Brain Waves During Sleep Predict Dementia Risk

A brain age index based on microstructures of sleep electroencephalography (EEG) data predicted dementia risk, a meta-analysis showed.

Across five cohorts and 7,105 participants, each 10-year increase in the EEG-derived index was tied to a 39% higher risk of dementia (HR 1.39, 95% CI 1.21-1.59, P<0.001) after adjusting for covariates, reported Yue Leng, PhD, of the University of California San Francisco, and co-authors.

The relationship remained significant after adjusting for comorbidities and apnea-hypopnea index scores (HR 1.31, P<0.001) and APOE4 gene status (HR 1.22, P=0.03), the researchers wrote in JAMA Network Open. Links were consistent across sex and age groups.

“Sleep gives us a unique window into the brain. This study shows that we can use brain activity during sleep to estimate how the brain is aging,” Leng told MedPage Today.

“Sleep isn’t just rest; it reflects how well the brain is functioning and maintaining itself,” she said. “By turning sleep EEG into a measure of ‘brain age,’ we may be able to detect risk of dementia earlier and more easily.”

The brain age index captures the difference between sleep EEG-based brain age and chronological age, Leng and colleagues noted, with negative values indicating a younger brain age and positive values reflecting an older one. The index incorporated 13 microstructural features of brain waves from EEG recordings of overnight, home-based polysomnography.

Earlier work has linked sleep macrostructure — like sleep efficiency or time spent in certain sleep stages — to dementia risk, but results have been inconsistent.

This study “advances the field by shifting attention from macrostructure to EEG microstructure (i.e., spectral power in specific bands, spindle and slow oscillation characteristics, and waveform properties across sleep stages),” observed Omonigho Bubu, MD, PhD, MPH, of the NYU Grossman School of Medicine in New York City, in an accompanying editorial.

“What makes this work particularly compelling is that brain age index is a general marker of neurophysiologic aging, developed in an independent cohort and then tested, prospectively, against dementia outcomes,” Bubu wrote. “That design enhances its plausibility as a robust marker of accelerated brain aging, rather than an overfitted prediction tool.”

Dementia research is moving toward combining genetic information, blood biomarkers, imaging, and cognitive testing into integrated risk assessments, Bubu pointed out. “Within this landscape, sleep EEG-based brain age index occupies a distinctive niche as it reflects real-time brain physiology, relies on widely used clinical technology, and condenses complex neural information into an interpretable brain age,” he added.

For their individual participant data meta-analysis, Leng and co-authors pooled information from five community-based longitudinal cohorts: the Multi-Ethnic Study of Atherosclerosis (MESA; 1,802 participants), the Atherosclerosis Risk in Communities (ARIC; 1,796 participants) study, the Framingham Heart Study-Offspring Study (FHS-OS; 617 participants), the Osteoporotic Fractures in Men Study (MrOS; 2,639 participants), and the Study of Osteoporotic Fractures (SOF; 251 participants).

Participants’ average ages at the time of the sleep study ranged from 59.5 to 82.7 years, and more than 90% were cognitively normal. The primary outcome for the pooled analysis was incident dementia, with death treated as a competing risk.

Over follow-up, 1,088 people developed dementia, with a median time to dementia ranging from 3.6 to 16.9 years.

The findings highlight the need to evaluate the predictive value of the brain age index as a digital marker for early dementia detection in community settings, the researchers said.

The five cohorts included in the study differed in population characteristics, data collection methods, dementia ascertainment, and follow-up times, which may have introduced heterogeneity and potential bias, they acknowledged. Only death was used as a competing risk, but other life events may affect follow-up or dementia ascertainment, they added. The analysis is observational and a causal relationship between brain age index and dementia can’t be inferred.

“Moreover, as a composite measure, brain age index itself is not a plausible therapeutic target,” Leng and colleagues wrote. “Rather, brain age index should be viewed as a prognostic marker for future dementia risk.”