Sleep is essential for both physical and mental health. Not getting the right amount of sleep or having poor quality sleep is linked to a wide range of poor outcomes, including all-cause mortality, poor cardiovascular and metabolic health, and mood disturbances including anxiety, depression and psychosis (Shah et al., 2025; Larsen 2025; Ferguson 2024). Poor sleep is also considered to be both a symptom of neurodegenerative diseases, such as Alzheimer’s disease, and a contributing factor to the condition’s onset and progression (Pathmanathan et al., 2025).
A new study by Bingtao Weng and colleagues, published in Sleep, looks at how changes in sleep patterns over time might relate to transitions from normal cognition to dementia. Using data from two large-scale longitudinal cohorts, the researchers investigated whether maintaining or changing sleep behaviours was associated with cognitive performance over time, and whether any particular sleep habits were linked to worse cognitive outcomes. These findings offer important insights into how sleep may serve as both a marker and a modifiable risk factor for cognitive decline.
We know sleep can get worse with aging, but do specific sleep patterns predict cognitive changes and dementia?
Methods
The study used data from the UK Biobank and China Health and Retirement Longitudinal Study (CHARLS), two large prospective cohorts. Self-reported sleep was assessed at two time points. Sleep dimensions measured included:
- UK Biobank: sleep duration, difficulty waking, chronotype, napping, insomnia, snoring, and daytime dozing.
- CHARLS: sleep duration, napping, and sleep quality.
Sleep duration was calculated as the sum of naps and nighttime sleep, with 7 to 8 hours considered optimal. Participants with very short (<4 hours) or very long (>12 hours) sleep were excluded. Chronotype was classified as “morningness” (preference for activity earlier in the day) or “eveningness” (preference for activity later in the day).
For each sleep dimension, participants were classified based on whether their responses remained stable or changed over time. Responses were analysed as binary outcomes (e.g., napping usually or not usually).
Two main analyses were performed:
- Change-to-change: Associations between sleep changes and changes in cognitive performance in people without dementia.
- Change-to-dementia: Associations between sleep changes and the risk of developing dementia over time.
For the UK Biobank cohort, overall cognitive scores were examined and hazard ratios calculated to see if changes in sleep predicted dementia. For the CHARLS cohort, the focus was on whether changes in sleep dimensions were linked to transitions between normal cognition, mild cognitive impairment (MCI), and dementia.
Results
8,994 participants were included in the change-to-change analyses, and 14,720 participants were included in the change-to-dementia analyses.
There were some differences between the UK Biobank and CHARLS cohorts. Participants in the CHARLS cohort were more likely to have persistent non-optimal sleep durations, whilst UK Biobank participants were more likely to maintain persistent optimal sleep durations. Napping was also much more common among UK Biobank participants (92%) than CHARLS participants (48%).
Improving sleep from non-optimal to optimal was linked with greater improvements in cognitive scores. Transitioning from eveningness to morningness was associated with improved overall cognition at follow-up, but also with a decline in reasoning ability.
Two patterns emerged that were associated with increased risk of all-cause dementia, after adjusting for potential confounders:
- Shifting from optimal to non-optimal sleep duration (hazard ratio, HR = 1.82, 95% CI = 1.21% to 2.75%, p= .005).
- Shifting from habitual napping to no napping (HR = 2.13, 95% CI = 1.17% to 3.88%, p= .015).
In terms of specific changes in cognitive status, declining overall sleep quality and transitioning to non-optimal sleep duration were also linked to cognitive decline in participants who had normal cognition at baseline (OR=1.06, p = 0.006), and napping cessation was associated with increased risk of progression from mild cognitive impairment to dementia (OR = 1.16, p = 0.001).
Participants who shifted from optimal to non-optimal sleep or stopped napping showed higher risk of cognitive decline and dementia.
