Scientists Develop ‘SleepFM’: AI Can Predict Risk of 130 Diseases from Your Sleep

New Delhi: Sleep does more than just relieve fatigue—it can provide early warnings about future health risks. Research conducted by Stanford University shows that patterns in a person’s sleep can indicate potential risks for diseases such as heart disease, stroke, dementia, kidney failure, and even mortality. The study, published in Nature Medicine, introduces an AI model called SleepFM, which analyzes bodily activity during sleep to predict the likelihood of developing 130 different diseases.

High Accuracy in Predictions
SleepFM has demonstrated a C-index accuracy of over 0.75, meaning it correctly predicts outcomes in approximately 75% of cases. For example:

• Heart attacks: 8 out of 10 predictions were accurate
• Dementia: 8.5 out of 10
• Kidney disease: 8 out of 10
• Stroke: 7.8 out of 10
• Breast cancer: 9 out of 10

The research analyzed sleep data from 65,000 individuals, covering a total of 585,000 hours of sleep, to train and validate the AI model.

What Happens During Sleep?
While we sleep, our brain, heart, lungs, and muscles remain active. Brain waves fluctuate, heart rate varies, breathing patterns change, and the body undergoes restorative processes. During a standard sleep test, these signals are recorded. Scientists explain that early signs of many diseases appear in the body long before symptoms are noticeable, and sleep patterns can capture these subtle indicators.

Sleep’s Role in Dementia and Heart Disease
The study found that disrupted sleep or irregular completion of specific sleep stages is associated with a higher risk of neurological disorders. Similarly, breathing irregularities and low oxygen levels during sleep were linked to heart and lung diseases. SleepFM identifies which aspects of sleep are most relevant for different diseases: brain activity matters more for neurological conditions, while heart rate and respiratory signals are key for cardiovascular diseases.

This breakthrough suggests that analyzing sleep through AI can become a powerful tool for early disease detection, enabling preventive care and timely medical interventions.