Ferina 2025 — Approaches for Predicting Challenging Behaviors in ASD: A Narrative Review
Citation: Ferina, J., Dando, E., Anderson, C., et al. (2025). Journal of Personalized Medicine, 15(10), 453.
DOI: 10.3390/jpm15100453
Type: Narrative review
Key Contribution
Comprehensive review of all approaches for predicting challenging behaviors in ASD. Covers wearables, non-wearable sensors, and manually recorded data. Includes short-term (seconds) to long-term (next-day) prediction horizons.
Covered Approaches
- Wearable biosensors (EDA, HR, ACC) — Goodwin, Imbiriba et al.
- Video-based emotion detection (SSBD, ARRBD datasets)
- Sleep data → next-day behavior prediction
- Environmental data (GI, pollen, weather, moon phase)
Key Gap Identified
Most approaches are short-lead (seconds to minutes). Longer-horizon prediction (next-day) is emerging but under-explored. Few studies on large, diverse populations.
Linked To
- Paper: romani-2026-review
- All Tier 1 datasets