An Aging Clock Built from Sleep Electroencephalography Data
Quality of sleep tends to decline with age for reasons both physical and neurological; sleep apnea is a concern for many older people. A broad body of literature connects sleep issues with risk of neurodegenerative conditions. Thus researchers can plausibly expect to take sleep assessment data from a population of people at various ages, and employ machine learning strategies to develop an aging clock derived from that data. Any sufficiently complex data set that changes with age can be used in this way. Researchers here report on an implementation of this approach to measuring the aging of the brain, and produce an aging clock that can predict dementia risk based on sleep electroencephalography results recorded during a sleep study.
Sleep disturbances are increasingly recognized as early indicators and potential modifiable risk factors for dementia. However, the macrolevel sleep architecture has shown inconsistent associations with cognitive impairment and incident dementia. These broad sleep metrics do not fully capture the complex and multidimensional nature of sleep physiology. In contrast, the microstructure of sleep electroencephalography (EEG) directly reflects the neural processes with explicit functional implications. To capture these complex patterns, we developed a sleep EEG-based brain age using a novel, interpretable machine learning approach that integrates multiple age-dependent EEG microstructures into a single agelike number. The difference between brain age and chronological age is termed the brain age index (BAI).
For this individual participant data meta-analysis, sleep study data from 5 community-based longitudinal cohorts were pooled. These cohorts included the Multi-Ethnic Study of Atherosclerosis (MESA; 2010-2013), the Atherosclerosis Risk in Communities (ARIC) study (1987-1989), the Framingham Heart Study-Offspring Study (FHS-OS; 1995-1998), the Osteoporotic Fractures in Men Study (MrOS; 2003-2005), and the Study of Osteoporotic Fractures (SOF; 2002-2004). This meta-analysis included 7,105 participants.
The median time to dementia was 4.8 years in the MESA cohort (n = 119 [6.6%]), 16.9 years in the ARIC cohort (n = 354 [19.7%]), 13.1 years in the FHS-OS cohort (n = 59 [9.6%]), 3.6 years in the MrOS cohort (n = 470 [17.8%]), and 4.6 years in the SOF cohort (n = 86 [34.3%]). Across the cohorts, each 10-year increase in BAI was associated with a 39% higher risk of incident dementia (hazard ratio [HR] 1.39) after adjustment for covariates. These associations remained after additional adjustment for comorbidities and apnea-hypopnea index scores (HR 1.31) and apolipoprotein E ε4 (HR 1.22), and they were consistent across sex and age groups.