Two Sleep and Activity Patterns Linked to Nearly Double the Dementia Hazard in 12-Year Study

Key Takeaways

  • A study of 3,973 older adults found that two combinations of disrupted daily rhythms, disturbed sleep, and activity patterns were associated with nearly twice the adjusted hazard of dementia.
  • During a median follow-up of 12.3 years, 379 participants received a dementia diagnosis.
  • Poor sleep was not consistently associated with worse cognitive outcomes among participants with stronger daily rest-activity rhythms and higher physical activity.
  • The most sedentary group experienced faster cognitive decline, but its association with diagnosed dementia weakened after adjustment for health conditions.
  • The findings cannot establish whether disrupted sleep and activity contribute to dementia or are early consequences of neurological changes.

Introduction

Middle-aged woman lying awake in bed at night, looking tired and thoughtful, with a digital alarm clock on her bedside table.

Disrupted sleep and daily activity patterns were associated with higher dementia hazard in a 12-year study of older adults. Researchers have not established whether these disturbances contribute to dementia or reflect early neurological changes.

Two older adults may sleep poorly but have very different daytime routines. One remains active and follows a relatively consistent pattern of movement and rest. The other spends much of the day inactive, with a less distinct separation between activity and sleep.

Could these differences help explain why some people experience greater cognitive decline than others?

A study published November 18, 2026, in JAMA Network Open suggests that the answer may depend on the broader sleep-wake cycle rather than sleep quality alone.

Researchers followed 3,973 older adults for a median of 12.3 years and identified two combinations of disturbed sleep and daily activity patterns associated with nearly twice the adjusted hazard of dementia diagnosis.

Yet not everyone with poor sleep had an elevated dementia hazard. Participants with stronger daily rest-activity rhythms and higher physical activity generally had similar cognitive outcomes regardless of sleep quality.

The findings raise an important question: do certain sleep-wake patterns contribute to cognitive deterioration, or could they be early signs of neurological changes already underway?

The observational study cannot distinguish between these possibilities, but it provides evidence that sleep, movement, and daily rhythms should be examined together when studying cognitive aging (Vidil et al., 2026).

What the Study Examined

Researchers analyzed participants from the Whitehall II study, a long-running investigation of British civil servants.

The dementia analysis included 3,973 adults aged 60 or older who had no recorded dementia diagnosis when their sleep and activity patterns were assessed in 2012 and 2013.

Their median age was 68.4 years, and 26% were women.

Participants wore wrist accelerometers for nine consecutive days. The devices measured movement, allowing researchers to characterize four aspects of the sleep-wake cycle:

  • Rest-activity rhythms: How strongly and consistently participants alternated between active and resting periods.
  • Daytime activity: Time spent moving or sedentary.
  • Sleep: Duration, efficiency, and fragmentation.
  • Chronotype: The timing of sleep and activity.

Using 36 measurements, researchers assigned participants to nine previously developed sleep-wake profiles.

Dementia diagnoses were identified through linked hospital, mental health, and mortality records until March 2025.

The researchers also examined changes in memory, reasoning, verbal fluency, and processing speed among 3,806 participants with cognitive testing data.

Two Sleep-Wake Patterns Stood Out

During follow-up, 379 participants, or 9.5% of the cohort, received a dementia diagnosis.

Researchers compared the nine profiles with a reference group characterized by stable daily rhythms, more light physical activity, and relatively good sleep.

Two groups showed significantly higher dementia hazards after adjustment for demographic, behavioral, and health-related factors.

Sleep-wake pattern Dementia cases Adjusted hazard ratio
Stable rhythms, more light activity, good sleep 52 of 626 Reference
Disrupted rhythms, low activity, poor sleep 61 of 362 1.94 (95% CI, 1.32–2.84)
Disrupted rhythms and restless sleep 35 of 232 1.96 (95% CI, 1.27–3.03)

The first higher-hazard group had weak, fragmented daily rhythms, little physical activity, and shorter, inefficient, frequently interrupted sleep.

The second group was different. These participants were relatively active during the day but experienced unusually restless sleep and a weaker distinction between their active and resting periods.

