Music Therapy After Stroke: Walking, Speech and Mood, Graded Separately

After a stroke, exactly one outcome has strong controlled evidence behind a music based intervention, and it is walking speed. Everything else that gets claimed for music after stroke sits at a much lower rung, and this page grades one outcome at a time so that the strong row cannot lend its credibility to the empty ones beside it. That structure is the argument. If the outcomes were discussed together, the page would read as a general endorsement of something that has only been demonstrated in one narrow place.

I run a weekly singing and songwriting session on a stroke rehabilitation unit, as a community practitioner rather than a registered therapist. Two winters ago a man who had been on the unit about three weeks, and who could not retrieve his wife’s name when she visited, sang the whole of a chorus. Not a phrase of it. All of it, in tune, with the words in the right order and the last line landing where it should. His wife was standing behind him. Nobody in that bay said anything for a moment, and then somebody did the thing everybody does, which is to say that it is amazing what music can do.

I have thought about that afternoon a great deal, mostly about what I have learned not to conclude from it. He sang a chorus he had known for fifty years, in the presence of a strong external cue, with the melody carrying the phrasing and the timing for him. He did not then say his wife’s name. He did not say it later that week. Nothing about the hour told me whether the singing helped his recovery, whether it would have happened without me there, or whether it would happen again. It is also precisely the kind of moment that gets filmed, shared and offered as proof of things it cannot prove, and I understand exactly why, because when I first saw a clip of my own father singing two verses I took it as evidence of something too. That clip is why I started reading this literature. What it actually showed was one afternoon.

Every claim below carries a label: Tier 1 supported, Tier 2 promising but limited, Tier 3 practice and anecdote. The labels are graded per outcome rather than per activity, and what each of them requires is set out on reading arts in health research.

The one strong claim, and it is about walking

Rhythmic auditory stimulation improved gait velocity by 11.34 m/min (95% CI 8.40 to 14.28) in the 2017 Cochrane review of music interventions for acquired brain injury, pooled from 9 trials and 268 participants, at moderate quality¹. Tier 1.

That is the whole of the strong evidence on this page, and it takes one paragraph because it is one narrow finding. The technique sets a beat to a person’s measured step cadence and then moves it towards a target while somebody watches how the body copes. It is closer to physiotherapy than to anything most people picture when they hear the words music therapy, and the detail, including who delivers it, who it suits and why the unit is never converted, is on rhythmic auditory stimulation.

Two boundaries have to travel with the figure wherever it goes. It is a difference between group averages, so it predicts nothing about how much faster a particular person will walk. And it is a result about gait velocity, full stop. Not balance, not falls, not independence, not confidence, and certainly not mood or speech. Every remaining section of this page describes an outcome that this number says nothing about.

Communication and aphasia

Music based approaches to aphasia are among the oldest ideas in the field, they are taught and used, and this site holds no locked effect size for any of them.

The best known is melodic intonation therapy. Broadly, short functional phrases are intoned rather than spoken, on two or three pitches, with a steady rhythm and often with the person tapping a hand in time, and the support is then withdrawn in stages so that the phrase moves back towards ordinary speech. It is delivered by speech and language therapists, by music therapists, and sometimes jointly, and it is a structured protocol rather than a general encouragement to sing. Anyone who has watched it will recognise how unlike singing along it looks.

The reason no number for it appears here is worth stating precisely, because silence would be read as scepticism. There is no effect size on this page, and that is a decision rather than an omission. For one to appear, this site would need a systematic review or an adequately powered randomised controlled trial reporting a named speech or language outcome, with an effect size, a confidence interval, a certainty rating, a sample size and a year, readable in the original rather than through a secondary source quoting it. Until that has been checked and locked, the label is Tier 3, of the not adequately studied here kind rather than the controlled evidence is null kind. Those two situations look identical from the outside and mean opposite things, so the distinction gets spelled out every time.

What that absence does not mean: that melodic intonation therapy fails, that a speech and language therapist offering it is doing something unsupported, or that a family should decline it. What it does mean is that nobody should be told a number, and if somebody quotes one, the useful question is which review it came from and whether it can be opened.

Mood after stroke, and the figure that keeps getting imported

Low mood after stroke is common, families ask about it more than about anything else, and there is no locked effect size for it here either.

The same conditions would have to be met: a review or an adequately powered trial, on a named mood outcome, in people after stroke, with all five of the things a Tier 1 label requires. What exists instead is a habit of borrowing, and the figure that gets borrowed is a specific one. The 2025 Cochrane review of music based therapeutic interventions in dementia reports depressive symptoms at SMD -0.23 (95% CI -0.42 to -0.04) at moderate certainty², which is Tier 1 and a real, small effect.

It is a real, small effect in people with dementia. Carrying it across to stroke changes the population, the condition, the reason for the low mood, the trials it came from and the outcome instruments used, while keeping the number and its authority intact. That is not a rounding error in a claim; it is a different claim wearing another one’s clothing, and it is exactly the failure this site was built to make visible. You will meet the borrowing in funding bids, in service descriptions and in perfectly sincere conversation. The rest of that dementia review, including the row that is null, is on what the dementia evidence shows; how the same borrowing happens in the other direction, into adult mental health, is on music therapy for depression and anxiety.

