Eeg-Guided Evaluation of Vr Attention Training for Reducing Stress Responses Using Machine Learning

This study investigates the feasibility of virtual reality-based attention training as an intervention for stress management. Mental stress remains a pervasive global concern and is far from being fully controlled, which motivates continued efforts to design, validate, and deploy robust tools for its detection and mitigation. Technology-driven approaches-including virtual reality (VR)-are central to these efforts; however, the effectiveness of VR depends critically on the character of the content delivered. Accordingly, we focus on an engagement-oriented content class, attention training (AT), and assess its immediate effect on VR's utility for stress management. We recruited ten undergraduate engineering students from the Arab Academy for Science, Technology, and Maritime Transport (AASTMT) in Alexandria, Egypt¹¹The protocol was approved by the IRB Approval Committee of VRapeutic Inc. (https://myvrapeutic.com/) and the Research Ethics Approval Committee of the Electronics and Communications Engineering Department, College of Engineering and Technology, Arab Academy for Science and Technology, Alexandria, Egypt (E.C.E. ALX/02/004).. Each participant completed a standardized cognitive stressor twice while electroencephalography (EEG) signals were recorded continuously: once after a brief VR based AT session and once without prior training. Using multiple feature representations and several conventional machine-learning (ML) classifiers, we observed a consistent reduction in the number of EEG segments classified as “stress” when the stressor was preceded by VR based AT. These findings constitute preliminary evidence in support of VR based AT as a stress-management strategy and underscore the need for replication with larger and more diverse samples. The resulting EEG dataset is publicly available to support reproducibility and future research²²https://github.com/alkabbany/stress_att_train_dataset.