What Happens in a Music Therapy Session: The Room, the Instruments and the Hour

A music therapy session is usually quieter, slower and more one to one than people expect: an hour or so in a room with instruments chosen to be playable without training, in which a registered therapist works towards goals written down before it started, mostly through improvised live music rather than songs anybody has to know. What makes it a clinical intervention, though, is not the hour. It is the referral and the assessment in front of it, and the notes, the supervision and the planned ending behind it.

I have watched a good deal of this work and I do none of it. I run community singing and songwriting sessions in two care homes and on a stroke rehabilitation unit, as a community arts practitioner, and I am not registered to call any of it therapy. What I notice most is the twenty minutes after the hour. When my session ends I stack the chairs, hand the register to the activities coordinator, say goodbye to two or three people in the corridor and drive home, and by teatime the afternoon exists nowhere except in my head and in whatever the residents remember of it. When the music therapist who works in the same building finishes, she goes to a small office and writes for twenty minutes, one entry per person, into a record that other clinicians read. Once a fortnight she takes that work to clinical supervision and someone senior asks her questions about it she would not otherwise have to answer. That is the difference between what she does and what I do, and it is almost entirely invisible from inside the room.

Where this page states what the evidence shows, the claim is labelled. Tier 1 means supported, Tier 2 means promising but limited, and Tier 3 means either that no adequately powered controlled study exists or that the controlled evidence is null, with the article saying which. How the labels are arrived at is explained on reading arts in health research.

Before the first session: referral, assessment and agreed goals

Nothing starts with a session. It starts with a referral from someone in a clinical or educational role, and then an assessment that produces the goals the rest of the work will be measured against.

The referral typically comes from a mental health team, a community learning disability team, a child and adolescent service, a hospice, a school’s special educational needs coordinator, a ward consultant or a rehabilitation team. It says who the person is, what the difficulty is, and what somebody hopes might change. Where the work is bought privately the referral may be a parent or the person themselves, but the assessment still happens, because a therapist who takes on work without one has nothing to aim at and nothing to review against. How to find a music therapist covers who to ask in each route.

The assessment usually runs across one or two sessions and is not treatment. The therapist is finding out how somebody responds to sound and to another person making sound near them: what they initiate rather than only accept, how they handle silence, how they handle a change in tempo or a piece stopping before they expected it, whether they can bear the ending of something. Family history matters here, because a person’s own music carries a great deal of biography, and so does the practical picture: hearing, vision, fatigue, medication timing, the hour of day when they are most themselves.

What comes out of it is written down. Goals in this work are rarely “enjoy music”, because that is not reviewable. They are closer to: sustain joint attention for longer within a session; initiate a musical exchange without prompting; tolerate transitions with less distress; use breath and voice in ways that support speech; express something about a diagnosis that has not been sayable in words. Those goals are discussed with the person, with a family member where that is appropriate, and with the referrer, and they set the frequency and the expected length of the course. The whole of that scaffolding is what the pillar page music therapy means when it calls this a clinical intervention.

The room and the instruments

The room is arranged so that music can be made by someone with no training, no strength and no plan, and the instruments are chosen on exactly that basis.

Expect hand drums, a large frame drum or ocean drum, tuned percussion such as chime bars or a xylophone, shakers, a few small wind instruments, and usually a piano or a guitar that the therapist plays. What these have in common is that they sound acceptable immediately. Strike a chime bar and you get a chime bar. There is nothing to get wrong, no embouchure to develop, no wrong note in a pentatonic set, and no gap between intention and result of the kind that makes people give up on an instrument at twelve. That is a deliberate clinical choice rather than a matter of budget: the instruments are there to remove the barrier of skill so that whatever the person does becomes usable material.

The room itself is usually plainer than people imagine. Chairs, floor space, instruments within reach or brought within reach, a door that closes. Where sessions happen at a bedside or in a lounge, the therapist works with a small portable selection and the acoustic of wherever they are, which is not ideal and is extremely common. Recording sometimes happens, for supervision or for review, and it requires its own consent that a relative usually cannot give on someone else’s behalf, which is one of several reasons why consent and safeguarding in arts in health is a harder subject in this field than it looks.

