Greater Manchester has become the UK’s first Centre of Excellence for Music and Dementia, backed by more than £1 million and a decade of prior research, a Shanghai hospital has used brain imaging to show exactly which neural circuit rhythmic music switches on in Parkinson’s patients who freeze mid-stride, and a meta-analysis of thirty randomised trials has quantified a dose-response relationship between minutes of music and a premature baby’s vital signs. Music therapy has spent decades on the softer edges of medicine, valued but rarely measured. In 2026, all three of those findings share the same theme: it is being measured properly, and it is holding up.

Music therapy covers a wide range of practice, from structured clinical sessions to informal singing groups, and that breadth has always made it hard to pin down scientifically. The three developments below are each narrower and more rigorous than that reputation suggests: a funded regional programme built on a decade of prior trial data, a brain-imaging study isolating a specific mechanism, and a dose-response meta-analysis, the kind of design normally associated with pharmacology rather than singing.

A Centre of Excellence Built on Ten Years of Data

In 2024, Greater Manchester became the UK’s first Centre of Excellence for Music and Dementia, hosted by the orchestra Manchester Camerata in partnership with the University of Manchester and the Alzheimer’s Society. The programme is backed by more than £1 million from the Power of Music Fund, established by the National Academy for Social Prescribing with the Utley Foundation and Arts Council England, alongside funding from Mayor of Greater Manchester Andy Burnham and NHS Greater Manchester. The money supports three years of weekly music cafés, Music in Mind and Singing for the Brain, across all ten of Greater Manchester’s boroughs, delivered by a trained volunteer workforce of 300 “Music Champions” and aiming to reach more than 1,000 people living with dementia and their carers. Manchester Camerata chief executive Bob Riley described it as “a colossal moment built on over ten years of work and research in partnership with The University of Manchester,” and the funding explicitly includes anonymised, data-driven research into whether embedding music support this way reduces pressure on frontline NHS and social care staff. That local programme sits inside a wider evidence base that has just been updated: a 2025 Cochrane review of music-based therapeutic interventions for people with dementia, led by Jenny van der Steen and colleagues, concluded that compared with usual care, music-based therapy probably improves depression and may improve overall behavioural problems, though the same review found only low or very low-certainty evidence of any benefit lasting four weeks or more after treatment ends. It is a genuinely mixed verdict, not an uncritical endorsement, which is exactly what makes the Cochrane update a useful anchor for a well-funded regional programme built to generate more data of its own.

Watching the Brain Unfreeze in Real Time

The second development explains, at the level of brain activity, why a simple metronome beat can get a Parkinson’s patient walking again. Freezing of gait is a common and distressing symptom of Parkinson’s disease in which a patient’s feet suddenly, briefly stick to the floor mid-stride, and rhythmic auditory stimulation, essentially walking to a steady external beat, has long been observed to reduce it without a full mechanistic explanation. A team led by Chen Liu at Xinhua Hospital, affiliated with Shanghai Jiao Tong University School of Medicine, published a study in Brain and Behavior in May 2025 that used functional near-infrared spectroscopy, a technique that measures brain activation through changes in blood oxygenation, on 45 participants: 21 Parkinson’s patients with freezing of gait, 24 without it, and 10 healthy controls. Without any auditory stimulation, patients with freezing of gait showed significantly lower activation than the other two groups in the supplementary motor area and the left primary motor cortex controlling the lower limb, regions central to planning and initiating movement. Rhythmic auditory stimulation significantly increased activation in exactly those regions and altered how they communicated with the rest of the brain. It is a small, exploratory study, but it does something rare for a low-cost, non-drug intervention: it identifies the specific neural circuit the therapy appears to be switching back on, rather than simply reporting that patients walked better.

A Dose-Response Curve for Premature Babies

The third development applies a method borrowed straight from drug trials to a NICU cot. Researchers led by Fatemeh Shahbazi at Hamadan University of Medical Sciences published a systematic review and dose-response meta-analysis in PLOS ONE in January 2025, pooling 30 randomised controlled trials covering 1,855 preterm infants, born between 27 and 37 weeks’ gestation, who received music therapy in neonatal intensive care compared with usual care. Rather than simply asking whether music therapy helped, the team modelled how outcomes changed with the duration of music exposure, using a two-stage random-effects spline model, the same broad statistical family used to characterise how a drug’s effect scales with dose. Across the pooled trials, longer exposure to music therapy was consistently associated with improved oxygen saturation, more stable heart rate, slower respiratory rate, longer sleep duration, more stable body temperature and lower blood pressure and behavioural distress scores, with the relationship holding up as a genuine dose-response curve rather than a simple yes-or-no effect. As the authors put it, “the existence of a dose-response relationship can indicate a causal relationship between music therapy and the improvement of neonatal outcomes.” That is a notably careful, statistically grounded claim for an intervention that costs almost nothing and carries essentially no risk in a NICU setting.

Measuring What Was Always Assumed

What connects a funded regional dementia programme, a brain-imaging study of a single Parkinson’s symptom, and a dose-response meta-analysis of premature infants is a shift in the kind of evidence music therapy is now being asked to produce. None of the three treats “music helps” as sufficient. Each asks a harder, narrower question: helps whom, through which mechanism, at what dose, and for how long. The Cochrane review’s caution about long-term effects is as much a part of that shift as the positive findings either side of it. Life sciences has spent years applying pharmacological standards of evidence to a therapy that predates modern medicine by millennia, and in 2026 that scrutiny is starting to produce answers precise enough to build funded programmes and clinical protocols around.

Sources include Manchester Camerata, the University of Manchester, the Cochrane Database of Systematic Reviews, Brain and Behavior, and PLOS ONE.