Both groups had approximately twice the adjusted dementia hazard of the reference group.

What Do Those Numbers Mean in Absolute Terms?

Dementia was diagnosed in:

  • 8.3% of participants in the reference group (52 of 626).
  • 16.9% of those with disrupted rhythms, low activity, and poor sleep (61 of 362).
  • 15.1% of those with disrupted rhythms and restless sleep (35 of 232).

These are observed proportions, not adjusted estimates of an individual’s probability of developing dementia.

The hazard ratios account for measured differences in demographic, behavioral, and health-related characteristics.

For example, a hazard ratio of 1.94 represents an estimated 94% higher dementia hazard relative to the reference group. It does not mean a 94% probability of developing dementia.

Because the study was observational, the differences cannot establish that the sleep-wake patterns themselves caused dementia.

Poor Sleep Alone Did Not Tell the Whole Story

One of the study’s most interesting findings involved participants who slept relatively poorly but maintained stronger daily rest-activity rhythms and higher physical activity.

They did not show a statistically significant increase in dementia hazard compared with the reference group.

Another group had weak daily rhythms and low physical activity but relatively good sleep. Its dementia hazard was also not significantly elevated.

These differences challenge the idea that a single measure, such as sleep duration or nighttime awakenings, adequately captures the relationship between sleep and cognitive aging.

The broader daily pattern appeared more informative.

However, the researchers identified clusters of characteristics rather than formally testing every possible interaction between sleep and activity.

The study therefore does not prove that exercise offsets the effects of poor sleep.

The Most Sedentary Group Experienced Faster Cognitive Decline

The researchers examined cognitive performance separately from dementia diagnoses.

Among 3,806 participants, those with disrupted daily rhythms, low activity, and poor sleep showed an additional 0.14-standardized-unit decline in global cognitive performance over 10 years compared with the reference group.

The most sedentary profile, which also had the weakest and most fragmented daily rhythms, showed an additional decline of 0.28 standardized units.

The researchers estimated that these differences were comparable in magnitude to approximately four and nine years of age-related cognitive decline, respectively.

These are statistical comparisons, not evidence that participants’ brains biologically aged by those amounts.

Lower Cognitive Scores Are Not the Same as Faster Decline

The study also distinguished between participants who already performed worse on cognitive tests when monitoring began and those whose performance deteriorated more rapidly afterward.

For example, the restless-sleep group had significantly lower cognitive scores at baseline than the reference group.

However, its subsequent rate of cognitive decline was not significantly faster after full adjustment.

This distinction matters because a lower score at one point in time does not necessarily indicate more rapid deterioration.

The most sedentary group also had a higher estimated dementia hazard, but that association was no longer statistically significant after adjustment for health-related factors.

Together, these results show why cognitive test performance and dementia diagnoses should not be treated as interchangeable outcomes.

Why Could Disrupted Sleep and Daily Rhythms Matter?

The body follows approximately 24-hour biological rhythms that help regulate sleep, alertness, hormone release, and other physiological processes.

These rhythms are coordinated by the brain’s circadian timing system, which responds to environmental signals such as light and helps organize daily activity and rest.

Researchers have proposed several ways disturbances in these rhythms might be connected to cognitive health.

One possibility involves sleep-related changes in inflammation and the regulation of amyloid-beta, a protein associated with Alzheimer’s disease pathology.

A review published in The Lancet Neurology examined evidence connecting circadian disturbances with neurodegenerative diseases, including the possibility that the relationship operates in both directions (Leng et al., 2019).

Physical activity may also play a role. Regular movement is associated with cardiovascular and metabolic health, both relevant to cognitive aging.

However, these proposed mechanisms do not establish that the patterns observed in the Whitehall II study caused dementia.

The researchers measured movement and rest, not the biological processes responsible for subsequent neurological changes.

Could Dementia Be Causing the Sleep Problems Instead?

This is one of the most important unanswered questions.

Alzheimer’s disease and other neurodegenerative conditions can begin developing many years before symptoms become severe enough for a clinical diagnosis.

