Enhancing biofeedback-driven self-guided virtual reality exposure therapy through arousal detection from multimodal...

Area of Science:

• Psychology
• Computer Science
• Biomedical Engineering

Background:

• Public-speaking anxiety (PSA) is a common social anxiety impacting many individuals.
• Virtual reality exposure therapy (VRET) offers a safe, controlled environment for anxiety treatment.
• Detecting physiological arousal in real-time is crucial for effective VRET but remains a challenge.

Purpose of the Study:

• To explore machine learning (ML) models for predicting arousal states using physiological data.
• To develop a pipeline for effective ML model and parameter selection in VRET.
• To implement a biofeedback framework for VRET to aid in anxiety management.

Main Methods:

• Utilized publicly available datasets including electroencephalogram (EEG) and heart rate variability (HRV).
• Investigated various ML models for arousal state prediction.
• Developed and tested a pipeline for ML model selection and parameter optimization.
• Implemented a biofeedback system providing heart rate and brain laterality index feedback.

Main Results:

• Successfully predicted arousal states using ML models with EEG and HRV data.
• Demonstrated the effectiveness of the proposed pipeline for ML model selection.
• Implemented a functional biofeedback framework for VRET.

Conclusions:

• ML models can effectively detect anxiety-induced arousal from physiological signals.
• The proposed pipeline facilitates optimal ML model selection for arousal detection in VRET.
• Biofeedback integrated with VRET shows promise for psychological intervention and anxiety reduction.