Simulated virtual reality experiences for predicting early treatment response in panic disorder

Area of Science:

• Psychiatry
• Psychology
• Medical Technology

Background:

• Panic disorder (PD) is a significant anxiety condition.
• Early treatment response in PD predicts better long-term outcomes.

Purpose of the Study:

• To evaluate a novel virtual reality tool, the Virtual Reality Assessment of Panic Disorder (VRA-PD), for predicting early treatment response in PD patients.
• To compare the predictive power of VR-based assessments versus traditional clinical measures.

Main Methods:

• 52 participants (25 with PD, 27 healthy controls) were assessed over 6 months.
• Measures included self-reported anxiety, heart rate variability during VR scenarios, and clinical questionnaires.
• A machine-learning model (CatBoost) classified participants into early responder, delayed responder, and healthy control groups.

Main Results:

• A combined model using VR and clinical data achieved 85% accuracy and a 0.71 F1-score.
• Models using only clinical data (77% accuracy, 0.56 F1-score) or VR data (75% accuracy, 0.64 F1-score) were less accurate.
• Key predictors included in-VR anxiety levels, heart rate variability, and scores from the Panic Disorder Severity Scale and Anxiety Sensitivity Index.

Conclusions:

• Virtual reality-based assessments are valuable for predicting early treatment outcomes in PD.
• VR assessments offer ecologically valid and individualized measures.
• VR tools can enhance clinical decision-making and support personalized mental healthcare for anxiety disorders.