AI Sleep-EEG Model Flags Higher Dementia Risk as Brain Age Rises

Dementia is often diagnosed only after cognitive symptoms emerge, limiting the window for early intervention. Researchers at the University of California, San Francisco, and Beth Israel Deaconess Medical Center developed a machine-learning model that estimates “brain age” from 13 microstructural features in noninvasive overnight electroencephalography, or EEG. The approach could offer a scalable digital marker for identifying elevated risk outside specialist clinics, while capturing subtle sleep physiology that conventional measures such as sleep duration and efficiency may miss.

The study, published in JAMA Network Open on March 19, 2026, pooled five longitudinal cohorts comprising 7,105 dementia-free participants aged 40 to 94. They were followed for 3.5 to 17 years, during which about 1,000 developed dementia. After adjustment for age, sex, education and lifestyle factors, each 10-year increase in the gap between EEG-derived brain age and chronological age was associated with a 39% rise in subsequent dementia risk. The result shows an association, not proof that accelerated brain aging causes dementia.