Conclusions
Shifting from non-optimal to optimal sleep duration and from eveningness to morningness chronotype was associated with better cognitive outcomes, whilst napping cessation and shifting from optimal to non-optimal sleep duration was associated with worse cognitive outcomes and higher risk of developing dementia.
The authors concluded that their findings:
highlight the critical role that changes in sleep dimensions play in both cognitive decline and the onset of dementia, offering valuable insights for the development of targeted interventions aimed at combating cognitive deterioration.
Improving sleep duration or timing may support cognition, while worsening sleep habits might signal increased dementia risk.
Strengths and limitations
A key strength of this paper is its use of two large prospective cohorts involving participants from across the UK and China, allowing for sufficient statistical power and demonstrating cross-cultural findings. The cohorts also had quite different sleeping patterns on several sleep dimensions, including how many participants engaged in frequent napping and had optimal sleep duration.
However, the use of these cohorts also limits the amount and type of data that the researchers were able to use. Cohort questionnaires often include relatively few questions for a given topic and may not reflect those considered most important for a specific question or population. For example, the CHARLS study included a question on nighttime sleep duration and post-lunch napping, whilst the UK Biobank data asked participants about 24-hour sleep duration, without separating nighttime sleep and naps. Similarly, chronotype was measured by asking participants if they were a morning or evening person, rather than about sleep-wake patterns. More information about sleep timing, given the importance of circadian rhythms in dementia and cognition, would have benefited the study. Moreover, self-reported sleep measures, particularly in older adults and those with cognitive impairment, can capture important reflections about sleep like subjective sleep quality and whether someone feels sleepy during the day, but are often poor proxy measures for objective sleep parameters like sleep duration or napping frequency (Landry et al., 2015).
Recognising the importance of sleep to health, the UK Biobank has added a much more in-depth sleep questionnaire to its study. It would be really interesting to use data from this dedicated sleep survey, combined with their actigraphy data that can objectively measure sleep-wake patterns, to replicate and extend the findings from the current study.
Whilst the paper itself briefly acknowledges the possibility of ‘reverse causation’ in its limitations, the abstract (often the only part people may read) concludes that optimal sleep duration, morning chronotypes, and preserving napping habits “are essential for dementia prevention”. This is a very strong statement and could mislead readers. Sleep is considered to have a bidirectional relationship with dementia, which means it is often difficult to tease apart where sleep is a risk factor of dementia and cognitive decline or whether it is a symptom of the underlying neurodegeneration. Whilst this study shows associations between changing sleep patterns and dementia, it is known that sleep patterns change as neurodegenerative conditions progress, and neurodegeneration typically precedes cognitive symptoms by many years. We cannot be confident that the changing sleep patterns observed here are not symptomatic of neurodegeneration – nor that efforts to change or maintain sleep patterns would prevent dementia. Clinical trials that actively seek to change or maintain sleep patterns, demonstrating meaningful clinical change, are needed to substantiate this conclusion – and we’re not quite there yet.
Large, cross-cultural cohorts allowed robust analysis of sleep changes and cognition, but self-reported measures and reverse causation limit certainty.
Implications for practice
Sleep is increasingly recognised as a potentially modifiable risk factor for dementia, and this study supports the idea that by changing sleep, we might be able to improve cognitive and clinical outcomes and possibly reduce risk or progression of dementia. It also highlights how napping, in addition to nighttime sleep, may be an important avenue for future interventions. Given that sleep has been excluded from the Lancet’s ‘modifiable risk factors for dementia’ list for having insufficient evidence, studies such as Weng et al. (2026) help us build the evidence base and support the rationale for large-scale clinical trials into improving sleep.