What a session actually looks like, stage by stage

There is no single answer, because the three stages after a stroke are three different clinical situations and the work looks nothing alike across them.

Acute. In the first days a person may be exhausted, medically unstable, and unable to tolerate much of anything. Where music therapy happens at all it tends to be short, quiet and at the bedside: single sustained sounds, familiar melodies at low volume, a therapist watching for the smallest signs of orientation, tolerance or response. The goals here are modest and are usually written in terms of alertness, engagement and comfort rather than recovery of function.

Rehabilitation. This is where most provision sits, and where the work becomes structured and goal directed. Sessions may run alongside physiotherapy on movement and timing, alongside speech and language therapy on voice, breath and phrasing, or on their own with goals about communication, mood or motivation. Instruments are chosen so that no skill is required. Group work appears here too.

Community. After discharge, anything that continues is normally not a clinical service at all. It is a stroke choir, a community group, a class, or private practice paid for by the person.

Around the clinical stages sits the half of the job that distinguishes therapy from activity: an initial assessment, written goals, session notes, clinical supervision, multidisciplinary team meetings and a planned discharge conversation. What happens in a music therapy session sets that out in more detail. And anything happening at a bedside runs straight into consent, which is harder after a stroke than almost anywhere, because the person whose agreement is needed may be the person whose communication has been affected. That problem is taken seriously on consent and safeguarding in arts in health.

Who provides it, and how patchy that is

Most stroke units in the United Kingdom do not have a music therapist, and the reason is arithmetic rather than attitude.

In July 2026 the Health and Care Professions Council register listed 6,103 arts therapists across all four protected titles, for the whole country³. Spread that number across mental health services, learning disability services, schools, hospices, prisons, care homes, community teams and acute hospitals, and stroke rehabilitation is competing for a very thin slice of a very small profession. Where a post does exist it is usually funded inside a rehabilitation service or by a hospital charity, which means it can also disappear at the end of a grant. The British Association for Music Therapy holds a directory of practitioners and is the practical place to start looking⁴, and the routes are set out on how to find a music therapist.

It is worth separating three things that all involve music inside a hospital and are constantly conflated. A registered music therapist is clinical staff working to assessed goals with named patients, which is the regulated profession set out on music therapy. A hospital arts programme is a managed programme run by an arts team or a charity, described on arts in hospitals. A musician playing on a ward is a third arrangement again, with its own set of practical problems, covered on live music on hospital wards. I am the third kind of person. Conflating the three is how a ward ends up believing it has a therapy service when what it has is me and a keyboard.

The long tail after discharge

Most of life after a stroke happens after the rehabilitation ends, and that is where the least is provided and where almost none of the evidence was gathered.

Services are organised around the weeks of rehabilitation, and so is nearly all of the research. But somebody discharged in the spring may live for decades afterwards with a weak arm, a changed voice and a mood that nobody is monitoring, and the honest answer about what is available then, in most of the country, is a community group if one exists locally and nothing if one does not. Stroke choirs, community singing groups and seated exercise classes fill part of that gap. They are not clinical services, they are not evidenced as clinical services, and they are frequently the only thing there. Dance after stroke covers the movement equivalent, where the same pattern holds.

I run one of those groups. I would defend it, I think the people who come are better off for coming, and I cannot show you a number that says so. That is Tier 3 of the not adequately studied kind, and it is where most of the useful things in this sector sit.

What to ask, and what not to expect

Ask specific questions about specific outcomes, because a general request for music produces whatever the service happens to have.

• Does this service have a music therapist, and if not, is there one anywhere in the trust?
• Is rhythmic cueing part of the gait work, and who delivers it here?
• If speech is the priority, has a music based approach been discussed with the speech and language therapist, and what would it be aimed at?
• What continues after discharge, and who arranges it?

What not to expect: a number for anything except walking speed, a service in most places, or a continuation after discharge. And what not to ask for: music therapy as a general good. It is a specific clinical intervention with specific goals, and the request that names an outcome gets a better answer than the request that names an activity.

Tier 3 runs in both directions here

On this page the empty rows are the majority, and that makes the standing caution more important than usual rather than less.

Overclaiming is the obvious risk and this field does plenty of it. The opposite error does more damage to the people it lands on: reading an absence of evidence as a finding of uselessness, and quietly withdrawing something from somebody on that basis. Almost nothing that happens in a stroke unit dayroom has been properly studied, for reasons that have nothing to do with whether it helps. Trials are hard to blind, funding is short, the outcomes people care about most are the hardest to measure, and the sector’s money goes into delivery rather than research. Where controlled evidence exists and is genuinely null, this site says null in those words. On this page, apart from the gait row, it mostly says nobody has looked properly, and those are different sentences.

The man who sang the chorus was discharged a few weeks later and I never found out what happened to him. That is the ordinary shape of this work and it is also why the anecdotes should not be trusted: the person telling you the story is never the person who followed it up. If a stroke team has arranged rehabilitation, none of this is a reason to change any part of it. Music after a stroke, at its best, is something added to a recovery that other people are managing, and it should never be the thing that replaces one of them.