A machine learning analysis of brain waves during sleep in over 7,000 participants has found that a “brain age” higher than a person’s actual age predicts a higher risk of developing dementia later on. Results published in JAMA Network Open show that every 10-year increase in brain age over actual age made dementia risk rise by 39%.
The machine learning model integrates measurements of multiple microstructural features found in brain waves from sleep electroencephalography (EEG) recordings. While earlier studies had shown no significant link between dementia and traditional measures of sleep quality, such as time spent in each sleep stage, analyzing these finer-scale metrics could provide a much more in-depth look at brain processes implicated in memory and cognition.
“Sleep disturbances are increasingly recognized as early indicators and potential modifiable risk factors for dementia,” said Yue Leng, MBBS, PhD, associate professor of psychiatry at the UCSF School of Medicine and senior author of the study. “However, 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, presenting an opportunity to develop digital markers for the early detection of dementia.”
Leng and colleagues analyzed data from 7,105 adults across five different studies, all of which had EEG recordings taken overnight at home. The machine learning model looked at 13 microstructural features of brain waves, including measuring sleep depth by looking at delta waves or quantifying bursts of brain activity known as spindles, which have been linked to memory consolidation during sleep.
As the brain ages, all of these sleep brain wave patterns are known to change. To better summarize the analysis of multiple brain wave features, the researchers compared each individual’s results to those expected at their age. The difference between the measured “brain age” and the patient’s chronological age is known as the brain age index (BAI).
Results showed that a higher BAI—indicating an older brain—was linked to a significantly higher risk of developing dementia. In particular, for each 10-year increase in BAI, the risk of dementia rose by 39%. This association remained significant after accounting for other factors known to influence dementia risk, including sex, age, education, BMI, physical activity, other medical conditions and genetic risk factors such as APOE carrier status.
Since sleep EEG recordings can be collected non-invasively at home, these findings point at the potential of this method to detect dementia risk in nonclinical settings with the use of wearable technologies. Although more work will be needed to confirm the causal links between sleep patterns and dementia risk, this work points at the possibility of reducing dementia risk by improving brain age, as previous studies have found that treating sleep disorders can make meaningful changes in sleep brain wave patterns.
“Better body management, such as lowering body mass index and increasing exercise to reduce the likelihood of apnea, may have an impact,” said Haoqi Sun, PhD, assistant professor of neurology at Beth Israel Deaconess Medical Center and lead author of the study. “But there’s no magic pill to improve brain health.”
