Brain electrical waves predict dementia years before it begins

Conventional sleep metrics, such as the duration of deep sleep or overall sleep efficiency, have been found to be weak predictors of dementia, with previous pooled analyses of multiple cohorts of participants failing to reveal a statistically significant association. However, a team led by Yue Lan, an assistant professor of psychiatry at the UCSF School of Medicine, took a different approach by training a machine learning algorithm on 13 microstructural characteristics of brain waves recorded during sleep using an EEG.

The researchers recruited about 7,000 participants aged 40 to 94, none of whom had been diagnosed with dementia at the time of inclusion in the study. Over a follow-up period of 3.5 to 17 years, approximately 1,000 participants developed dementia. The model compared the estimated “brain age” with the participants’ actual ages and found a direct correlation: the greater the difference between the estimated “brain age” and the actual age, the higher the risk of dementia. A ten-year gap increased the likelihood of disease by almost 40%, and if the brain’s age lagged behind the passport age, the risk decreased.

Among the wave patterns that the model used for calculation, there were well-studied markers of cognitive health. Delta waves, which are characteristic of deep sleep, and sleep spindles, short high-frequency bursts of activity associated with memory consolidation, made a significant contribution to the assessment. One of the most interesting results was the observation of the excess (kurtosis) indicator: Sharp, large peaks on the EEG were associated with a reduced risk of dementia.

The association between the “old” brain and dementia persisted even after the researchers took into account the participants’ education, smoking, body mass index, physical activity level, comorbidities, and genetic predisposition, suggesting that the new measure has independent predictive value.

Since sleep EEG recording does not require invasive procedures, researchers see a direct path to practical application through wearable devices capable of recording brain signals at home and assessing its biological age without visiting a clinic. “Brain activity during sleep provides a measurable window into how the brain ages,” Yue Leng noted.

The results raise another question: is it possible to slow down brain aging by improving sleep quality? Early studies have shown that treatment of sleep disorders changes wave patterns, and therefore the estimated age of the brain. Haoci Sun, the first author of the article and an associate professor of neurology at Beth Israel, noted that reducing body mass index and engaging in regular physical activity can reduce the likelihood of sleep apnea and indirectly impact brain health, although there is currently no magic pill for brain rejuvenation.