AI Can Predict 130 Health Issues From One Night of Sleep

One of the greatest opportunities for artificial intelligence (AI) machine learning is in the field of health and disease diagnostics. A new breakthrough study demonstrates how AI can predict a person’s risk of developing over a hundred serious medical conditions from data collected noninvasively from just a single night of sleep.

“This study underscores the potential of sleep-based foundation models for risk stratification and longitudinal health monitoring,” wrote Stanford University co-corresponding authors James Zou and Emmanuel Mignot in collaboration with co-authors Rahul Thapa, Magnus Ruud Kjaer, Bryan He, Ian Covert, Hyatt Moore IV, Umaer Hanif, Gauri Ganjoo, M. Brandon Westover, Poul Jennum, and Andreas Brink-Kjaer.

Why sleep?

Sleep is an essential part of maintaining not only physical health, but also psychological well-being, as it affects emotional regulation, cognition, resilience, focus, and memory. An estimated 50 percent of insomnia cases are linked to psychological stress, anxiety, or depression, and obsessive-compulsive disorder (OCD) is often linked with poor sleep, according to the National Alliance on Mental Illness (NAMI).

There are more than 80 types of sleep disorders, according to the Cleveland Clinic. The most common ones include chronic insomnia, obstructive sleep apnea, restless legs syndrome, REM sleep behavior disorder, narcolepsy, delayed sleep phase syndrome, and shift work sleep disorder.

Who’s not sleeping well?

There are a lot of people who are not sleeping well in the U.S. and globally. By 2034, the sleep disorder market is expected to reach USD 72 billion, growing at a compound annual growth rate of 10 percent from 2025 to 2034, according to Global Market Insights. The American Brain Foundation estimates that 50-70 million people in the U.S. alone have sleep or wakefulness disorders. Roughly 1 in 3 adult Americans reported not getting enough rest or sleep daily, per the U.S. Centers for Disease Control and Prevention. Globally, there are nearly one billion adults aged 30-69 years old with sleep apnea, according to a 2019 study by Benjafield et al. published in The Lancet Respiratory Medicine. Sleep apnea is just one sleep disorder.

Overnight sleep: An AI data goldmine

The researchers created a multimodal AI model called SleepFM that was trained on polysomnography (PSG) data, which is data captured noninvasively during an overnight sleep study. The Greek prefix “poly” means “many,” and polysomnography records many physiological signals.

During PSG, brain waves are noninvasively and painlessly recorded via electroencephalogram (EEG), in addition to blood oxygen levels via pulse oximetry, eye movements via electro-oculogram, heart rate via electrocardiogram, breathing, and leg movements via electromyogram. Polysomnography is the gold standard for diagnosing sleep behaviors such as sleepwalking, sleep apnea, other sleep-associated breathing disorders, long-lasting insomnia, periodic limb movement disorder, narcolepsy, and REM sleep behavior disorder.

When it comes to AI model performance, having massive datasets with high-quality training data may improve the overall accuracy. In this study, the AI model was trained on polysomnography data from roughly 65,000 participants across multiple cohorts on more than 585,000 hours of curated recordings. The cohorts include polysomnography data from the Stanford Sleep Clinic (SSC), Outcomes of Sleep Disorders in Older Men (MrOS), the Multi-Ethnic Study of Atherosclerosis (MESA), and BioSerenity. Data from the Sleep Heart Health Study (SHHS) were used to fine-tune the algorithm.

“Our model uses 5 to 25 times more data than previously trained supervised sleep or biosignal models,” the researchers wrote.

To train SleepFM, the team used a self-supervised learning algorithm that did not require labeled data. The researchers then tested the AI on over a thousand disease phenotypes. The AI model performed particularly well in predicting Alzheimer’s disease and Parkinson’s disease, both of which are neurodegenerative diseases. According to the scientists, their AI model accurately predicted 130 health conditions that include dementia, stroke, heart failure, chronic kidney disease, myocardial infarction, atrial fibrillation, and all-cause mortality.

“This work shows that foundation models can learn the language of sleep from multimodal sleep recordings, enabling scalable, label-efficient analysis and disease prediction,” the researchers concluded.

Copyright © 2026 Cami Rosso. All rights reserved.