ARES-SI: adaptive reinforcement-enhanced sampling for suicidal ideation

Suicidal ideation (SI) fluctuates rapidly within individuals, but standard ecological momentary assessment (EMA) uses fixed schedules that miss critical risk windows. We present ARES-SI (Adaptive Reinforcement-Enhanced Sampling for Suicidal Ideation), a reinforcement learning system that personalizes EMA timing to individual risk patterns. ARES-SI combines random forest models to predict active SI probability and non-response likelihood with a Q-learning algorithm to determine optimal assessment intervals (30 min for high-risk states, 3 h for low-risk states, and next day to reduce burden). We trained the system using EMA data from 98 participants across three studies with varied sampling frequencies (4–15 assessments per day; 2359 positive active SI responses) over 28-day protocols. In validation testing ( N = 34; 374 positive active SI responses), the random forest models achieved an AUC of 0.865 for active SI and 0.750 for non-response. ARES-SI demonstrated 65% higher action selection accuracy than random scheduling (0.781 vs. 0.474; macro F1 = 0.650 vs. 0.377), with stronger detection of high-risk windows (30-min recall = 0.810 vs. 0.392) and efficient low-risk scheduling (3-h precision = 0.972 vs. 0.876). These findings provide initial evidence that reinforcement learning can adapt assessment timing to momentary risk.