KAIST Develops AI That Alerts to Stroke Premonitory Symptoms

An AI neural model that analyzes stroke premonitory symptoms and alerts in advance has been developed. KAIST researchers, in collaboration with experts from Seoul National University and Korea University Hospital, created an AI framework to identify early risk stages and evaluate imminent diagnostic conditions for cerebrovascular diseases. The model utilized 14-day interval data from 1,224 elderly individuals over 16 weeks, analyzing activity patterns, sleep characteristics, circadian rhythms, indoor environments, and personal health data to detect diagnostic risk stages and imminent diagnostic periods.