I found a study last week that I want to share with you. It looks at what happens inside the adult brain during deep sleep and it connects to something we deal with all the time – the sound machines.

The study comes from MIT. Researchers played very short bursts of pink noise to sleeping adults. Each burst lasted 50 ms and was timed to a specific phase of the brain’s slow waves. The slow waves increased in amplitude when the timing was right. Earlier studies had already shown that sound timed to slow waves can do this (Ngo et al., 2013). What this study added is that the waves of cerebrospinal fluid moving in and out of the brain increased in amplitude as well.

Why is that interesting? Cerebrospinal fluid is the clear liquid that surrounds the brain and spinal cord. During sleep cerebrospinal fluid helps transport waste from the brain, including proteins linked to Alzheimer’s disease. The MIT team now wants to test whether making these fluid waves larger can improve cognition, help people with insomnia get more restorative sleep and slow cognitive decline in normal ageing and in the early stages of Alzheimer’s. None of that has been tested yet. But if it works, this would be a non-invasive way to support the brain’s own cleaning process using nothing more than a quiet sound at the right moment.

Also parents will see the headlines and some will ask whether the white noise machine in the nursery does the same thing. The answer is no and I would explain why.

This whole topic needs some background to make sense…

Why scientists think sleep cleans the brain

The idea that sleep helps “clean” the brain gained attention after a mouse study published in Science by researchers at the University of Rochester. Xie et al. (2013) found that the interstitial space (the tiny, fluid-filled spaces between brain cells) was around 60% larger during sleep than during waking. This gave cerebrospinal fluid more room to exchange with interstitial fluid, the fluid surrounding the cells.

The researchers measured waste clearance too – how quickly substances were removed from brain tissue. One substance they tracked was beta-amyloid, a protein that can accumulate in Alzheimer’s disease. It cleared faster in sleeping mice than in awake mice. This finding offered a possible explanation for one of sleep’s functions. Sleep might change the conditions inside the brain in ways that help it remove waste.

Then came a result that challenged this explanation.

In another mouse study, published in Nature Neuroscience, Miao et al. (2024) used a fluorescent tracer ( this is like a dye that researchers can track) to investigate movement and clearance within the brain. They reported that the tracer cleared more slowly during sleep than during waking.

The disagreement goes beyond two studies reporting different results. It concerns how scientists measure clearance itself. Fluid entering brain tissue, a tracer moving between brain regions and a substance leaving the brain are different events. Researchers disagree about how well particular experiments distinguish them. That dispute led to a published exchange in 2025. Plá et al. challenged Miao’s methods and interpretation, including whether the measurements adequately separated tracer redistribution from removal. Franks and Wisden rejected those criticisms and defended their finding of reduced clearance during sleep.

So there is evidence supporting sleep-related waste clearance, but the mechanism and its measurement remain disputed. For the MIT sound study, this distinction is important because making cerebrospinal fluid waves larger is a measurable finding. Showing that those larger waves remove more waste requires a separate test.

The first fluid waves seen in a sleeping human brain

Several years after the mouse study, Laura Lewis and her team studied cerebrospinal fluid during human sleep. In Fultz et al. (2019), participants fell asleep inside an MRI scanner wearing EEG electrodes, which recorded the brain’s electrical activity. The MRI tracked blood-related changes and cerebrospinal fluid movement.

During non-REM sleep, the researchers detected large, rhythmic waves of CSF inflow at the fourth ventricle, a fluid-filled cavity in the brain. These waves followed a striking sequence – first came slow electrical waves, then changes in the blood-oxygenation signal, then a wave of incoming CSF.

The study showed that the sleeping brain’s electrical rhythms and fluid movements were closely linked. It raised a question that led to the new MIT study – if slow electrical waves come before CSF waves, could making the electrical waves larger make the fluid waves larger too?

What the MIT team did with 14 sleeping adults

Levitt et al. (2026) tested exactly that. The paper was published in Science Translational Medicine with Levitt as first author and Lewis as senior author.

The technique is called closed-loop auditory stimulation. It uses EEG to follow the brain’s electrical activity and plays a short sound at a particular point in a slow wave. Earlier research had shown that this could make slow waves larger and improve memory (Ngo et al., 2013). The new question was whether the fluid waves would grow too.

