Sleep EEG & Dementia Risk: Machine Learning Predicts ‘Brain Age’
A machine-learning analysis of brain waves recorded during sleep can identify individuals at high risk of developing dementia, according to a study published March 19 in JAMA Network Open. Researchers at UC San Francisco and Beth Israel Deaconess Medical Center in Boston found that a disparity between a person’s “brain age,” as estimated from sleep signals, and their chronological age is a significant predictor of dementia risk.
The study, which followed approximately 7,000 participants aged 40 to 94 over a period of 3.5 to 17 years, revealed that for every 10-year increase in brain age compared to actual age, the risk of developing dementia rose by nearly 40 percent. Conversely, individuals whose brain age was younger than their chronological age exhibited a lower risk. Approximately 1,000 participants developed dementia during the study period.
Researchers utilized a machine-learning model integrating 13 microstructural features of brain waves obtained through electroencephalography (EEG) recordings. The analysis revealed that conventional sleep metrics, such as total sleep time and sleep stage duration, were less informative than the fine-scale patterns within brain waves themselves. Previous analyses of the same participant cohorts had not established significant links between dementia risk and these traditional sleep measures, according to the study.
“Broad sleep metrics don’t fully capture the complex multidimensional nature of sleep physiology,” said Yue Leng, MBBS, PhD, associate professor of psychiatry at the UCSF School of Medicine and senior author of the study. The research suggests that analyzing these subtle brain-wave patterns during sleep offers a more nuanced understanding of brain health and aging.
Specific EEG patterns linked to brain age and cognitive health include delta waves, associated with deep sleep, and sleep spindles, which play a role in memory consolidation. Notably, the presence of kurtosis – sudden, large spikes observed on EEG – was associated with a lower risk of dementia.
The correlation between an “older” brain age and increased dementia risk persisted even after researchers accounted for factors such as education level, smoking habits, body mass index, physical activity, and pre-existing health conditions, as well as genetic predispositions. This suggests that brain age, as determined by sleep EEG, represents an independent risk factor for dementia.
Researchers believe the non-invasive nature of EEG recordings opens the possibility of using brain age as a screening tool for dementia risk in non-clinical settings, potentially through the use of wearable technologies. “Brain age is calculated from sleep brain waves,” Leng explained. “We know that brain activity during sleep provides a measurable window into how well the brain is aging.”
The study also raises the possibility of influencing brain aging through improved sleep health. Leng noted that previous research has demonstrated that treating sleep disorders can alter sleep-related brain-wave patterns. Haoqi Sun, PhD, assistant professor of neurology at Beth Israel Deaconess Medical Center, who developed the model, cautioned that while lifestyle factors like maintaining a healthy weight and engaging in regular exercise may have a positive impact, “there’s no magic pill to improve brain health.”
The UCSF Memory and Aging Center provides comprehensive evaluations and personalized treatment for dementia and related cognitive disorders, utilizing advanced diagnostic methods and participating in ongoing research to develop new therapies. The center is part of the Weill Institute for Neurosciences, which has been instrumental in advancements in dementia care and research.