Sleep Features Predict Cognitive Performance in Older Adults

Post by Deborah Joye

What's the science?

Sleep is an important feature of our lives that is crucial for the health of our brain and body. Not getting enough sleep can leave us sluggish and forgetful. Chronic sleep problems can make it difficult, perhaps impossible, to maintain our health and as we get older, sleep problems become increasingly likely. Cognitive difficulties also become more and more likely as we get older. One possibility is that age-related decreases in sleep and cognitive performance are associated with one another. Are there systematic changes in sleep features that can predict better or worse cognitive performance as we age? This week in Nature Human Behavior, Djonlagic and colleagues demonstrate that multiple aspects of brain function during sleep are associated with underlying age-related elements of cognitive and physical health in older adults.

How did they do it?

To investigate which aspects of brain function during sleep are associated with cognitive performance in older adults, the authors first analyzed data from the Multi-Ethnic Study of Atherosclerosis (MESA) – a diverse sample that included self-reported sleep measures from late-middle aged and elderly adults across the US. A subset of MESA participants also underwent polysomnography, which measures various features of brain and body function during sleep. To test whether their findings generalized to other samples of older adults, the authors repeated their analysis in another large independent dataset, the Osteoporotic Fractures in Men Study (MrOS) – a study of men 65 years or older who also underwent polysomnography. In both datasets, participants also completed measures of global cognitive function, processing speed, working memory, attention, and psychomotor ability, allowing direct comparison of objective sleep measures and cognitive performance. The two independent cohorts total almost 4000 older adults, both men and women.

What did they find?

The authors first found that, as expected, quite a few sleep measures changed with age. Briefly, older people tended to wake up more after going to sleep and have lower sleep efficiency (meaning less of the total time spent in bed is spent sleeping). Older people also exhibited reductions in sleep spindle frequency and intensity, as well as changes in the timing of unique sleep spindle features. Sleep spindles are short bursts of high-intensity, high-amplitude neural activity that occur during very slow brain activity (“slow oscillations”) during sleep and are thought to be important for benefits like memory consolidation and synaptic plasticity.

Next, the authors found that from roughly 170 sleep measures used in both datasets, 23 predicted processing speed and grouped into three broad areas: 1) sleep time and the ability to stay asleep; 2) the frequency and timing of sleep spindles, and 3) slow wave activity. First, better cognitive performance was associated with more time spent in REM sleep, higher sleep efficiency, and fewer times waking after falling asleep. Second, the authors found that more frequent and more intense sleep spindles (both fast and slow types) were associated with better cognitive performance. Finally, the authors observed several measures of slow wave activity that were associated with better cognitive performance including shorter slow oscillation duration and a stronger relationship between the timing of sleep spindles during a slow oscillation.

The author’s rigorous and multi-level analysis also revealed that regardless of chronological age, people who performed better cognitively had sleep measurement profiles that looked more like those in younger, healthier people. Perhaps even more interesting, the authors found that participants with diabetes/hypertension tended to have sleep measurements that looked like those seen in older (but healthier) individuals. Importantly, the authors note that subjective measures of sleep collected by self-report (which is a frequent tool of sleep study design) were only loosely associated with objective measurements gathered through polysomnography, providing a cautionary reminder when relying solely on the interpretation of self-report data.

What's the impact?

This study demonstrates that sleep features in older adults undergo age-related changes and that some features can reliably predict cognitive performance. Associations between sleep and cognitive performance have been reported before, but the present study significantly increases the scope of sleep features comprehensively tested and analyzed. A more detailed understanding of how sleep features change with age may allow the development of a “brain age” index that can compare an individual’s sleep features with their chronological age to determine possible pathology. Sleep behavior and brain activity are also modifiable, suggesting future treatment routes for age-related cognitive problems.

Djonlagic et al., Macro and micro sleep architecture and cognitive performance in older adults, Nature Human Behavior (2020). Access the original scientific publication here.