A study of 2,410 adults found that regularly sleeping nine hours or more raises blood levels of the main protein used to detect early Alzheimer’s. The curve spikes sharply past ten hours

The relationship between sleep and Alzheimer’s disease has become one of the clearest stories in brain health research over the past decade. Not sleeping enough lets the brain’s waste-clearance system fall behind. Amyloid and tau proteins, the molecular hallmarks of Alzheimer’s, accumulate when sleep is insufficient. Studies consistently show that people who sleep less than six or seven hours a night carry higher dementia risk over time. The advice that followed was intuitive: protect your brain by sleeping enough.

A new study published in Alzheimer’s & Dementia, the journal of the Alzheimer’s Association, has examined that relationship at higher resolution using a well-established population dataset, and the picture it reveals is more complicated than the standard guidance suggests.

Researchers at UT Health San Antonio analyzed data from 2,410 older adults enrolled in the Framingham Heart Study, one of the longest-running cardiovascular and neurological cohort studies in the United States. Rather than treating sleep as a single variable with a linear effect on brain health, the team used restricted cubic spline modeling, a statistical technique specifically designed to detect curves and thresholds that simple averages would miss.

They were looking for the shape of the relationship between sleep duration and p-tau181, a phosphorylated form of tau protein now detectable in blood that has emerged as one of the most reliable early biomarkers of Alzheimer’s-related pathology. Unlike amyloid scans or spinal fluid tests, p-tau181 measured from a routine blood draw is becoming a practical tool for detecting early disease processes, sometimes years before symptoms appear.

The shape they found was not what the standard sleep-and-dementia narrative would predict.

Blood levels of p-tau181 were lowest in participants who reported sleeping seven to eight hours per night. The association began rising at eight and a half hours. Beyond ten hours of regular sleep, the curve spiked sharply upward. And short sleep, six hours or less per night, showed no significant association with elevated p-tau181 at all.

What p-tau181 actually measures

Understanding why this finding matters requires understanding what p-tau181 is and what elevated levels mean.

Tau is a protein that normally stabilizes the internal skeleton of neurons. In Alzheimer’s disease, tau becomes abnormally phosphorylated, meaning chemical tags are added to it in specific locations that cause it to detach from the neuronal skeleton, aggregate into tangles, and spread through the brain in a pattern that correlates closely with cognitive decline. The degree of phosphorylation at the 181 position on the tau molecule has emerged as a particularly specific indicator of Alzheimer’s pathology.

Until recently, measuring tau required either a lumbar puncture, which is invasive, or a specialized PET brain scan, which is expensive and not widely available. The development of blood-based p-tau181 assays has changed that. Blood levels of p-tau181 now appear in the Framingham Heart Study data and others as an accessible early signal that Alzheimer’s-related biological processes may be underway, potentially years before any cognitive symptom becomes apparent.

When the researchers found that long sleep duration was specifically associated with elevated p-tau181, they then tested whether that association might simply reflect other factors. They tracked four blood-based brain health biomarkers simultaneously: p-tau181, GFAP (a marker of astrocyte activation), NfL (a marker of neuronal damage), and Aβ42/40 (a ratio measuring amyloid accumulation). All four were examined against sleep duration.

When the analysis controlled for kidney function, the associations for GFAP, NfL, and Aβ42/40 largely disappeared. Kidney function turned out to explain most of the apparent sleep-related signal in those proteins. But p-tau181 held. The association between long sleep and elevated p-tau181 survived adjustment for kidney function, suggesting it reflects something specific to Alzheimer’s-related biology rather than general health status.

Why long sleep might signal early neurodegeneration rather than cause it

The most important interpretive point in the study is the direction of causality, which the data cannot establish.

The researchers are explicit that sleeping long hours probably does not cause Alzheimer’s. The more plausible interpretation runs the other way: early neurodegeneration, occurring silently before any cognitive symptoms appear, may itself alter sleep architecture in ways that increase total time in bed or time asleep. Disrupted sleep quality, fatigue from subclinical brain changes, or altered circadian regulation driven by early disease processes may all produce longer reported sleep duration as a downstream effect rather than as a cause.

