliminary step was addressed by developing a regression model, trained on the MAT and tested on the VR game, to predict the stress state of the user. Notably, the predictions can drive a stimulation with either auditory or mechanical stimuli to mitigate the effect of stress.