Can sleep reveal whether the brain is aging faster than the body? New research links EEG and the risk of dementia

Can sleep reveal whether the brain is aging faster than the body? New research links brain waves during sleep to the risk of dementia

New research published on March 19, 2026, in the journal JAMA Network Open opens a new chapter in understanding the connection between sleep and cognitive decline. Scientists from the University of California, San Francisco and Beth Israel Deaconess Medical Center in Boston developed a machine-learning model that estimates the so-called “brain age” from brain activity during sleep. When the age estimated in this way was higher than the participants’ actual chronological age, the likelihood of later developing dementia also increased. According to the published results, every additional ten years of difference between an “older” brain and actual age was associated with approximately a 39 percent higher risk of incident dementia. In the opposite direction, when the estimated brain age was lower than the actual age, the risk was lower. This is a finding that has attracted great attention because it is based on noninvasive measurement of signals collected during sleep, without the need for more complex and more expensive diagnostic procedures.

What the researchers actually measured

Instead of relying only on the usual sleep indicators, such as total sleep duration, the share of individual sleep stages, or overall sleep efficiency, the authors analyzed the finer structure of brain waves recorded by electroencephalography, or EEG. The model included 13 microstructural features of brain activity during sleep in the assessment, therefore a series of small but informative patterns that are often lost in classic summaries of polysomnographic findings. This is precisely where the main novelty of the paper lies: the researchers argue that the “broader” picture of sleep does not capture the brain’s complex physiology well enough, while a more detailed wave analysis can reveal what standard indicators fail to recognize. This approach is especially important in the field of dementia, where early risk assessment is one of the key open questions of modern medicine.

The analysis included 7,105 adults from five long-term cohort studies in the United States. At the start of follow-up, the participants did not have dementia, and their average age differed among the cohorts, from the late fifties to over 80 years. Depending on the cohort, follow-up lasted from approximately 3.6 to 16.9 years, and a total of 1,082 new cases of dementia were recorded. It is precisely this combination of a larger sample, multi-year follow-up, and data from several shared cohorts that gives the study more weight than smaller or shorter papers that had previously tried to link sleep and cognitive outcomes.

Why “brain age” is more important than ordinary age

Chronological age tells us how much time has passed since birth, but it does not necessarily tell us how biologically “worn” individual organ systems are. In recent years, the concept of biological age has increasingly been used in cardiology, oncology, and neurology to explain why two people of the same age can have very different health risks. In this research, scientists applied that concept to the brain and tried to estimate whether it ages faster or slower than the rest of the organism. The results suggest that a signal carrying additional information about future cognitive health can be read from brain activity during sleep.

It is important to emphasize that this is not a diagnosis of dementia, but an assessment of risk. In other words, the model does not claim that a person with an “older” brain will necessarily develop dementia, but that on average the probability is higher than in a person whose brain patterns during sleep look “younger.” The authors additionally showed that this association did not disappear even after they took into account a series of other factors in statistical models, including sex, education, body mass index, smoking, physical activity, certain comorbidities, sleep apnea, and genetic risk associated with the APOE ε4 variant. This does not prove causality, but it suggests that the observed signal is not merely a by-product of several known risk factors.

Which sleep patterns drew particular attention

Among the EEG features that contributed to estimating brain age are patterns already known to be connected with memory and brain recovery. Among them are delta waves, typical of deep sleep, and so-called sleep spindles, short episodes of faster brain activity that are often linked to memory consolidation. That is precisely why the finding is not only statistically interesting but also biologically interesting: the models did not “extract” an arbitrary signal, but patterns that make sense from the perspective of sleep neuroscience proved important.

One of the more notable details in the paper concerns so-called kurtosis, a statistical measure that in this context reflects the tendency of the EEG signal to occasionally produce pronounced, large peaks. In the analysis, higher kurtosis, especially in the N2 sleep stage, was associated with a lower risk of dementia. The authors state that this could reflect the activity of related patterns such as K-complexes, large waves that are part of normal sleep architecture. This part of the finding is of particular interest to researchers because it shows that a protective signal does not necessarily have to be only “more sleep” or “longer sleep,” but a qualitatively specific way in which the brain organizes its electrical activity during the night.

What classic sleep indicators miss

One of the more important messages of the paper is that traditional sleep metrics may not be sufficient when trying to predict long-term neurological outcomes. In earlier pooled analyses of similar cohorts, no clear and statistically significant association was found between the risk of dementia and general measures such as time spent in individual sleep stages, overall sleep efficiency, or simple ratios of wakefulness and sleep. That does not mean those measures are not useful in clinical practice, but that by themselves they may not capture the “signature” of brain aging well enough. The new paper therefore suggests a shift from the coarse architecture of sleep toward a fine analysis of its microstructure.