“A lot of people worry about whether their sleep habits are affecting their brain health,” said lead author Vanessa M. Young, a postdoctoral research fellow at the Glenn Biggs Institute for Alzheimer’s and Neurodegenerative Diseases at UT Health San Antonio. “In plain terms, if you regularly find yourself sleeping nine to ten hours or more a night, it may be worth mentioning to your doctor as a useful conversation starter about your sleep quality and overall brain health.”

This interpretation, that long sleep is a behavioral marker of early disease rather than a driver of it, has precedent. Prior research has documented that people in the preclinical phase of Alzheimer’s, before any measurable cognitive decline, frequently report changes in their sleep, including both disrupted nighttime sleep quality and increased total sleep time. A person sleeping ten hours who was previously comfortable on seven may be experiencing early disease-related fatigue without knowing it.

What the short-sleep finding means

The absence of an association between short sleep and elevated p-tau181 in this dataset is worth examining directly, because it appears to contradict the most commonly cited evidence linking insufficient sleep to Alzheimer’s risk.

The researchers do not interpret this as evidence that short sleep is harmless. The Framingham Heart Study population is older, and older adults who habitually sleep less than six hours and have remained cognitively intact may represent a selected group of people whose biology is more resilient to sleep restriction. Additionally, the study measured p-tau181 at a single time point, meaning it cannot capture whether short-term sleep deprivation over weeks or years eventually produces the same signal that showed up on the long end of the duration spectrum.

The finding also does not contradict the glymphatic hypothesis, the biological mechanism by which sleep helps clear amyloid and tau from the brain during the night. That mechanism operates during sleep and would predict that some amount of sleep is protective. The question this study raises is whether the benefit plateaus or even reverses past a certain duration threshold, or whether reported long sleep duration at a single time point is simply a proxy for poor sleep quality rather than a measure of actual restorative sleep time.

What the 8.5-hour inflection point means practically

The study identifies 8.5 hours as the approximate point where p-tau181 levels begin to rise, with the sharpest increase appearing above 10 hours. This is not a clinical diagnostic threshold. It is a statistical pattern across a population, and individuals vary enormously.

The practical message from the researchers is not that everyone who sleeps nine hours is developing Alzheimer’s. It is that habitual long sleep, particularly when it represents a change from a person’s normal pattern, is worth discussing with a physician. It may reflect nothing more than a recovery phase after illness or unusual exertion. But in an older adult who has gradually shifted toward sleeping longer without explanation, it may be worth checking against other clinical signs of early cognitive change.

The study is observational and relies on self-reported sleep duration, which is subject to both inaccuracy and confounding. A person who spends ten hours in bed but sleeps poorly would report long sleep duration while actually experiencing significant sleep deprivation at the level of restorative sleep. The study cannot separate time in bed from actual sleep time, which limits how precisely the 8.5-hour threshold should be interpreted.

What it adds to the field is a specific, rigorous look at the non-linear shape of the sleep-Alzheimer’s relationship using one of the most validated brain health biomarkers now available. The story is not simply more sleep equals better brain health. The relationship curves, and the curve matters.

The study “Non-linear associations between sleep duration and plasma p-tau181 in the Framingham Heart Study” was authored by Vanessa M. Young, Crystal Wiedner, Andrée-Ann Baril, Matthew P. Pase, Agustin Ruiz, Arash Salardini, Christopher R. Frei, Tiffany Kautz, Rebecca Bernal, Stephanie Yiallourou, Lachlan Cribb, Alexa Beiser, Antonio L. Teixeira, Jayandra Jung Himali, and Sudha Seshadri at UT Health San Antonio and collaborating institutions. Published May 19, 2026 in Alzheimer’s & Dementia.

Source: UT Health San Antonio / Glenn Biggs Institute for Alzheimer’s and Neurodegenerative Diseases. DOI: 10.1002/alz.71499