The glymphatic system
The glymphatic system is the brain's proposed waste-drainage network: fluid flowing along the outside of blood vessels that is said to wash out proteins such as amyloid-β, the one that clumps in Alzheimer's disease, mostly during deep sleep. It is one of the most quoted reasons to protect deep sleep, and one of the least settled. A 2024 mouse study found brain clearance lower during sleep, and the most direct human test found no difference between a night's sleep and a night awake.
Don't buy a product or adopt a routine on the promise that it boosts overnight "brain cleaning". Protect total sleep, keep a regular schedule and get screened for sleep apnea if you snore; those have outcome evidence whichever way the clearance debate goes.
What it is said to do
The idea is that cerebrospinal fluid, the clear fluid around the brain and spinal cord, flows into the brain along the spaces around arteries, mixes with the fluid between brain cells, and carries waste out. On this account the flow runs hardest during non-rapid-eye-movement (NREM) sleep, and deep, slow-wave sleep in particular, which is why it is so often offered as the reason sleep protects against dementia. The Sleep architecture article covers what deep sleep is and why its measured share of the night is a shaky number.
In mice, the evidence now points both ways
The challenge. A 2024 study in male mice was designed to avoid the flaw critics see in the earlier experiments, which tracked dye flowing into the brain rather than waste leaving it. It found the movement of molecules through brain tissue was the same whether the mice were asleep, awake or anaesthetised, and that clearance was reduced during sleep and under anaesthesia, with deeper slow-wave activity going with less clearance, not more.[1]
The other side. A 2025 study from the group behind the original idea reported that slow, rhythmic pulsing of blood vessels, driven by the brain chemical norepinephrine, drives clearance during NREM sleep. That mechanism could explain part of the 2024 anaesthesia result, since one of the three anaesthetics used, dexmedetomidine, suppresses norepinephrine; it does not concede the sleep half.[2] It relies, though, on the same inflow measurements the 2024 group argues confuse fluid entering the brain with waste leaving it. The question is contested, not settled.
In people, two small studies
A night of disrupted deep sleep. In a crossover trial, 17 healthy adults aged 35–65 had their slow-wave activity selectively broken up with sound through earphones for one night, and spinal fluid was sampled the next morning. Amyloid-β rose about 10% in the responders, a nine-person subgroup, and the effect was tied to the slow-wave disruption rather than to total sleep time.[3] The authors read this as deep sleep reducing amyloid release, not speeding its removal. Tau, the other Alzheimer's protein, did not move; what went with higher tau was poorer home sleep efficiency over the previous six nights, in 14 people. One night, a biomarker, not an outcome.
Sleep versus a night awake. A 2026 randomised crossover trial in 39 healthy adults aged 49–66 measured amyloid-β and tau in the blood before and after a night of normal sleep and a night without sleep. The direct comparison found nothing: the overnight change was statistically indistinguishable between the two nights on every one of the ten comparisons reported, for tau as well as amyloid.[4] The paper's positive headline is a model result: feeding the device's readings, chiefly a measure of electrical resistance in brain tissue plus brainwave (electroencephalogram, EEG) and cardiovascular signals, into a statistical model does predict higher morning levels, and the authors read the null comparison as clearance and production cancelling out. Hold it lightly. The readings came from an investigational in-ear device rather than established instruments, and the trial was funded by the company that makes the device, which employs the first author and three co-authors, with two further authors at the company that sells the blood tests; the senior author, one of the scientists who first proposed the glymphatic system, also declares a financial interest.
What hasn't been shown
No study has yet shown that boosting glymphatic flow, by any means, lowers the risk of dementia. The observational link between losing deep sleep and dementia is itself in doubt: a 2025 analysis of five American studies, 4,657 people and 998 dementia cases, found no association between how much deep sleep people had on one night, or any other measure of sleep structure, and later dementia.[5] It did not test the decline over years that the best-known study linked to dementia.
What to do
- Treat "sleep cleans your brain" as a hypothesis, not a reason to buy a device or supplement that claims to boost it.
- Spend the effort on what has outcome evidence: about seven hours of sleep, a regular schedule, and screening for sleep apnea if you snore. See Sleep architecture.
- Don't read a tracker's deep-sleep score as a clearance score. Consumer devices can't measure deep sleep reliably; see Sleep trackers.
Further reading
- Miao A, et al. Brain clearance is reduced during sleep and anesthesia. Nat Neurosci 2024.[6]
- Hauglund NL, et al. Norepinephrine-mediated slow vasomotion drives glymphatic clearance during sleep. Cell 2025.[7]
- Ju YS, et al. Slow wave sleep disruption increases cerebrospinal fluid amyloid-β levels. Brain 2017.[8]
- Dagum P, et al. The glymphatic system clears amyloid beta and tau from brain to plasma in humans. Nature Communications 2026 — n=39; funded by Applied Cognition, whose device produced the measurements.[9]