The headline, plus Kiviniemi: "The treatment seemed to enhance the glymphatic system's ability to clear beta amyloid from the brain to the blood."

Blood concentrations went up. Brain concentrations were never measured, and no tracer was placed in the cerebrospinal fluid to follow it out. The authors write that the findings are "consistent with transient redistribution of brain derived proteins toward circulation" and should be "interpreted cautiously in light of our currently limited mechanistic knowledge of the solute clearance pathway." They also raise, on their own, a possible transient increase in blood brain barrier permeability.

A rise in blood is evidence that more amyloid arrived or that less left. The headline picked one. The paper did not.

Aβ42 "increased by 35 per cent, on average, after the intervention."

Serum Aβ42 went from 2.6 to 3.5 pg/ml (n = 23, p = 0.03) after correcting for a 5.2 percent blood dilution caused by nitroglycerin. Aβ40 rose by a similar fraction (38.2 to 50.5 pg/ml) and GFAP by about 30 percent. The larger proteins, neurofilament light and pTau181, did not move. The arithmetic checks out. The standard deviations are wide (Aβ42 pre 2.6 ± 1.4), and the dilution correction was estimated in the 12 person NO only subgroup and applied uniformly.

Tested on "the same participants, at least a week before or after," the solo interventions "seemed to have less than half the effect on the glymphatic system than when they were combined."

The comparison is the brain volume showing a T2 dark fluid MRI change, which was 2 to 6 times larger with the combination. The size of the change per voxel was about the same for nitroglycerin alone and the combination. The NO only session ran in 12 people, was exploratory, and was outside the randomization. The vibration alone MRI comparison drew on a partially overlapping cohort.

Blood was drawn only around the combined session. Nobody knows whether nitroglycerin alone would have raised Aβ42 by the same 35 percent, which is the question a reader most wants answered before buying the mattress.

"Brain scans suggest the treatment enhanced the flow of fluid along glymphatic channels."

Whole brain T2 dark fluid signal fell 4.6 percent, the CSF space grew by about 6.5 ml, and signal rose in the posterior venous sinuses in proportion to blood amyloid (r about 0.3). The authors read this as protein shifting from brain toward veins. It is an inference from signal intensity, a clinical sequence repurposed for a new job. Ultrafast MREG pulsation changes after the combined treatment were sparse, and a 13 person sham group (different people, lying still for 45 minutes) showed no significant change.

Prior mouse studies show vibration activates Piezo1 in vessel wall muscle cells, which "helps push waste fluid" through the glymphatic system.

The Piezo1 link is from rodent work, and the paper places Piezo1 mainly in vascular and lymphatic endothelium. Nothing in the human study measured Piezo1. The device is labeled "Piezo1 targeted" because of how its frequency is modulated, which is a design intention.

Six participants had nausea or headache, possibly from expectation; Kiviniemi warns about fainting in people with low blood pressure.

Four headaches, three with nausea, and one participant withdrawn for vomiting. Systolic pressure fell about 15 percent (134 to 114 mmHg). All symptoms resolved within 30 minutes. That is a fair safety record for healthy 44 year olds with no regular medications. Migraine sufferers were excluded at the door.

Christopher Brown (Southampton) suggests the intervention may slow hepatic breakdown of Aβ42, raising blood levels with no change in brain clearance. He proposes CSF tracers as the test.

The paper does not test it. The strongest predictor of the amyloid rise was the size of the blood pressure drop (r = −0.57 to −0.63). That fits the authors' story. It fits Brown's just as well. See below.