How a single night’s sleep could help AI predict your risk of over 100 diseases

The team has developed an artificial intelligence model, known as SleepFM, that analyses detailed physiological data recorded during sleep to assess future health risks. The system was trained on nearly 600,000 hours of sleep data collected from about 65,000 people, making it one of the largest studies of its kind. The data included brain activity, heart rate, breathing patterns, eye movements, leg movements and other signals captured overnight using medical-grade sensors.

Researchers say sleep offers a unique window into overall health because it records many of the body’s core functions continuously over several hours. “We record an amazing number of signals when we study sleep,” said Dr Emmanuel Mignot, a senior author of the study published in Nature Medicine. “It’s very data-rich.”

While artificial intelligence has increasingly been used in areas such as cardiology and cancer detection, sleep has received far less attention, despite its importance to physical and mental health. “From an AI perspective, sleep is relatively understudied,” said Dr James Zou, an associate professor at Stanford and co-author of the research. He said SleepFM shows that sleep data can be used to predict a wide range of health outcomes, from cardiovascular conditions to mental health disorders.

The researchers stress that the technology is not meant to replace doctors, but to support earlier detection and prevention. In the future, such models could help identify health risks before symptoms appear, allowing patients and clinicians to act sooner.

Stanford’s team says more research and clinical testing will be needed before sleep-based AI tools are used widely in healthcare.