← All posts

Thoughts / The craft

Does music therapy work for autism? What the research (really) says

July 7, 2026 · Thoughts / The craft

It's the first question. Always. It arrives almost before the parent has sat down: "This music therapy thing, for my son, does it work?"

And I'd love to answer with a nice full "Yes", round and reassuring, the kind that puts your mind at rest. It would be convenient. It would also be dishonest.

So let's do something harder and, I think, more useful: let's actually look at what the research says. Without selling you anything, and without selling short what I believe in either.

The research is encouraging. But it's patchy

In 2025 a systematic review came out that lined up the randomised controlled trials — the ones with the most rigorous design — on music therapy and autism: nine RCTs, more than one thousand three hundred children. The snapshot is honest and a little ruffled: some studies show significant improvements in social communication, others, even larger ones, find no clear change on the same indicators, while still picking up positive signals on specific aspects of social response.

Then, at the start of 2026, a meta-analysis came along that put eighteen studies back together and found something more encouraging: scores that improve significantly on the social, behavioural, sensory, emotional and even verbal level. The authors themselves, though — and I like this — call for caution: in some analyses the literature is still thin.

Translated: there's a direction, and in the most recent syntheses you can see it better. But it's still patchy. Often small samples, different ways of measuring, protocols that don't resemble each other. No magic wand, no "sound pill" to take twice a day.

For some people this honesty drains the enthusiasm. For me, it sparks the curiosity. Because inside those patches there's a red thread worth more than a promise.

The red thread: the relationship

When researchers went looking not at whether but at what makes the difference, an interesting predictor turned up: the relationship that gets built in sound between the child and the person beside them. Not the "right" piece, not the most high-tech instrument. The quality of the encounter.

And here music does something chemistry can't: we make it together. When a child and I find ourselves inside a sound dialogue — he offers, I take it in, I send it back, I wait — I'm not administering a remedy. I'm offering a place to exist in time, where his initiative counts, where a hesitation isn't a mistake to be corrected but a phrase to be listened to. It's that togetherness that sets us apart from the pill.

The beautiful thing is that today someone is trying to measure it, this relationship. In New York, NYU's Nordoff-Robbins Center has just won an award for a study that, with wearable sensors, tracks over fifteen weeks the synchrony of movement between the child and the therapist while they improvise together: the hypothesis is that when that synchrony grows, social communication blossoms too. And there's a detail that moved me: for children who can't bear contact, they're working on a way to read synchrony straight from video, so the sensors can come off. Even the measurement, here, learns to be gentle.

That's why I'm wary of anyone who promises "the music that cures autism". There's no music that's good for everyone. There is, if anything, a music child — it lives inside each of us, it wants to play and to be heard — and our craft is to go and find him right where he is.

On your own you go nowhere

And here comes the part I want to say loud and clear, because it's the most misunderstood.

Music therapy isn't an island, and the music therapist isn't a lone healer who walks into the room and works the miracle. It works when it sits inside an orchestra: the family, the school, the child neuropsychiatrist, the speech therapist, the educator. Not in competition — in dialogue.

And I don't say this out of faith alone. A 2026 qualitative meta-analysis brought together forty-two studies to understand what actually makes the collaboration between the music therapist and the special education teacher work. The ingredients that come back most often — in more than eight studies out of ten — are three: shared goals, frequent communication, clear roles. Nothing esoteric. The same things that make a good orchestra work.

That's why, before we even switch an instrument on, we sit down at a table and define the goals. Together. What are we looking for, for that child: the start of a communicative exchange? Tolerating closeness? Regulating when the world becomes too much? And at that same moment, at that same table, we decide something else too: how we'll know whether we're getting there. Because setting a goal and choosing how to observe it are twin gestures — they're born together. You don't need a laboratory: sometimes it's a grid, a diary, a small mark that counts the times the child takes the initiative. Yes, I know, "measuring" sounds cold. But the New York study up there reminds us: giving shape to what changes isn't the opposite of tenderness, it's the serious way of caring for it. Without shared goals — and without an agreed way of reading them — the most beautiful session in the world stays a lovely private concert. With goals, and with a patient observation of how that child changes over time, music becomes a journey.

So, does it work?

Let's go back to the parent sitting in front of me.

Honest answer: music therapy doesn't "cure" autism, and no serious professional will promise you that. But it can open channels of communication, give rise to moments of shared joy and of initiative, give the child back a place where he feels competent. And that, inside a team rowing in the same direction, is no small thing. It's almost everything.

Health, after all, isn't only a symptom going down. It's being able to take part. It's having someone who answers you when you play.

On that, I've been betting for thirty years.

Sources

Related

Read also

Episode

#07 Does music therapy work?

Episode

#18 Music therapy in a team: is it possible?

Episode

#27 Evidence-based music therapy