What music therapy, music medicine and sound healing each mean

The word “sound” is doing a lot of work in 2026. One sentence can cover a hospital program, a wellness class and a device awaiting trial results. Here is the vocabulary, who delivers each, and a rough evidence grade.

Music therapy

What it is. A clinical profession. A music therapist assesses a patient, sets goals (speech after a stroke, agitation in dementia, anxiety before surgery) and uses live music, singing, movement or improvisation in sessions to reach them. In the US the credential is board certification (MT-BC) after a degree and a supervised internship.

Who delivers it. A credentialed therapist, in person, usually in a hospital, care home, school or clinic.

Evidence grade. Moderate, and specific. A 166-person Finnish randomized trial published in Alzheimer’s & Dementia: TRCI in August compared 10 weeks of group music therapy with a digital alternative and standard care in people with dementia. Music therapy produced modest gains in naming, verbal learning and executive function, and both music arms preserved autobiographical memory while controls declined. The primary general-cognition outcome was null. “Music improves cognition in dementia” overstates it; “music therapy helps some specific functions” is what the trial supports.

Music medicine

What it is. Recorded music, chosen with or by the patient, given in a clinical setting by staff who are not music therapists. A nurse hands over headphones. The Finnish trial’s digital arm, pre-recorded sessions run on screen by a non-therapist, sits on the border.

Who delivers it. Any clinician, or the patient.

Evidence grade. The strongest in this list for immediate outcomes. A 150-patient randomized trial at a Milan stroke unit, published in JAMA Network Open in July, gave inpatients an hour a day of personalized vocal music through a tablet. Depression scores were 2.68 points lower, anxiety 1.86 lower, insomnia 2.83 better than standard care, with no adverse events. Unblinded and self-reported, with no motor-recovery endpoints, but cheap and safe.

Sound healing and sound baths

What it is. A wellness practice: singing bowls, gongs, chimes, sometimes voice or didgeridoo, played over people lying down. Marketed for relaxation and sometimes for far more.

Who delivers it. Anyone. There is no license, no standard training and no required health screening.

Evidence grade. Low. The studies cited in press coverage are single-session trials of around 50 people measuring mood immediately afterward, and the popular “brainwave” explanation is not established. The AJC’s roundup notes the cautions practitioners themselves give: epilepsy, heart-rhythm conditions and implanted devices. An hour of rest in a quiet room is pleasant; that is the honest claim.

Vibroacoustic therapy

What it is. Low-frequency sound, typically in the tens of hertz, delivered as vibration through a bed, mat or chair so the body feels it rather than hears it.

Who delivers it. Physical therapists and wellness centers alike; equipment ranges from clinical tables to consumer mats.

Evidence grade. We have not reviewed a trial we would cite, and we will not grade it on reputation. Treat it as unproven until we do. Standard cautions apply to strong vibration, including pregnancy and implanted devices.

40 Hz light-and-sound devices

What it is. A headset and glasses that pulse light and sound together at 40 Hz, the “gamma” rhythm, for about an hour a day at home. The idea comes from mouse work in Li-Huei Tsai’s lab at MIT beginning in 2016.

Who delivers it. Cognito Therapeutics’ Spectris device is the lead product. It holds FDA Breakthrough Device designation, which speeds review but is not approval, and it is used in trials.

Evidence grade. Promising and unproven. In a 76-patient feasibility study, treated patients showed less EEG slowing over 24 weeks, and the company reports 77 percent less functional decline, 76 percent less cognitive decline and 69 percent less brain-volume loss. Every one of those numbers comes from 76 people and a company press release. The MIT human follow-up is five people. The decisive test is HOPE, a sham-controlled trial of about 670 patients at 70 US sites, fully enrolled, with topline results promised “before the end of the year” after slipping from mid-2026.

Ultrasound neuromodulation

What it is. Low-intensity focused ultrasound aimed through the skull at a deep brain structure, briefly changing its activity. It is the only non-invasive tool that reaches deep targets with millimetre precision.

Who delivers it. Research neuroscientists, in a lab, with imaging.

Evidence grade. Mechanistic, not therapeutic. In April, Oxford researchers reported in Neuron that stimulating the basolateral amygdala in healthy volunteers made them more willing to approach ambiguous neutral faces, with reduced amygdala-network communication on imaging: the first causal evidence in humans that the amygdala resolves emotional ambiguity. It is not a treatment for anything yet, and the press release gives no sample size.

What this means for you

If someone offers you “sound therapy,” ask which of these six they mean and who is delivering it. A board-certified therapist and a tablet of your own songs have trial evidence behind them. A gong has a pleasant hour. A 40 Hz headset and an ultrasound beam are experiments, and the date to watch is Cognito’s HOPE readout before the end of 2026.