Artificial intelligence is opening a new door in early disease detection. A new AI system can predict the risk of developing 130 different diseases years before symptoms appear. The system uses data from just one night of sleep in a laboratory.
The model is called SleepFM. Researchers developed it at Stanford University. It analyzes brain waves, heart activity, and breathing patterns during natural sleep. These signals often change long before illness becomes visible.
James Zou, a data scientist at Stanford and one of the study’s authors, said researchers have long underestimated sleep in medical research. He said that AI allows scientists to use sleep data to protect lives over the long term. Importantly, the predictions appear years before the first signs of disease.
To build the model, researchers analyzed 585,000 hours of sleep data from 65,000 people. They then linked this information to 25 years of medical records. This helped them understand how short-term sleep signals can develop into serious diseases later in life.
The system can identify risks for many major conditions. These include Parkinson’s disease, dementia, heart attacks, prostate cancer, and breast cancer. In many cases, SleepFM performed better than existing advanced medical devices. As a result, researchers believe it could change how doctors approach prevention.
In the past, sleep laboratories mainly focused on diagnosing snoring or breathing disorders. However, this research shows that sleep labs can now predict long-term health risks. This shift could lead to earlier treatment and better outcomes for patients.
The study was published in the journal Nature Medicine. A team of Stanford specialists led the research. Experts say this technology could reduce healthcare costs and improve survival rates through early intervention.
Overall, the findings highlight sleep as a powerful and untapped health signal. With artificial intelligence, one night of sleep may soon help predict and prevent diseases years in advance.
