Fragmented sleep predicts slower mental processing speed the next day in older adults

New research published in the journal Sleep Health has found that sleep fragmentation, which refers to the amount of time spent awake in bed after initially falling asleep, is linked to slower mental processing speeds the next day in older adults. The findings suggest that the continuity of sleep may be more relevant to daily cognitive functioning in later life than the total number of hours slept.

Sleep is widely recognized as a fundamental pillar of physical health, yet its specific relationship to cognitive maintenance in older age remains a subject of scientific inquiry. Much of the existing data regarding sleep and aging relies on information collected at a single point in time, often asking participants to recall their general sleep habits over weeks or months.

This approach provides a broad overview but often misses the dynamic, day-to-day variations in rest and mental acuity that define daily living. The authors of the current study sought to fill this gap by observing how fluctuations in sleep quality on a specific night relate to cognitive performance on the immediately following day.

By focusing on older adults without dementia, the research team aimed to clarify if sleep disturbances act as a precursor to cognitive difficulties, which could inform preventative strategies for conditions like Alzheimer’s disease.

“We are interested in the role of sleep in Alzheimer’s disease and cognitive impairment with aging,” said study author Orfeu M. Buxton,
a professor of biobehavioral health and director of the Sleep, Health & Society Collaboratory at Pennsylvania State University.

“A Lancet Commission report in 2024 described modifiable factors for Alzheimer’s disease. They concluded that it wasn’t clear if sleep problems come before or after cognitive impairment, due to lack adequate data.”

“We wanted to expand our understanding of the role of sleep in cognition in older adults, and inform this consensus. Is sleep a modifiable factor that could be used to prevent or slow cognitive decline?”

The research team analyzed data from the Einstein Aging Study, utilizing a systematic random sampling method to recruit older adults residing in Bronx County, New York.

The final analysis included 261 participants who were at least 70 years old and free of dementia. The group was diverse, with approximately 47 percent identifying as non-Hispanic White and 40 percent as non-Hispanic Black. The average age of the cohort was around 77 years.

Over a period of 16 days, participants engaged in a rigorous ambulatory assessment protocol designed to capture real-world data. They wore a device called an Actiwatch on their non-dominant wrist continuously to objectively measure sleep patterns.

Unlike self-reported sleep diaries which can be inaccurate due to memory errors, actigraphy uses sensitive motion sensors to estimate sleep and wake intervals based on physical activity. This allowed the researchers to derive precise metrics regarding sleep timing and quality.

To measure brain function, participants used study-provided smartphones to complete brief cognitive tests six times per day. These assessments occurred in the morning, four times throughout the day at semi-random intervals, and once before bed.

This high-frequency testing allowed the researchers to collect a substantial amount of data, totaling over 20,000 valid cognitive assessments across the sample. The tests were specifically designed to evaluate domains often affected by aging and mild cognitive impairment.

The cognitive battery included four specific tasks. The Grid Memory task assessed visuospatial working memory by asking participants to recall the location of dots on a grid after a brief delay. The Symbol Search task measured processing speed, requiring users to quickly identify matching symbol pairs.

The Color Dots task evaluated visual working memory by asking participants to recall the color and location of specific items. Finally, the Color Shapes task tested the ability to bind features together, requiring participants to detect changes in combinations of shapes and colors.

The researchers examined several specific sleep characteristics derived from the actigraphy data. They calculated total nighttime sleep duration and sleep timing, defined as the midpoint of the sleep cycle. They also measured napping behavior, including the frequency and duration of daytime naps.

A primary focus was Wake After Sleep Onset, or WASO, which quantifies the total minutes spent awake during the night after the individual has initially fallen asleep. Higher WASO indicates more fragmented and disrupted sleep.

The statistical analysis utilized multilevel linear mixed-effect models to separate the results into two categories: between-person effects and within-person effects.

Between-person effects compare one individual’s average performance to another’s. Within-person effects compare an individual’s performance on a specific day to their own average performance, allowing the researchers to see how daily deviations in sleep impacted the next day’s brain function.

When looking at differences between people, the researcher found that participants with higher average levels of sleep fragmentation performed worse on several cognitive measures. Individuals who typically experienced more minutes of wakefulness at night exhibited slower average processing speeds on the Symbol Search task.

These individuals also demonstrated poorer performance on working memory tasks and the visual memory binding assessment. These associations held true even after the researchers adjusted for potential confounding variables.

To ensure the connections between sleep and cognition were robust, the statistical models included adjustments for various factors that could influence the results. The researchers controlled for age, sex, race and ethnicity, years of education, and income level. They also accounted for health-related variables such as history of stroke or heart attack and symptoms of depression.

Importantly, the analysis adjusted for sleep-disordered breathing and oxygen levels, which were measured using a single night of home pulse oximetry.

The within-person analysis revealed that on days following a night where a participant experienced more sleep fragmentation than their personal average, they exhibited slower processing speeds.

For every 30-minute increase in wakefulness during the night compared to the person’s norm, there was a measurable slowing of reaction time on the symbol search task the next day. This finding provides evidence that the immediate aftermath of a restless night includes a reduction in mental quickness.

The data indicated that total sleep duration was not a significant predictor of cognitive performance in this sample. Neither the total number of hours slept nor the frequency of naps showed a statistical association with the outcomes of the daily cognitive tests.

The results suggest that for this age group, the continuity and consolidation of sleep may play a stronger role in immediate cognitive functioning than the sheer quantity of sleep obtained.

“People with disrupted sleep had, on average, slightly worse cognitive performance,” Buxton told PsyPost. “We also found a small but significant relationship between last night’s sleep and today’s performance. Having a better night’s sleep with fewer disruptions is related to better cognitive performance the next day.”

The lack of association between sleep duration and cognition aligns with some previous research suggesting that sleep quality often outweighs quantity in older populations. As people age, sleep naturally becomes lighter and more fragmented, but the degree of that fragmentation appears to be the critical factor for mental performance. The study did not find that variations in sleep timing or napping habits influenced the next day’s cognitive scores in a significant way.

The authors discuss potential biological mechanisms that might explain why broken sleep harms cognitive speed. One theory involves the glymphatic system, a waste clearance pathway in the brain that is most active during deep sleep.

Chronic sleep fragmentation may interrupt this cleaning process, potentially allowing metabolic waste products to accumulate. Even short-term disruptions might interfere with synaptic function or neuronal communication, manifesting as slower reaction times the following day.

There are some limitations to the study. The sixteen-day assessment window, while intensive, is relatively short and may not account for seasonal variations in sleep or long-term health changes. Additionally, the study focused on a community-based sample, which increases the generalizability of the findings compared to clinical samples, but the results may not apply to younger adults or populations with different demographic profiles.

“We looked at effects day to day,” Buxton said. “We are still most interested in these types of relationships over time in years, and as cognitive function may decline. Does sleep quality decline before cognitive function differently than sleep after Mild Cognitive Impairment or Alzheimer’s Disease?”

The study, “Within- and between-person associations of sleep characteristics with daily cognitive performance in a community-based sample of older adults,” was authored by Orfeu M. Buxton, Qi Gao, Jonathan G. Hakun, Linying Ji, Alyssa A. Gamaldo, Suzanne M. Bertisch, Martin J. Sliwinski, Cuiling Wang, and Carol A. Derby.