Broken sleep predicts amyloid buildup in the brain, autopsy study finds

Researchers have long suspected that poor sleep and Alzheimer's disease feed each other in a vicious cycle, but separating cause from consequence has proven stubbornly difficult. A new study published June 25, 2026 in Alzheimer's & Dementia offers some of the most direct evidence yet: objectively measured sleep fragmentation correlates with amyloid-beta density in brain tissue confirmed at autopsy, with the relationship concentrated in people who were still in the early stages of disease.

Sleep fragmentation predicts amyloid-beta density in post-mortem brain tissue

The core finding comes from Rudolf MA, Ng YT, and Lieberman AP, who analyzed data from 798 participants in the Rush Memory and Aging Project (MAP), a long-running cohort of older Chicago-area adults. Participants wore actigraphy devices for 7 to 10 days each year, and their brains were donated for neuropathologic examination after death. The team quantified sleep fragmentation using a metric called kRA, an actigraphy-derived probability that a person transitions from rest to activity during a given interval. A higher kRA means more fragmented, less consolidated sleep.

The result: a 1 standard deviation increase in kRA corresponded to a 0.09 unit increase in amyloid density, and that association held up after adjusting for age at death, sex, education, and APOE genotype. For the average participant, a 1 SD rise in sleep fragmentation translated to an 18% relative increase in the cortical area testing positive for amyloid-beta immunostaining. The growing body of work on the bidirectional link between sleep and amyloid-beta has long suggested both directions of influence, and this study adds hard neuropathologic confirmation to that picture.

Why the finding matters for early Alzheimer's detection

The association was not uniform across the full disease spectrum. Statistical significance appeared only in participants with low AD neuropathologic change or no to mild cognitive impairment, meaning people who had not yet accumulated heavy amyloid burden or severe cognitive decline. In those with more advanced disease, the signal faded.

That pattern carries practical implications. Research on sleep abnormalities and Alzheimer's risk has increasingly emphasized that intervening before pathology becomes entrenched may be the most effective window. If sleep fragmentation is most tightly coupled to amyloid accumulation in the early phase, it could serve as both a warning sign and a modifiable target before the disease process becomes self-sustaining. Amyloid-beta may itself disrupt sleep through neuronal hyperexcitability and impaired slow-wave sleep, while disrupted sleep in turn impairs the glymphatic system's overnight clearance of amyloid from brain tissue, a mechanism that even a single night of sleep deprivation can stress, raising amyloid-beta levels measurably.

How the Rush MAP cohort made stronger evidence possible

Previous studies linking poor sleep to Alzheimer's pathology have faced two persistent problems: reliance on self-reported sleep quality, and reliance on biomarkers like PET imaging or cerebrospinal fluid assays that can miss early-stage pathology. The Rush MAP approach addressed both.

Participants underwent annual actigraphy alongside cognitive assessments and sleep questionnaires over years, with a mean age at enrollment of 81.72 years and mean age at death of 91.70 years. Post-mortem neuropathologic evaluation measured five quantitative outcomes: amyloid density, neurofibrillary tangle (NFT) immunohistochemical density, diffuse plaque count, neuritic plaque count, and NFT silver-stain count. Autopsy has higher sensitivity for early-stage amyloid deposition than imaging or fluid biomarkers, making this a genuine methodological upgrade. Earlier cohort work using actigraphy to measure sleep fragmentation in 737 older adults found elevated Alzheimer's risk with higher fragmentation, but lacked the neuropathologic confirmation that autopsy provides.

The timing of the sleep measurements also mattered. The kRA value from the year closest to death showed a stronger association with amyloid density than either the mean across all years or the first-year measurement. That proximal reading also associated positively with neuritic NFT density and neuritic plaque count, suggesting the relationship between sleep and pathology intensifies as disease approaches.

Subjective sleep reports told the opposite story

One of the more striking results in the paper is what subjective sleep quality did not show. Poorer scores on a modified Pittsburgh Sleep Quality Index (PSQI), a self-report questionnaire, were paradoxically linked to lower amyloid density and lower counts on several other neuropathologic measures, not higher ones.

This is not entirely surprising to researchers in the field. People with advancing Alzheimer's disease tend to underreport sleep disturbances, a form of recall bias rooted in the cognitive impairment itself. The small negative correlation between PSQI scores and kRA in this dataset suggests the two measures are capturing genuinely different things. Objective actigraphy picks up fragmentation the sleeper may not consciously register; questionnaires capture what the sleeper believes about their sleep. The divergence is a methodological caution for any study that relies solely on self-report to characterize sleep in an aging or cognitively impaired population.

Open questions about causality and who the findings apply to

The authors are careful to note that observational data cannot determine the direction of effect. Sleep fragmentation could be promoting amyloid accumulation, amyloid could be disrupting sleep architecture, or both could be true simultaneously and in a stage-dependent way. The finding that the association is strongest in early disease is consistent with a model where fragmentation is most relevant when pathologic changes are just emerging, but the data cannot confirm it.

Several other caveats limit interpretation. The cohort was predominantly White and drawn from one metropolitan area, which constrains generalizability. Information on chronic sleep medication use and the severity of medical conditions was unavailable. The time interval between the last actigraphy recording and death was not tracked. The kRA metric captures rest-activity disruption broadly and may not isolate sleep fragmentation from other sources of nocturnal movement.

Despite those constraints, the study's combination of longitudinal actigraphy, large sample size, and gold-standard autopsy neuropathology represents a significant step beyond what earlier work could achieve. The practical message for future research is pointed: if sleep fragmentation is most consequential in the pre-dementia window, clinical trials testing sleep interventions in cognitively normal or mildly impaired older adults may have the best chance of showing an effect on downstream amyloid burden.

Evidence rating

EMERGING — 73/100

Caveats:

• Observational design: the study cannot determine whether sleep fragmentation causes amyloid accumulation, results from it, or both.
• Cohort was predominantly older White adults from Chicago (mean age at death ~92), limiting generalizability to younger or more diverse populations.
• Funding source was not disclosed by the authors.
• Subjective sleep quality showed the opposite pattern to objective measures, suggesting self-reported sleep data may be unreliable in this population.

Longitudinal observational cohort (n=798) using actigraphy and post-mortem neuropathologic evaluation. Demonstrates an association between objective sleep fragmentation and amyloid-beta density but cannot establish causal direction. Findings are consistent with prior work but have not been independently replicated using this gold-standard autopsy methodology.

References

1. Mark A. Rudolf, Yee To Ng, Andrew P. Lieberman (2026). Sleep fragmentation correlates with amyloid beta deposition at brain autopsy. Alzheimer's & Dementia. https://doi.org/10.1002/alz.71616
2. Fuxin Ren, Jianfen Luo, Wen Ma, et al. (2019). Hearing Loss and Cognition Among Older Adults in a Han Chinese Cohort. Frontiers in Neuroscience. https://doi.org/10.3389/fnins.2019.00632
3. Sleep Abnormalities and Risk of Alzheimer's Disease – PMC. NIH. https://pmc.ncbi.nlm.nih.gov/articles/PMC8909181/
4. Lack of sleep boosts levels of Alzheimer's proteins – WashU Medicine. medicine.wustl.edu. https://medicine.wustl.edu/news/lack-of-sleep-boosts-levels-of-alzheimers-proteins/
5. Sleep Fragmentation and the Risk of Incident Alzheimer's Disease and Cognitive Decline in Older Persons – PMC. NIH. https://pmc.ncbi.nlm.nih.gov/articles/PMC3711956/