During that period, changes in brain function may affect sleep, movement, and daily rhythms.

An irregular sleep-wake pattern could therefore be an early consequence of neurological disease rather than a cause.

The researchers investigated this possibility through additional analyses.

First, they excluded participants with fewer than five years of follow-up to reduce the possibility of reverse causation.

The associations remained statistically significant for the two main higher-hazard profiles:

Sleep-wake pattern Adjusted hazard ratio after five-year restriction
Disrupted rhythms, low activity, poor sleep 1.63 (95% CI, 1.03–2.56)
Disrupted rhythms and restless sleep 2.29 (95% CI, 1.41–3.70)

This suggests the associations were not explained entirely by dementia diagnosed shortly after the initial measurements.

However, the analysis cannot eliminate reverse causation. Biological changes associated with dementia may begin long before diagnosis.

The researchers also repeated the analysis after excluding 702 participants with mild cognitive impairment at baseline.

The association for the restless-sleep profile remained statistically significant, with a hazard ratio of 1.79.

The association for the low-activity, poor-sleep profile weakened and was no longer statistically significant.

This difference matters. It raises the possibility that some of the observed relationship reflected cognitive problems already developing when sleep and activity were measured.

Further adjustment for genetic susceptibility and blood-based Alzheimer’s biomarkers in a smaller sample did not eliminate the two main associations, but it could not resolve the question of causality.

What Do Earlier Studies Show? Are There Contradictions?

The new findings broadly agree with research linking weaker daily activity rhythms to dementia, although results have not been consistent across every measure.

A US study published in Neurology examined 2,183 older adults from the Atherosclerosis Risk in Communities study.

Participants wore monitoring patches that recorded activity patterns, and researchers followed them for a median of approximately three years.

During follow-up, 176 participants developed dementia.

Each standard-deviation decrease in relative rhythm amplitude, a measure of the distinction between active and resting periods, was associated with 54% higher adjusted dementia hazard.

Greater rhythm fragmentation was also associated with higher hazard, although the magnitude was smaller: 19% per standard-deviation increase (Wang et al., 2026).

These findings support the broader observation that weaker and more fragmented daily activity rhythms can precede dementia diagnoses.

However, the studies differed in several ways.

The US participants were older, the follow-up was shorter, and the researchers examined individual rhythm measurements rather than the combined sleep-wake profiles used in the British study.

Not every rhythm measure showed a significant association. For example, the US study did not identify a statistically significant association for its measure of day-to-day rhythm stability.

This is not a direct contradiction. It indicates that different aspects of circadian activity may have different relationships with dementia.

The British study adds another complication: some participants with weak daily rhythms did not show significantly higher dementia hazards.

Similarly, poor sleep did not consistently predict worse outcomes among participants whose overall rhythms and activity levels were relatively strong.

The evidence therefore supports a relationship between disrupted daily rhythms and cognitive health, but it does not establish that all irregular sleep or activity patterns carry the same implications.

Could Other Health Problems Explain the Association?

Several alternative explanations deserve consideration.

Undiagnosed sleep disorders: Obstructive sleep apnea, restless legs syndrome, and periodic limb movement disorder can disturb sleep and produce unusual nighttime activity patterns. The researchers lacked detailed information on sleep disorders.

Underlying illness: Cardiovascular disease, diabetes, chronic pain, and reduced mobility may affect sleep and activity while also being associated with cognitive outcomes.

Depression and medication use: Depressive symptoms and medications affecting the central nervous system can alter sleep, movement, and alertness. The researchers adjusted for these factors, but residual confounding remains possible.

Early neurological changes: Preclinical dementia may alter sleep and activity years before diagnosis.

These explanations are not mutually exclusive. Several could contribute to the associations.

What Were the Study’s Main Limitations?

The study’s strengths include objective movement measurements, long follow-up, and analyses of both dementia diagnoses and cognitive performance.

Nevertheless, several limitations affect interpretation.

Short monitoring period: Sleep and activity were assessed over nine days, which may not represent patterns throughout the following decade.

