US EditionStanford AI Predicts 130 Diseases from One Night of Sleep
SleepFM, trained on 585,000 hours of data from 35,000 patients, predicts risk for 130 diseases including cancers and neurological disorders with strong accuracy.
- On January 6, 2026, Stanford Medicine researchers published a Nature Medicine paper introducing SleepFM, an AI that estimates risk for 130 diseases from one night of sleep.
- Polysomnography captures brain, heart, respiratory, eye and muscle signals overnight, providing detailed physiological data that much went unused until recent AI advances enabled full analysis.
- Trained on nearly 600,000 hours of polysomnography from about 65,000 participants, SleepFM used five-second segments and leave-one-out contrastive learning to harmonize data modalities.
- Strong performance for major condition groups indicates SleepFM spots subtle mismatches among brain, heart and breathing signals, achieving a C-index higher than 0.8 for cancers, pregnancy complications, circulatory diseases and mental disorders.
- Using decades of clinic records, the researchers linked Stanford Sleep Medicine Center records with electronic health records with up to 25 years follow‑up, but no single signal suffices so researchers are working on interpretation and wearable integration.
33 Articles
33 Articles
Researchers Develop AI Model That Predicts Disease Risk From Sleep Data
The AI system, described in a paper in the journal Nature Medicine, was initially tested on standard tasks involving sleep analysis, such as tracking different stages of sleep or diagnosing severity...
AI Revolutionizes Health Prognosis Through Sleep Data Analysis
Researchers have created an AI model, 'SleepFM', that predicts health risks across 130 conditions by analyzing sleep data. Developed with six lakh hours of data from 65,000 participants, SleepFM excels in forecasting issues like cancer and mental disorders, offering new potential in early disease detection.
Stanford researchers develop SleepFM, an AI that analyzes sleep data to predict risks of serious diseases such as dementia and heart attack.
Researchers say AI can accurately predict 130 diseases from one night of sleep
Researchers at Stanford University have developed an artificial intelligence system that can forecast a persons risk of developing certain diseases and even death by analyzing data from just one night of sleep.The research team examined sleep recordings from more than 65,000 people for the study, introducing their SleepFM AI tool, which they said can accurately predict risks for 130 diseases years before diagnosis.RELATED STORY | National Sleep …
Coverage Details
Bias Distribution
- 86% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
