People whose sleep is more fragmented — waking up frequently throughout the night, as measured by wrist-worn actigraphy — show higher levels of amyloid beta plaque density in their brains at autopsy, according to a new study published in Alzheimer’s & Dementia.
The finding offers some of the strongest evidence yet that objectively measured sleep disruption in living humans is directly tied to the neuropathological hallmark of Alzheimer’s disease, years or even decades before death.
The association, however, came with a twist: it was only detectable in people who were still cognitively normal or mildly impaired and who had relatively low overall Alzheimer’s pathology at autopsy. In those with advanced disease, the link disappeared — suggesting that sleep disruption may be most relevant in the earliest stages of the disease process, when intervention might still matter most.
The Rush MAP cohort
The study drew on data from the Rush Memory and Aging Project, one of the longest-running prospective cohort studies of aging and dementia in the United States. Beginning in 1997, the Rush MAP has enrolled thousands of older adults from retirement communities and subsidized housing facilities in the Chicago area, all of whom agreed to annual clinical evaluations and brain donation after death.
For this analysis, researchers led by Dr. Mary Rudolf and colleagues at the University of Michigan examined 798 participants (mean age 81.7 years at baseline) who had undergone at least one full week of wrist actigraphy annually, completed sleep quality questionnaires, and received cognitive assessments each year. After death, all participants received standardized brain autopsies with quantitative measurement of amyloid beta plaque density and tau neurofibrillary tangle burden.
The design is unusual and powerful. Most studies linking sleep to Alzheimer’s pathology rely on either cross-sectional biomarker measurements (PET scans, CSF assays) in living participants or retrospective sleep reports from caregivers after diagnosis. Very few have the combination of longitudinal actigraphy — an objective measure of sleep-wake patterns — and the gold-standard neuropathological quantification that only autopsy can provide.
Objective vs. subjective sleep
The central finding was that higher actigraphy-measured sleep fragmentation (more frequent nighttime awakenings and transitions between sleep and wake states) was significantly associated with greater amyloid beta density at autopsy, after adjusting for age, sex, education, APOE4 carrier status, and other confounders.
But the same relationship did not hold for subjective sleep quality. In fact, the opposite pattern emerged. Participants who reported poorer subjective sleep quality on standardized questionnaires actually had lower Alzheimer’s-related neuropathologic burden at autopsy — a paradoxical finding that the authors attribute to the well-known mismatch between how people perceive their sleep and how it registers on objective monitoring.
Older adults with early cognitive changes may lose insight into their sleep quality, or subjective reports may capture different aspects of the sleep experience — distress about poor sleep, for example — that bear a different relationship to brain pathology than objective fragmentation.
Recent sleep matters more
The study also revealed a temporal pattern with practical implications. While the analysis examined both mean sleep fragmentation averaged over all available study years and fragmentation measured at the last actigraphy assessment before death, the latter showed a significantly stronger association with amyloid density. Recent sleep disruption appeared to matter more than long-term average disruption.
This could mean that sleep fragmentation accelerates amyloid accumulation relatively quickly, or it could reflect the fact that study participants were oldest (and thus closest to the pathological end stage) at their final actigraphy assessment. Either way, it points to sleep as a dynamic, modifiable factor whose relationship to brain health may change over the course of the disease.
Why it matters
Alzheimer’s disease has no disease-modifying therapies that work for most patients. The amyloid cascade hypothesis has driven drug development for two decades with limited success, in part because by the time cognitive symptoms appear, amyloid deposition is already extensive.
Sleep disruption has emerged as a promising modifiable risk factor. Experimental studies in both animals and humans show that sleep, particularly deep non-REM sleep, facilitates clearance of metabolic waste products from the brain, including amyloid beta, via the glymphatic system. Chronic sleep fragmentation may impair this clearance, allowing amyloid to accumulate. If confirmed, improving sleep continuity — not just total sleep time — could become a preventive strategy.
The current study strengthens this hypothesis by showing that the sleep-amyloid link is specific to objective fragmentation, detectable before advanced pathology sets in, and measurable with widely available actigraphy devices that could be deployed in large-scale screening or intervention trials.
Strengths and limits
The study’s main strengths are its large sample size (798 autopsied brains), long follow-up with repeated actigraphy measurements, and the use of autopsy-based quantitative neuropathology — the definitive standard for measuring Alzheimer’s pathology. Very few studies in the sleep-Alzheimer’s literature have all three of these features simultaneously.
Several limitations should be noted. The Rush MAP cohort is predominantly white, well-educated, and drawn from a single metropolitan area, which limits generalizability. The observational design cannot establish causality — sleep fragmentation could cause amyloid accumulation, amyloid pathology could cause sleep fragmentation, or both could be driven by a shared underlying factor. The study examined amyloid beta but not tau pathology, leaving open the question of whether sleep disruption relates to tangles as well as plaques. Finally, actigraphy is a proxy measure of sleep continuity; it does not capture sleep architecture (time spent in specific stages) or the electrophysiological features thought to be most relevant to glymphatic clearance, such as slow-wave activity.
Bottom line
Objectively measured sleep fragmentation in older adults is associated with higher amyloid beta density at autopsy, but only in those with mild or no cognitive impairment — suggesting a window of opportunity for intervention. The paradoxical absence of this relationship for subjective sleep quality underscores the importance of objective monitoring.
Source
Rudolf MA, Ng YT, Lieberman AP. “Sleep fragmentation correlates with amyloid beta deposition at brain autopsy.” Alzheimer’s & Dementia. 2026 Jul;22(7):e71616. DOI: 10.1002/alz.71616. PMID: 42340807.