Broken-up deep-sleep brain waves tracked with tau buildup and weaker memory, a Berkeley study found

Deep sleep produces slow electrical waves that normally sweep from the front of the brain toward the back, and a University of California, Berkeley study reports that older adults whose waves stop traveling carry more tau protein and remember less the next day. Omer Sharon, a postdoctoral researcher at Berkeley’s Center for Human Sleep Science, led the work, published in Nature Neuroscience (DOI 10.1038/s41593-026-02415-9). The paper calls the isolated events lonely, nontraveling slow waves, a technical label that says nothing about social life.

Stay ahead on AI: Join 600,000+ professionals who read The Deep View, the daily newsletter that keeps you up to date with artificial intelligence.

Subscribe to The Deep View →

We may earn a commission if you sign up.

Slow waves that stop traveling

The study centers on episodic memory, the recall of personal experiences, which tends to weaken with age. Tau is a protein tied to Alzheimer’s disease, and slow waves during non-REM deep sleep normally begin in the frontal cortex and spread across the brain in a coordinated way, a pattern thought to help move the day’s memories into storage. Sharon has described that sweep as one of the most synchronized pieces of electrical activity in a healthy brain. The new work asks what happens to that coordination when tau accumulates in the very region where the waves start.

The team recorded sleeping brain activity with electroencephalography in healthy young adults in their early 20s and in older adults from roughly their mid-60s to mid-70s. In the young group, clusters of slow waves traveled across distances about the width of a hand on the scalp. In the older group, the waves covered shorter distances and behaved more independently, less like a coordinated sweep and more like separate flickers.

The ScienceDaily write-up of the paper, dated Nov. 5, says greater tau buildup in the frontal cortex was associated with a breakdown in the longer traveling waves. The older participants did not have Alzheimer’s disease, and their memory was within the normal range for their age, so the pattern was found in people without a diagnosis.

Nine authors are listed on the paper: Sharon, Xi Chen, Jason Dude, James Westphal, Chelsea Brown, Vyoma D. Shah, Yo-El S. Ju, Jagust and Matthew P. Walker. Walker is the Berkeley sleep scientist under whom Sharon worked, and the funding came from the National Institute on Aging at the National Institutes of Health, along with a Glenn Foundation for Medical Research postdoctoral fellowship and a Weill Neurohub fellowship for Sharon.

Frontal tau, word pairs and spinal fluid

Tau was measured with PET scans in the older participants, an effort in which William Jagust, a Berkeley professor emeritus, took part. Memory was tested with word pairs learned in the evening and recalled the next day. People with more of the shorter, isolated waves generally remembered fewer of those associations. A subset retested years later showed that rising frontal tau went with less coordinated waves and weaker overnight retention.

A separate group of older adults, analyzed with Yo-El Ju of Washington University in St. Louis, gave spinal fluid samples. More isolated waves went with a higher tau-to-amyloid ratio there too. Spinal fluid cannot pinpoint tau in the frontal cortex, so that result points the same direction without matching the PET evidence.

The study also leans on a measure that is hard to pin down. Sleep quality is difficult to define or measure exactly, and the researchers worked around that by analyzing the shape and spread of individual waves instead of relying on total hours or a general score. The tradeoff is that the finding applies to one specific feature of deep sleep, not to sleep as a whole, so it does not support broad claims that more sleep or longer sleep would protect memory.

Association, with the direction still open

An earlier preprint version of the work, covered by Alzforum, described 61 young adults averaging about 20 and 74 adults around 75, all with tau and amyloid PET scans and two overnight sessions. Those counts come from the preprint report and may differ in the published paper. Sharon said at that stage that pinpointing the cause would take more longitudinal data.

The published version keeps the same caution. SciTechDaily’s account, dated Sept. 11 to match the paper’s release, reports Sharon as saying the team cannot yet tell which change comes first, and adds that the answer will shape how interventions are designed. The findings do not establish that tau directly causes memory loss, or that disrupted waves are the mechanism.

The work spread quickly. A Baptist Health roundup on Sept. 25 summarized it for patients, and the newsletter The Careful Read praised the independent second cohort while warning that the word lonely could be misread as evidence about isolation. Its author had checked only the abstract, title and figure captions, so the methods went unreviewed there.

The timing matters for readers meeting the story in November: the paper itself appeared in September, and the Nov. 5 ScienceDaily item is a later write-up of it. Whether tau damages the waves, whether weakened waves let tau accumulate, or whether a third process drives both is the question the authors leave unresolved.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

Stay ahead on AI: Subscribe to The Deep View →

We may earn a commission if you sign up.