There is a genuine and active field of research using ultrasound in Alzheimer's disease. It is real science, published in serious journals. But it does not currently support the claim as worded.

The memory-restoration finding is from mice, not patients. The single most-quoted study, and almost certainly the ultimate origin of the "restores memory" phrasing, is the 2015 University of Queensland paper in Science Translational Medicine. Its title is literally "Scanning ultrasound removes amyloid-β and restores memory in an Alzheimer's disease mouse model." The abstract describes a non-pharmacological approach for removing amyloid-beta and restoring memory function in a mouse model, using repeated scanning ultrasound without any additional therapeutic agent. That result has never been replicated as memory restoration in human patients.

The strongest human plaque-reduction result required a drug. The 2024 NEJM trial by Rezai's team at West Virginia University treated three participants with mild Alzheimer's. Focused ultrasound was applied alongside six monthly infusions of aducanumab, an anti-amyloid antibody, to temporarily open the blood-brain barrier so more antibody could reach the target region. Amyloid reduction was greater in the ultrasound-treated regions than in the matching untreated regions of the opposite hemisphere: differences of roughly 48%, 49% and 63% across the three participants over 26 weeks. The ultrasound was a drug-delivery enabler. It did not dissolve plaques on its own.

Cognition did not improve in that trial. Two participants had no neurologic, cognitive or behavioral changes at last follow-up. The third showed cognitive decline at 30 days post-follow-up. The authors state that given the small number of participants it is difficult to determine whether the cognitive changes were related to disease or to the procedure. Nothing in this trial resembles memory restoration.

Standalone ultrasound results in humans are mixed and small. A preliminary Korean study of eight patients receiving transcranial focused ultrasound to the right hippocampus found no evidence of blood-brain barrier opening on imaging, yet reported statistically significant improvement on immediate recall and recognition memory on a verbal learning test. A 2025 randomized trial of transcranial pulse stimulation in JAMA Network Open reported improved cognitive scores in a younger subgroup, improvement in depressive symptoms, and increased functional brain activation and connectivity. Note that the cognitive benefit was a subgroup finding, not a whole-sample result.

Pooled RCT evidence is currently null. A systematic review and meta-analysis in Neurological Sciences pooled three randomized controlled trials totaling 84 participants and found that transcranial ultrasound stimulation did not significantly improve global cognition compared with sham stimulation (SMD 0.31, p = 0.54).

A separate technology may be conflating the picture. The phrase "sound waves" also maps onto MIT's GENUS work using 40 Hz light and sound. In mouse models that approach reduced amyloid and tau and improved cognition. Human phase II work at MIT and spinoff Cognito Therapeutics has reported slowing of brain atrophy and improvement on some cognitive measures versus untreated controls. This is early-stage and distinct from focused ultrasound, but shares the "sound" framing.