Such a shift also has practical consequences. If it turns out that detailed EEG sleep patterns are precisely the more stable and more sensitive early markers of cognitive decline, this could change the way risk is assessed in the future in older people, but also in middle-aged people who still have no symptoms at all. In theory, an overnight recording of brain activity could become one of the elements of preventive assessment, similarly to how blood pressure, blood sugar, or cardiovascular risk factors are monitored today.

Could findings like these soon end up in smart devices

The authors of the paper believe that one of the greatest advantages of this approach is the fact that EEG signals during sleep can be collected noninvasively. Today, such data are most often obtained in sleep laboratories or as part of research, but the development of wearable technologies opens the possibility that at least some similar measurements could in the future also be carried out outside clinical institutions. This, however, does not mean that smart bracelets or home devices will soon reliably “predict” dementia. Between a research result and broad application in the population stands a series of steps: additional validation, device standardization, accuracy checks in different populations, and the very sensitive question of how to communicate such information to people without unnecessary intimidation.

That is precisely why scientific potential should be distinguished from a finished product. The study shows that a useful signal about brain aging can be extracted from sleep, but it still does not prove that the model is ready for routine home use or for independent clinical decision-making. In medicine, it is especially important to avoid a situation in which a screening tool would begin to be perceived as a diagnostic verdict. Dementia is a complex syndrome with different causes and patterns of development, and such an index represents only one piece of the puzzle.

The broader picture: why early risk assessment matters

The World Health Organization describes dementia as a syndrome caused by a number of diseases that over time damage nerve cells and the brain, leading to a decline in cognitive abilities beyond what is considered the usual consequence of biological aging. The organization also emphasizes that age is the strongest known risk factor, but not the only one, and that dementia is not an inevitable consequence of old age. Factors associated with lower risk include physical activity, not smoking, avoiding harmful alcohol use, controlling body weight, and maintaining healthy levels of blood pressure, cholesterol, and blood sugar. In that context, the new research on sleep gains additional weight because it tries to identify an early, measurable signal that could fit into the broader framework of prevention.

The U.S. National Institute on Aging had also previously warned that lack of sleep in middle age is associated with a higher later risk of dementia. In one earlier analysis, people in their fifties and sixties who slept six hours or less had a higher risk of a subsequent dementia diagnosis than those who slept about seven hours. The new research goes a step beyond the number of hours of sleep alone and shows that it is not only crucial how long someone sleeps, but also what the neurophysiological “imprint” of that sleep is like. For public health, this is an important message: the quality and structure of sleep may carry information that is just as important, and perhaps even more important, than duration alone.

What can and cannot be concluded from the study

Although the results are strong and statistically convincing, they do not mean that a causal link has been definitively proven. It is not entirely clear whether an unfavorable pattern of brain activity during sleep itself accelerates the process of neurodegeneration, or whether it partly reflects very early, still clinically invisible changes that have already begun in the brain. It is also possible that both processes are interconnected. That is precisely why the authors and expert commentators emphasize that this is a promising biomarker of risk, not a final answer to the question of how dementia arises.

The composition of the included cohorts should also be kept in mind. Some were highly specific, for example a cohort that included only men and another that included only women, and racial and ethnic representation was also not equal in all groups. Some outcomes in individual cohorts were determined in part through hospital diagnostic codes, which may have limited sensitivity. None of this nullifies the findings, but it is a reminder that it will be important to confirm them in other populations as well, with additional follow-up methods and in real clinical conditions.

Can better sleep slow brain aging

That is the question likely to interest the wider public the most, but there is still no simple answer to it. The authors of the paper note that previous studies suggest that treating sleep disorders can change the patterns of brain waves associated with sleep. They also mention the importance of reducing the risk of sleep apnea, among other things through body-weight regulation and more movement. But they do not offer a “magic pill” for brain health either. The fairest reading of the study would be this: sleep is probably one of the important windows into the state of the brain, and possibly one of the levers that can be influenced, but it has not yet been proven that every change in sleep will directly and measurably reduce the future risk of dementia.

Still, the findings further reinforce the message that medicine has been repeating more and more often in recent years: sleep is not a passive pause in the organism’s functioning, but an active biological process deeply connected with memory, recovery, and long-term brain health. If it is confirmed that a reliable estimate of “brain age” can be obtained from EEG during sleep, this could in the future help doctors identify earlier the people who need closer monitoring, more aggressive treatment of sleep disorders, or a broader preventive program. For now, the most important message is that sleep quality can no longer be viewed as a secondary habit, but as one of the serious indicators of what is happening to the brain long before the first clear symptoms of forgetfulness and cognitive decline appear.

Sources:

• - JAMA Network Open – original scientific paper on the association between the EEG-based brain age index during sleep and the risk of incident dementia (link)
• - UC San Francisco – official announcement on the research results, the authors, and the main findings of the study (link)
• - World Health Organization – overview of dementia as a public health priority, symptoms, risks, and prevention options (link)
• - National Institute on Aging – earlier official announcement on the association between shorter sleep in middle age and a higher later risk of dementia (link)