They had 14 healthy adults take afternoon naps in an MRI scanner with EEG on. The hard part was technical. The magnetic fields in an MRI scanner interfere with the EEG signal. The team built a system that cleaned the EEG in under 100 ms and then predicted when the next slow wave peak would occur so they could play the sound at the right moment. The sound was a 50 ms burst of pink noise. If you have never heard pink noise it is a soft hiss like steady rain. Same idea as white noise but with more energy in the lower frequencies and less in the higher ones.

The sound was not loud enough to wake a sleeping person. Both the electrical slow waves and the cerebrospinal fluid waves increased in amplitude – meaning they became larger. The fMRI data showed the link between the two. Slow waves cause blood vessels in the brain to constrict and dilate and this acts as a pump that moves cerebrospinal fluid.

Lewis has said that as far as her team knows there was no method to increase these fluid waves during sleep before this.

Why the Alzheimer’s headlines are ahead of the data

This was 14 people taking naps. It was a proof of concept. Whether the method improves memory or thinking or changes amyloid levels in the brain has not been tested. So we do not know yet whether larger fluid waves do anything useful for the person. The team plans to test it in older adults and in people with early signs of neurological disease but that has not happened.

The headlines about fighting Alzheimer’s go further than the data. What we actually have is evidence that a tiny well timed sound can change fluid flow in the sleeping brain. Whether it keeps anyone healthy is a separate question that nobody has answered.

Levitt has started a company that hopes to develop a device, possibly a headband, that would do this at home. I mention it so you know it exists. There is no evidence yet that such a device would improve anyone’s health.

What this means for sound machines

The MIT method depends on timing. One short burst at the right phase of the slow wave. A sound machine in a nursery plays continuous sound with no timing at all. The researchers used live EEG and a predictive algorithm to time the sounds. They used MRI to measure the resulting fluid movements. A sound machine on the nightstand does none of that. So if a parent asks whether their white noise machine is cleaning their child’s brain – there is no evidence for that. The two are different things.

What we know from research about sound machines is about volume and distance. Hugh et al. (2014) measured 14 infant sleep machines at maximum volume and found that every one of them exceeded 50 dBA at 30 cm, which is the recommended noise limit for infants in hospital nurseries. Their advice was simple. Place the machine as far from the baby as possible, keep the volume as low as possible and use it for as short a time as needed. Used that way sound machines remain a widely used tool for settling. The MIT study does not change that advice or provide new evidence about the benefits or safety of sound machines for babies.

References

Fultz, N. E., Bonmassar, G., Setsompop, K., Stickgold, R. A., Rosen, B. R., Polimeni, J. R., & Lewis, L. D. (2019). Coupled electrophysiological, hemodynamic, and cerebrospinal fluid oscillations in human sleep. Science, 366(6465), 628–631. https://doi.org/10.1126/science.aax5440

Hugh, S. C., Wolter, N. E., Propst, E. J., Gordon, K. A., Cushing, S. L., & Papsin, B. C. (2014). Infant sleep machines and hazardous sound pressure levels. Pediatrics, 133(4), 677–681. https://doi.org/10.1542/peds.2013-3617

Levitt, J., et al. (2026). Closed-loop auditory stimulation in phase with slow waves during sleep enhances cerebrospinal fluid flow in humans. Science Translational Medicine. https://doi.org/10.1126/scitranslmed.aea5469

Miao, A., Luo, T., Hsieh, B., Edge, C. J., Gridley, M., Wong, R. T. C., Constandinou, T. G., Wisden, W., & Franks, N. P. (2024). Brain clearance is reduced during sleep and anesthesia. Nature Neuroscience. https://doi.org/10.1038/s41593-024-01638-y

Ngo, H.-V. V., Martinetz, T., Born, J., & Mölle, M. (2013). Auditory closed-loop stimulation of the sleep slow oscillation enhances memory. Neuron, 78(3), 545–553. https://doi.org/10.1016/j.neuron.2013.03.006

Xie, L., Kang, H., Xu, Q., Chen, M. J., Liao, Y., Thiyagarajan, M., O’Donnell, J., Christensen, D. J., Nicholson, C., Iliff, J. J., Takano, T., Deane, R., & Nedergaard, M. (2013). Sleep drives metabolite clearance from the adult brain. Science, 342(6156), 373–377. https://doi.org/10.1126/science.1241224