Your brain is sealed inside the skull, consumes enormous energy and produces metabolic by-products around the clock. It has no ordinary lymphatic plumbing threaded through its tissue, so keeping its chemical environment stable depends partly on the movement of cerebrospinal fluid, or CSF, and interstitial fluid around brain cells.
Sleep has become closely associated with this maintenance system. During non-REM sleep, slow neural, vascular and CSF rhythms become tightly coordinated, creating conditions thought to help move soluble waste, including amyloid-beta and tau, the proteins most closely associated with Alzheimer’s pathology.
A 2025 Vanderbilt-led experiment found that focused-attention meditation altered several CSF signals in a direction the researchers compared with sleep, even though participants remained awake. The result is striking physiology, but it is not evidence that meditation prevents or treats Alzheimer’s disease, and it is not medical advice. The study measured fluid dynamics, not protein clearance, dementia risk or clinical outcomes.
Why scientists think sleep helps clean the brain
The most familiar model is the glymphatic system. In that account, CSF travels along spaces surrounding arteries, exchanges with interstitial fluid inside brain tissue, and helps carry dissolved material toward routes that ultimately drain outside the brain. Astrocytes and their aquaporin-4 water channels are thought to help organise the exchange.
A landmark 2013 study in mice reported that sleep or anesthesia increased the space between brain cells and accelerated the removal of metabolites, including amyloid-beta. The finding gave a compelling physical explanation for one restorative function of sleep.
Human evidence is less direct because researchers cannot routinely inject tracers into healthy brains. In 2019, an MRI and EEG study found large CSF waves during non-REM sleep. Slow electrical waves were followed by changes in cerebral blood volume and then by CSF movement. The sequence suggested that sleep synchronises neural and vascular events into a pump-like rhythm.
The Vanderbilt study used three small groups
The peer-reviewed study in the Proceedings of the National Academy of Sciences was led by Bryce Keating, David Vago, Manus Donahue and colleagues. It combined phase-contrast MRI, which quantified movement through the narrow cerebral aqueduct, with blood-oxygenation-level-dependent MRI near the skull base.
The central group contained 23 adept meditators. Each first spent a scanning block in passive mind wandering and then performed focused-attention meditation, repeatedly returning attention to a chosen object while disengaging from distractions. The researchers also enrolled 13 meditation-naive adults for repeat mind-wandering scans and 14 meditation-naive adults for a slowed-breathing control.
Those controls addressed two obvious alternatives. A second scan could differ simply because time had passed, and meditation usually slows respiration. Breathing can mechanically influence venous return and CSF movement, so a meditation comparison without a respiratory control would be difficult to interpret.
Less backflow was interpreted as more efficient motion
During focused attention, the experienced meditators’ average breathing rate fell from 13.0 to 10.4 breaths per minute. Their average heart rate fell from 73.0 to 67.8 beats per minute. Absolute CSF motion through the cerebral aqueduct also declined, from 4.60 to 4.17 millilitres per minute.
At first, less total motion sounds like poorer circulation. The crucial detail is that the reduction came mainly from less regurgitant flow, meaning less CSF reversed direction during the diastolic phase of the heartbeat. Aging and some neurodegenerative conditions are associated with more hyperdynamic, back-and-forth aqueduct flow. The Vanderbilt team therefore interpreted the meditation pattern as directionally opposite to those less efficient dynamics.
A second signal moved the other way. Near the skull base and cervicomedullary junction, low-frequency CSF fluctuations became stronger during focused attention, especially between 0.0614 and 0.0887 hertz, or about 3.7 to 5.3 cycles per minute. These fluctuations were inversely related to grey-matter blood signals. As one rose, the other tended to fall, resembling the alternating blood-and-CSF pattern seen in sleep.
Slower breathing did not reproduce the full result
The meditation-naive breathing group deliberately reduced its respiratory rate while remaining in a mind-wandering state. That manipulation did not produce a significant change in absolute aqueduct CSF motion. The repeatability group, which completed the same passive state twice, did not show the meditation-associated change either.
These comparisons make it less likely that the central result was merely a second-scan effect or a mechanical consequence of slower breathing. They do not identify a single cause. Focused attention simultaneously changes arousal, autonomic balance, neural oscillations, vascular tone and subjective mental state. Any combination of those processes could contribute to the CSF pattern.
The order also matters. The experienced meditators completed mind wandering first and focused attention second, rather than receiving the conditions in a counterbalanced random order. The repeatability controls reduce concern about simple time effects, but they cannot remove every order-related explanation.
CSF dynamics are not the same as waste removal
The experiment did not measure amyloid-beta, tau or any other waste molecule leaving the brain. It did not use a tracer to calculate clearance through tissue. It measured motion and signal fluctuations in specific CSF compartments, then compared their direction with patterns previously reported in sleep, aging and neurodegeneration.
That distinction is especially important because the glymphatic field remains contested. A 2024 Nature Neuroscience study in male mice found that fluorescent-molecule clearance was reduced, not increased, during sleep and anesthesia. The authors argued that movement through tissue and removal from tissue must not be treated as the same process.
Other evidence points in the opposite direction. A 2025 human study using an implanted wireless sensor found overnight sleep-related changes in brain parenchymal resistance consistent with increased glymphatic exchange. Different tracers, species, brain regions, anesthetics and definitions of clearance may explain part of the disagreement. The broad conclusion is that brain-fluid physiology changes with sleep; exactly how that translates into net waste removal is still being resolved.
A waking model, not a substitute for sleep
The Vanderbilt study provides the first direct evidence that focused-attention meditation can modulate human CSF dynamics while people remain awake. Its strongest features are the use of two imaging approaches and control groups designed to test repeat scanning and slower respiration.
Its limits are equally clear. The decisive meditation group contained only 23 experienced practitioners. The design cannot tell us whether beginners would show the same response, how long the response lasts, whether other meditation styles work similarly, or whether repeated practice changes brain health. It also cannot show that meditation removes Alzheimer’s-linked proteins.
Science Blog has previously reported direct imaging evidence of waste-clearance pathways in the living human brain. The new work adds a different layer: a deliberately maintained waking state may tune part of the fluid machinery that researchers usually study during sleep.
That makes focused attention a useful experimental model. Scientists can ask what happens to neurofluid circulation when arousal decreases but consciousness is retained, then separate neural, respiratory and vascular contributions more carefully. Future studies will need larger samples, counterbalanced conditions, direct measures of solute clearance and longitudinal clinical outcomes.
For now, the careful conclusion is fascinating enough. Meditation did not make the awake brain literally sleep, and it did not prove that harmful proteins were washed away. It produced a distinctive pattern of reduced CSF backflow and stronger slow fluid oscillations that resembles part of sleep physiology. That resemblance opens a research path, not an Alzheimer’s prescription.