Carbon dioxide leaves the body with every breath, and small amounts are always present in the air. Now, a preliminary 12-person experiment suggests that carefully timed pulses of a much higher concentration may help move Alzheimer’s-linked proteins out of the brain and into the bloodstream.
That does not make carbon dioxide a cure, and it certainly does not make inhaling the gas at home safe. The finding is an early clue that doctors may one day be able to activate the brain’s natural cleanup system while a patient is awake.
The brain has a cleanup network
The brain produces metabolic leftovers as its cells work, much like a busy kitchen accumulates scraps during dinner. A fluid pathway known as the glymphatic system moves cerebrospinal fluid around blood vessels and through brain tissue, helping carry some of that material away.
This pathway becomes especially active during deep sleep. Sephira Ryman, a neuropsychologist leading the work at the University of New Mexico, said that “respiration slows and the brain exhibits an oscillatory pattern of activity called slow waves that drives the glymphatic pathway to clear the brain.”
Aging, poor sleep and neurodegenerative disease may make this process less efficient, although scientists are still working out exactly how glymphatic flow operates in people.
That uncertainty matters because beta-amyloid and abnormal tau are central hallmarks of Alzheimer’s, but they are not the disease’s only moving parts.
What the 30-minute test involved
The experiment included eight people with normal cognitive function and four with elevated tau markers. Three people in the second group had mild cognitive impairment, a condition involving noticeable memory or thinking problems that do not yet fully prevent everyday independence.
Each participant wore a medical mask during a half-hour session. The system alternated every 35 seconds between ordinary air containing about 0.05% carbon dioxide and a medical gas mixture containing 5%, roughly one hundred times as much.
Why switch back and forth? Carbon dioxide temporarily widens blood vessels in the brain, while returning to normal air lets them move back toward their usual size. That repeating expansion and contraction, called vasomotion, acts like a pump and pushes cerebrospinal fluid through the brain’s drainage routes.
MRI scans captured the pumping effect
Magnetic resonance imaging showed that the alternating gas produced rhythmic vascular changes similar to patterns seen during deep sleep. The important feature may be the rhythm, not simply exposure to carbon dioxide.
The researchers drew blood before the session and again 15 minutes afterward. Amyloid beta rose in the bloodstream after the intervention, while tau clearance appeared strongest among participants who already had elevated tau or cognitive difficulties.
At first glance, more Alzheimer’s-linked protein in blood may sound like bad news. In this setting, the team interpreted the rise as a sign that material had moved away from brain tissue and entered circulation for later removal, although a blood increase alone cannot prove a lasting reduction inside the brain.
A dermatologist performs a laser facial treatment, a procedure commonly used to improve skin texture, pigmentation, and signs of aging.
Earlier research points beyond Alzheimer’s
A related peer-reviewed study used three roughly 10-minute sessions of intermittent carbon dioxide in people with Parkinson’s disease and healthy older adults.
MRI measurements showed increased cerebrospinal fluid inflow, and later blood samples contained higher levels of alpha-synuclein, amyloid beta, tau and other brain-related markers.
That earlier result supports the idea that the technique may not target only one protein. Jeff Iliff, a researcher at the University of Washington, has noted that a broader boost to waste clearance could also matter for alpha-synuclein, which forms damaging clusters in Parkinson’s disease.
But broader is not the same as proven. A treatment that briefly changes biomarkers must still show that it reduces harmful deposits, protects neurons or slows the memory loss that patients and families experience.
This is not a home experiment
No serious side effects were reported during the short, closely monitored test. Even so, 5% carbon dioxide equals 50,000 parts per million, a level that can cause intoxication during continuous exposure and sits above the concentration the National Institute for Occupational Safety and Health classifies as immediately dangerous to life or health.
High carbon dioxide can trigger headache, dizziness, breathing difficulty, a racing heart, loss of consciousness and asphyxiation. The trial used brief pulses, normal-air breaks, medical equipment, MRI monitoring and trained staff, so trying to imitate it with a bag, tank, mask or breath-holding routine could be deadly.
The next step is not a prescription. Researchers need larger, controlled trials that test repeated sessions, long-term safety, changes inside the brain and, most importantly, whether the approach preserves memory or slows disease progression.
A promising mechanism with a long road ahead
The study offers a striking idea. Instead of designing a drug that binds to one target, doctors might eventually stimulate the brain’s own fluid movement and help several kinds of waste leave at once.
Still, twelve participants and one short experiment cannot answer the big question. For now, the result is best viewed as a proof of concept that turns a normal response to carbon dioxide into a possible medical tool, not evidence of an Alzheimer’s cure.
The related peer-reviewed work was published in npj Parkinson’s Disease, and the new preliminary findings were presented at the Alzheimer’s Association International Conference 2026.