To nap or not to nap, that is the question. When we talk about sleep (e.g., at public events or with people with memory problems), whether napping is good or bad for you often comes up. Whilst this study suggests that stopping napping is associated with cognitive decline, others have found that having longer naps is associated with an increased risk of cognitive impairment (Fang et al., 2023), but short duration naps may be beneficial (Li et al., 2023). Nap timing could also matter, with one study suggesting morning naps are linked to higher risk of Alzheimer’s disease whilst afternoon naps might be protective (Gao et al., 2025). Whilst practicing good sleep hygiene and getting sufficient sleep is recommended for overall health and specifically brain health, more research is needed before we can confidently advise people to nap (or not to nap).
One thing that we should all be really mindful of in discussions around modifiable risk factors, like sleep, is the messaging. We should take care not to ‘blame’ dementia on poor sleep or create unnecessary worry. Efforts to control sleep often backfire, in what is known in behavioural sleep medicine as “the paradox of effort”. Many people who sleep well also develop dementia, and people who sleep poorly may never develop dementia.
Nonetheless, improving sleep remains a promising avenue for supporting cognitive health and brain function. There are approved treatments, including talking therapies such as CBT-I and medications, and lifestyle changes that improve sleep for various populations and could be tested in people with or at-risk of dementia. New, targeted sleep interventions could also be developed and evaluated.
Sleep offers a unique opportunity and should be explored thoroughly as an avenue for dementia prevention and treatment.
Sleep offers a promising, low-risk approach for maintaining cognitive health and reducing dementia risk, but care must be taken not to overstate the evidence or create unnecessary worry about sleep.
Statement of interests
Victoria Gabb’s team works across several different sleep and dementia research studies. She is currently supporting colleagues preparing to open a clinical trial looking at whether by improving sleep duration, and specifically duration of slow wave sleep, we might slow progression of Alzheimer’s disease. Her role in the study is as a Patient and Public Involvement lead.
AI has not been used during the writing of this blog.
Editor
Edited by Éimear Foley. AI tools assisted with language refinement and formatting during the editorial phase.
Links
Primary paper
Weng, Bingtao, Chen, Haizhen, Zheng, Ying, Wei, Jiahe, Xue, Pei, Benedict, Christian, Chen, Wei, & Tan, Xiao (2026) Changes in sleep dimensions, cognitive transitions, and incident dementia: a two-cohort longitudinal study. Sleep 2026, 49(1):zsaf233. doi:10.1093/sleep/zsaf233
Other references
Fang W, Le S, Han W. et al (2023) Association between napping and cognitive impairment: A systematic review and meta-analysis. Sleep Med 2023, 111, 146-159. doi: 10.1016/j.sleep.2023.09.022
Ferguson, A. Is short sleep linked to risk of psychosis and could inflammation be a factor? The Mental Elf, 03 Dec 2025.
Gao C, Zheng X, Cai R. et al (2025) Timing and intraindividual variability of daytime napping and Alzheimer’s disease in older adults. Commun Med 2025, 5, 219. doi: 10.1038/s43856-025-00936-1
Larsen, T. The genetic link behind sleep problems, cognitive dysfunction, and neuroticism in ‘treatment-resistant depression’. The Mental Elf, 17 Sep 2025.
Li J, McPhillips MV, Deng Z. et al (2023) Daytime Napping and Cognitive Health in Older Adults: A Systematic Review. J Gerontol 2023, 78 (10), 1853–1860. doi: 10.1093/gerona/glac239
Landry GJ, Best JR, Liu-Ambrose T (2015) Measuring sleep quality in older adults: a comparison using subjective and objective methods. Front Aging Neurosci 2015, 7 (2015). doi: 10.3389/fnagi.2015.00166
Pathmanathan J, Westover MB, Sivakumaran S. et al (2025) The role of sleep in Alzheimer’s disease: a mini review. Front Neurosci 2025, 19 (2025). doi:10.3389/fnins.2025.1428733
Shah AS, Pant MR, Bommasamudram T. et al (2025) Effects of sleep deprivation on physical and mental health outcomes: an umbrella review. Am J Lifestyle Med 2025. Online ahead of print, doi:10.1177/15598276251346752
