Wearables may flag PTSD flare-ups before symptoms escalate
Detecting short-term post-traumatic stress disorder symptom increases among veterans: machine-learning analysis integrating wearable sensor and daily self-report data.
AI Summary
This study examined whether wearable activity trackers, combined with brief daily questionnaires, could detect short-term increases in post-traumatic stress disorder (PTSD) symptoms among recently discharged veterans. Seventy-four veterans provided data over 87 days, and each personβs wearable readings were compared with their own first 14 days to identify meaningful changes from their usual patterns.
A machine-learning model performed best when it combined 17 features, including mood and perceived stress reports alongside changes in sleep continuity, activity variability, and autonomic regulation. The strongest model showed good discrimination, with a precision-recall AUC of 0.86 and a receiver-operating-characteristic AUC of 0.89. The findings suggest that personalized wearable data may help flag periods of increased PTSD symptoms earlier than infrequent screening. However, the participants had elevated PTSD symptoms and problematic cannabis use, so the results do not show that cannabis caused or relieved PTSD symptoms, nor that the model applies broadly to all cannabis users or people with PTSD.
π‘ Key Findings
π
Original Abstract
Related Research
Similar Studies
More Mental Health research papers you might find interesting.
Brain blood-flow reductions in acute anorexia follow a cortical hierarchy
CBD showed limited adaptive-functioning differences versus placebo in boysβ¦
Resilience linked to depression outcomes and THC positivity
Explore More Research
Stay informed about the latest cannabis science.