The first ten minutes, realistically

The first ten minutes are frequently unremarkable, and people who arrive expecting a dramatic breakthrough often read that as a session going badly. It is not.

Typically there is a greeting, sometimes a short piece of music that marks the beginning and comes back each week as a boundary, and then a long stretch in which the therapist offers something small and waits. The waiting is the work. A person may take an instrument and put it down again. A person may talk for six minutes about the traffic. A person may do nothing observable at all while the therapist plays quietly, watching breath, gaze and hands. In work with someone in the later stages of dementia the first response may be a change of breathing rather than a movement, and the therapist is looking for exactly that.

What almost never happens is instruction. Nobody is told how to hold anything, corrected, counted in, or asked to try again. If you are watching for the first time and expecting something that resembles a lesson or a performance, the absence of both is the strongest signal that you are in the right place. My own sessions are noisier, faster and more led than this, because I am running an activity for a group of twenty and the activity is the point. The distinction is worked through on music therapy against community music.

Improvisation, and why nobody needs to be musical

Most of what happens is improvised, and improvisation here means a shared, unwritten musical exchange rather than a jazz solo. It is the default because it is the only material that comes from the person rather than from the repertoire.

If somebody taps twice, the therapist can answer, match, delay, or leave a space where the third tap would go and see whether it arrives. That exchange carries information about initiation, turn taking, tolerance of another person and capacity to be affected by them, all of which are clinical observations rather than aesthetic ones. Known songs are used too, particularly with older people, where a song from someone’s twenties can open a route into speech, memory and mood that conversation cannot. But a familiar song is a tool used deliberately, not the structure of the hour.

Nobody is assessed on their playing and nothing has to sound good. I have heard people who have not played anything since school apologise for their drumming, and I have watched a therapist decline to reassure them and instead simply play along with what they were doing until the apology stopped mattering. Musical ability belongs to the therapist, which is why entry to the training assesses it so heavily, as described on training as a music therapist.

Individual work and group work

Individual sessions are the norm where the goals are personal and the material is difficult; groups are used where the goal itself is social, and they are not simply a cheaper version of the same thing.

One to one work allows the therapist to follow one person exactly, which is what makes the fine grained observation possible. Group work does something different: it puts turn taking, waiting, tolerating other people and being heard by them at the centre, and those can be precisely the goals. In a group the therapist is managing several relationships at once and often works with a co worker, and the notes afterwards are correspondingly harder to write.

The pairing that confuses people most is a therapy group and a community group, which can look identical and are not. A therapy group has a closed membership agreed at assessment, goals for each member, notes on each member and a planned ending; a community group has an open door and none of those. Dementia choirs and singing groups describes the community version, which is my own territory, and it is worth having on its own terms.

What happens outside the hour

This is the part that makes the difference, and it is the part nobody watching a session ever sees.

After each session there are notes. They are written for a clinical record rather than for the participant, they are framed against the agreed goals, and they include the things that did not work as well as the things that did. They are what allows the work to be reviewed rather than merely repeated, and they are what a colleague would read if the caseload changed hands.

Then there is clinical supervision, which is not line management and is not optional. A registered therapist takes their work regularly to a supervisor and is asked about it: what they are avoiding, whose distress they are carrying, why a particular session unsettled them, whether the goals still make sense. Working closely with people who are frightened, dying, confused or very angry produces reactions in the practitioner, and supervision exists so that those reactions are examined somewhere useful rather than acted out in the room. Registration with the Health and Care Professions Council, which listed 6,103 arts therapists across all four protected titles in July 2026¹, carries obligations of exactly this kind.

Then there are multidisciplinary meetings, reviews with referrers, notes read by nurses and reports written for education plans and case conferences. On a ward this is also how a therapist finds out that the person they are about to see has had a bad night, is nil by mouth, or has had a family visit that changed everything. Programmes that put musicians into hospital buildings without any of that structure are doing something different and sometimes valuable, described on live music on hospital wards.

