Timed pink noise boosts the brain's nighttime fluid waves, MIT study finds
In 14 healthy sleepers, 50-millisecond sound pulses enlarged slow brain waves and the cerebrospinal fluid pulses linked to waste clearance.

Very short bursts of pink noise, played at precisely the right instant during deep sleep, can make the brain's slow electrical waves larger. The waves of cerebrospinal fluid that follow them grow larger as well, according to new research from MIT.
The study, published in Science Translational Medicine, involved 14 healthy volunteers. Scientists tracked each sleeper's brain electrical activity and fluid movement at the same time, and the sounds were soft enough that no one woke up. The team did not show whether the bigger fluid waves carried more waste out of the brain or left people better rested. Those are the two outcomes that would make the finding matter outside a lab.
Laura Lewis, an MIT associate professor of electrical engineering and computer science, led the work as senior author. Joshua Levitt, who recently finished a PhD at Boston University after a stint as a visiting graduate student in Lewis' lab, was lead author. Lewis said the team found "we were able to increase the size of the CSF flow wave during sleep." She added that, as far as the researchers know, no earlier method had done that.
What fluid flow has to do with sleep
Cerebrospinal fluid, or CSF, is the clear liquid that cushions the brain and spinal cord. It delivers nutrients such as glucose and carries off the byproducts brain cells release as they burn energy. During waking hours, lactic acid and proteins the brain no longer needs build up, and sleep appears to be when much of that material gets cleared.
The idea of sleep as cleanup time took hold through animal research over roughly the past decade, much of it in mice, which suggested the brain flushes waste more efficiently during sleep. In 2019, Lewis' lab brought the question into people. Using functional MRI, the team watched CSF move into and out of the sleeping brain and found that its pulses track slow electrical waves. Those waves dominate deep, non-REM sleep and grow stronger as sleep deepens.
The new scans add a piece of mechanism. Slow electrical waves tracked with blood vessels constricting and dilating, which works like a pump that pushes fluid out of the brain. If slow waves drive the fluid, then strengthening the waves should strengthen the flow. Earlier studies had already shown that well-timed sounds can deepen slow waves.
Lewis compares the approach to pushing a child on a swing: a push at the right point in the arc sends the swing higher, while a mistimed one does little. "The challenge is: How do you find just the right time?" she said.
From an MRI scanner to a possible headband
Each sound burst lasted 50 milliseconds. Pink noise covers the full range of frequencies people can hear, as white noise does, but it emphasizes the lower ones, which gives it a gentler character closer to steady rain or a distant waterfall. The goal was to land each burst at the peak of a sleeper's slow wave.
That turned out to be the hard part. The researchers tracked brain waves with EEG while measuring fluid with fMRI, and the scanner's strong magnetic fields scramble EEG signals. The team built a way to strip that interference out in under 100 milliseconds. It paired that with an algorithm that predicts when the next wave peak will arrive, which offsets the lag created by processing the signal.
That engineering may prove to be the study's most durable contribution, because it gives researchers a tool for testing cause and effect in sleep rather than only observing it. It also shows how much the result depends on timing. Nothing in this work suggests that a pink noise track or a sound machine running all night would have the same effect. The volunteers heard brief pulses keyed to their own brain activity, not a steady background hum.
The researchers next want to learn whether stronger CSF flow makes sleep more restorative, particularly for people with insomnia, and whether it actually improves the brain's ability to clear harmful substances. Levitt tied that second question to dementia. He said Alzheimer's disease and related conditions are caused in part by the buildup of amyloid and tau, and that better waste clearance might help prevent the plaques that lead to disease.
That is a long chain of ifs. Bigger fluid waves would need to remove more waste. More waste removal would need to lower amyloid and tau. Lower levels would then need to protect memory. This study established only the starting point, that timed sound can enlarge the fluid waves, and only in 14 healthy adults lying in a scanner. Treatments aimed at amyloid have also produced mixed results in drug trials, so even a clean clearance effect would not guarantee a clinical benefit.
Levitt has launched a company aiming to build a home device, possibly a headband, that would time sounds to a sleeper's brain waves to increase CSF flow. A headband could plausibly read EEG, but it would have no fMRI to confirm that fluid was moving. Any product will need its own evidence that it improves sleep or brain health, not just that it changes a brain-wave reading. With the lead author's company in the picture, independent replication in larger and more varied groups will carry extra weight.
Funding came from the National Institutes of Health along with a mix of private foundations, individual scholar awards and fellowships, and an internal MIT research fund, among them a Panasonic-backed fellowship focused on AI and wellness.
For now, the experiment shows that researchers can deliberately turn up a fluid rhythm that until recently they could only watch. Whether turning it up helps a tired or aging brain is the question the next studies will have to answer.




