Assessment of PTSD in military personnel via machine learning based on physiological habituation in a virtual immersive environment

Posttraumatic stress disorder (PTSD) is a complex mental health condition triggered by exposure to traumatic events that leads to physical health problems and socioeconomic impairments. Although the complex symptomatology of PTSD makes diagnosis difficult, early identification and intervention are crucial to mitigate the long-term effects of PTSD and provide appropriate treatment. In this study, we explored the potential for physiological habituation to stressful events to predict PTSD status. We used passive physiological data collected from 21 active-duty United States military personnel and veterans in an immersive virtual environment with high-stress combat-related conditions involving trigger events such as explosions or flashbangs. In our work, we proposed a quantitative measure of habituation to stressful events that can be quantitatively estimated through physiological data such as heart rate, galvanic skin response and eye blinking. Using a Gaussian process classifier, we prove that habituation to stressful events is a predictor of PTSD status, measured via the PTSD Checklist Military version (PCL-M). Our algorithm achieved an accuracy of 80.95% across our cohort. These findings suggest that passively collected physiological data may provide a noninvasive and objective method to identify individuals with PTSD. These physiological markers could improve both the detection and treatment of PTSD.