How a course ends

A course of music therapy ends on a planned date that has been named in advance and worked towards, rather than stopping when the funding runs out or the term does.

Endings are prepared over several sessions. They get talked about, and often played about, because for many people referred to this work separation is precisely the difficult thing, and an ending done carelessly can undo a good deal. A discharge summary usually goes back to the referrer, sometimes with a recommendation about what might follow.

There is also an evidence reason to take endings seriously, and it is not a comfortable one. The 2025 Cochrane review of music based therapeutic interventions in dementia, covering 30 studies and 1,720 participants randomised, found no evidence of any effect persisting four weeks after treatment ends². That is Tier 3, and specifically the kind of Tier 3 that means persistence has not been demonstrated rather than that it has been tested and ruled out. It is not a reason to decline a course. It is a reason to ask, before the course starts, what is meant to happen afterwards, and it is one argument for community provision continuing after clinical work finishes rather than the two being treated as rivals. Music therapy for dementia sets out the review in full, including a small effect on depressive symptoms at SMD -0.23 (95% CI -0.42 to -0.04) at moderate certainty, which is Tier 1 and is a difference between group averages rather than a prediction about any individual.

How the session changes by setting

The structure survives everywhere. The hour looks very different depending on the building.

On a ward, sessions are short, interrupted and opportunistic. Drug rounds, physiotherapy, meals, ward rounds and visitors all outrank a music therapist, and a session may end after eleven minutes because a porter has arrived. On the stroke unit where I work, a good deal of the music therapy is closely goal directed and physical, and the technique with the strongest evidence in the whole of arts in health belongs here: rhythmic auditory stimulation improved gait velocity by 11.34 m/min (95% CI 8.40 to 14.28) in 9 trials and 268 participants, at moderate quality³, which is Tier 1. Note how narrow that claim is: gait velocity and nothing else, not mood and not communication. Music therapy after stroke covers the wider picture, where the evidence is much thinner.

In a care home the session may happen in a bedroom, a lounge with the television on, or a corridor alcove, and the therapist spends real effort on getting a door shut and a television off. Who else is in the room is rarely fully controllable, which affects both the work and the consent position. In a school, sessions are timetabled, often shorter, frequently attached to targets in an education plan, and the therapist reports into a system with its own paperwork. In a hospice, sessions bend entirely around how someone is that day, family are often present, and a session may become work with the family instead. Whatever the setting, the notes still get written and the supervision still happens.

What it costs and who pays

Most people who receive music therapy never see a price, because it arrives through a service that has already paid for it.

Where the work is commissioned by the NHS, a hospice, a school, a local authority or a charity, it is free to the person receiving it, and the binding constraint is whether a post exists rather than what it costs. Privately, the British Association for Music Therapy recommends from £52.00 upwards for an individual session⁴, which is a recommended floor, not an average and not a typical price, and real rates vary with region, setting and whether the work is individual or group. In the United States the 2021 AMTA workforce analysis reports a median of $79 per hour⁵; quote the median rather than the mean and date it 2021, because it is not current.

Ask what a quoted price covers before comparing two of them. Assessment sessions, travel to a home or a school, report writing, attendance at reviews and the therapist’s own supervision are all real costs, and different practitioners fold them in differently. The comparable figures are not the hourly rates but what a term of work would total.

A last thought rather than a summary. If you get the chance to sit in on a session, watch the beginning and the end rather than the middle. The middle is where the music is, and it is the part that photographs well. The beginning tells you whether the person was expected and prepared for, and the end tells you whether the hour has anywhere to go afterwards. Neither of those is about music at all, and both are what separate a clinical intervention from a pleasant afternoon. Whatever a session does or does not do, it sits alongside somebody’s treatment rather than in place of it, and nobody should be stepping away from care they need in order to make room for it. If you want the same walk through for the visual modalities, it is on what happens in an art therapy session.