Researchers at Stanford Medicine have developed an advanced AI model called SleepFM that can predict the risk of more than 130 diseases using data from just one night of sleep.

The model analyses detailed sleep signals collected during polysomnography tests, including brain waves, heart rate, breathing patterns, eye movements, and muscle activity. SleepFM was trained on nearly 600,000 hours of sleep data from around 65,000 people, allowing it to detect health patterns that are often missed in routine medical checks.

During testing, the AI showed strong accuracy in predicting risks for serious conditions such as heart disease, cancer, stroke, dementia, diabetes, and even pregnancy-related complications - sometimes years before symptoms appear.

According to researchers, sleep acts like a continuous health monitor, quietly reflecting what’s happening inside the body. While SleepFM is still in the research stage and not yet used in hospitals, experts believe it could soon support early diagnosis, preventive care, and personalised health monitoring through sleep analysis.