Limited generalizability: Participants were former British civil servants who were generally healthier than the wider population. Women represented only 26% of the sample.

Incomplete sleep-disorder information: The researchers could not adequately distinguish the influence of conditions such as sleep apnea or nighttime movement disorders.

Dementia identification: Diagnoses came from health records rather than repeated comprehensive clinical examinations of every participant. Some cases may have been missed.

No dementia subtype analysis: The study could not establish whether the associations differed for Alzheimer’s disease, vascular dementia, or other conditions.

Complex profiles: The clustering method identified combinations of characteristics but did not formally establish how individual sleep and activity factors interacted.

Observational design: Even with extensive adjustment, the findings cannot establish that changing sleep or activity patterns would reduce dementia incidence.

What Should Older Adults Take From the Findings?

The study does not mean that someone who sleeps poorly or follows an irregular routine is destined to develop dementia.

It also does not establish that a wearable device, a strict bedtime, or a particular exercise routine can prevent the disease.

Instead, the findings raise the possibility that changes in sleep and daytime activity may provide useful information about cognitive health when considered together.

Persistent sleep disturbances, pronounced daytime inactivity, or substantial changes in usual sleep-wake patterns may warrant discussion with a healthcare professional, particularly when accompanied by memory problems or difficulties with everyday activities.

Treatable sleep disorders and other medical conditions can affect these patterns.

For now, the evidence supports investigating sleep and activity together, not prescribing a specific routine as a proven dementia-prevention strategy.

Final Thoughts

This 12-year study suggests that the relationship between sleep and dementia is more complicated than simply getting too little sleep.

Two combinations of disrupted daily rhythms and disturbed sleep were associated with nearly twice the adjusted dementia hazard. Meanwhile, participants with stronger rest-activity rhythms and higher physical activity generally did not show worse cognitive outcomes simply because their sleep was poorer.

The findings also highlight an unresolved question: whether these patterns contribute to neurodegeneration or emerge as early consequences of changes already occurring in the brain.

For older adults, persistent changes in sleep and daily activity may be worth discussing with a healthcare professional, particularly when accompanied by memory difficulties or changes in everyday functioning.

The study does not prove that improving sleep or increasing physical activity will prevent dementia. Its main contribution is showing why researchers may need to consider the entire 24-hour sleep-wake cycle when investigating cognitive aging.

Funding and Disclosures

The Whitehall II study received support from the US National Institute on Aging, the UK Medical Research Council, and the Wellcome Trust.

Additional researcher support came from the European Union, Fondation Alzheimer, and Fondation Vaincre-Alzheimer.

Séverine Sabia disclosed grants from the European Commission, Fondation Alzheimer, and Fondation Vaincre-Alzheimer during the study. No other disclosures were reported.

The authors stated that funders had no role in the study’s design, data collection, analysis, interpretation, manuscript preparation, or publication decision.

References

Leng, Y., Musiek, E. S., Hu, K., Cappuccio, F. P., & Yaffe, K. (2019). Association between circadian rhythms and neurodegenerative diseases. The Lancet Neurology, 18(3), 307–318. https://doi.org/10.1016/S1474-4422(18)30461-7

Vidil, S., Meneghel Danilevicz, I., Dugravot, A., Cavaillès, C., Garcia Ocampo, N., Singh-Manoux, A., & Sabia, S. (2026). Sleep-wake cycle profiles and incident dementia and cognitive decline in older adults. JAMA Network Open, 9(11), e2644843. https://doi.org/10.1001/jamanetworkopen.2026.44843

Wang, W., Wanigatunga, A. A., Etzkorn, L. H., Rabinowitz, J. A., Palta, P., Pike, J. R., Dougherty, R. J., Zipunnikov, V., Marino, F. R., Crainiceanu, C., Spira, A. P., Schrack, J. A., & Chen, L. Y. (2026). Association between circadian rest-activity rhythms and incident dementia in older adults: The Atherosclerosis Risk in Communities Study. Neurology, 106(2), e214513. https://doi.org/10.1212/WNL.0000000000214513