Music Therapy

Abstract

Music therapy is the clinical use of music to address physical, psychological, cognitive, and social needs; MeSH classifies it under psychotherapy. This article treats it as a scientific object whose central claim, that organised sound can measurably change mind, body, and behaviour, can be tested rather than assumed. It traces the field from the pioneering work of E. Thayer Gaston, Paul Nordoff and Clive Robbins, and Helen Bonny to the Cochrane reviews that estimate its effects across depression, dementia, autism, stroke, and cancer, and to the neuroscience of music-evoked emotion and rhythmic entrainment that is beginning to explain how it works. Three interactive demonstrations model the dose-response relationship between sessions and benefit, the entrainment of movement to a beat, and the pooling of small trials across conditions.

Keywords: music therapy, rhythmic entrainment, music-evoked emotion

Music therapy rests on a claim as old as music itself and as demanding as any in clinical science: that structured sound can reach a person when other means cannot, and can change how they feel, move, think, and relate. What separates the modern discipline from that ancient intuition is the willingness to submit it to trial, to ask whether a course of music therapy produces measurable benefit, for whom, in what conditions, and by what causal route. This article follows that arc from the profession's mid-century founding to the current evidence base and the neuroscience that is beginning to explain, rather than only demonstrate, its effects.

What Music Therapy Is

Music therapy is the evidence-based clinical use of music to accomplish individualised goals within a therapeutic relationship, delivered by a credentialed professional. It is a type of psychotherapy: the National Library of Medicine's Medical Subject Headings file it beneath the psychotherapy descriptor, and it belongs to the broader family of creative arts therapies alongside art, dance/movement, and drama therapy. What distinguishes it within that family is its medium. Where verbal therapy works through conversation and art therapy through a made object, music therapy works through organised sound, treating rhythm, melody, and harmony as both the material to be shaped and the means of change.

The field is conventionally divided into two modes of practice. In active or expressive methods the client makes music, singing, drumming, or improvising with the therapist, so that the interaction itself becomes the treatment; the Nordoff-Robbins tradition of creative music therapy is the paradigm case. In receptive methods the client listens to music the therapist selects or performs, often to induce a state, evoke imagery, or support relaxation, as in Helen Bonny's Guided Imagery and Music. The practice requires no musical training or talent from the client, and a session may involve no performance in any ordinary sense, only sound used deliberately toward a clinical end. Sessions are conducted individually or in groups, and the therapist's choices of tempo, key, dynamics, and structure are the clinical instruments.

Origins and Pioneers

Music has been used for healing across cultures and centuries, but music therapy as an organised profession is a mid-twentieth-century development, prompted in the United States by the treatment of veterans after the Second World War, when musicians brought into hospitals produced effects that clinicians could not ignore. The figure most associated with placing the emerging field on an academic footing is E. Thayer Gaston, a music educator and researcher at the University of Kansas often called the father of American music therapy, who helped build the training, professional structures, and research orientation that turned an ad hoc practice into a discipline.

Two clinical traditions gave the field its characteristic methods. Paul Nordoff, a composer, and Clive Robbins, a special educator, developed through the 1950s and 1960s an improvisational approach, creative music therapy, in which therapist and client make music together in real time so that the musical relationship itself carries the therapeutic work; their method remains one of the most influential in the world. On the receptive side, Helen Bonny, a music therapist and researcher, developed the Bonny Method of Guided Imagery and Music, in which a client in a relaxed state listens to carefully sequenced classical programmes and follows the imagery and feeling that arise, a technique that drew music therapy into the exploration of consciousness and inner experience. Between the expressive and the receptive, the active and the listening, these founders established the range the field still occupies.

Theoretical Foundations

Beneath the clinical methods lies a claim that neuroscience has increasingly borne out: that music is not a peripheral pleasure but engages core systems of emotion, reward, and movement in the brain. Music reliably evokes emotion, and it does so through identifiable neural machinery. Reviewing the evidence, Koelsch showed that music-evoked emotions engage the same limbic and paralimbic structures, the amygdala, nucleus accumbens, hippocampus, and orbitofrontal cortex, that underlie emotion generally, so that music can act directly on the brain's affective circuitry rather than merely accompanying a mood (Koelsch, 2014). The most direct evidence that music engages the reward system came from Salimpoor and colleagues, who used positron emission tomography to show that the peak pleasure of music, the spine-tingling chills, is accompanied by dopamine release in the striatum, with the anticipation of a musical climax and the climax itself recruiting anatomically distinct parts of the reward pathway (Salimpoor et al., 2011). If music can drive the emotion and reward systems, then using it deliberately to alter feeling has a mechanistic rationale its founders intuited but could not demonstrate.

A second foundation is rhythmic. The human motor system spontaneously synchronises to a regular auditory pulse, a phenomenon of entrainment that couples perception and action. Thaut and colleagues traced this to the reciprocal connections between the auditory and motor systems and showed that a steady beat can serve as a timekeeper for movement, stabilising and priming the motor output, which is the neurobiological basis of Rhythmic Auditory Stimulation for gait (Thaut et al., 2015). These two foundations, emotional and motor, mark the two great channels through which music therapy is thought to act: on affect through the limbic system, and on movement through auditory-motor entrainment.

The Question of Outcome

Whether music therapy produces benefit beyond attention and the passage of time is a question addressed by the field's highest evidential standard, the Cochrane systematic review, several times over. The application with the longest record is depression. Maratos and colleagues' original Cochrane review pooled the controlled trials then available and found that music therapy was accepted by people with depression and associated with improvements in mood, while cautioning that the trials were few and of variable quality (Maratos et al., 2008). The updated review by Aalbers and colleagues, incorporating newer and larger trials, reached a firmer conclusion: music therapy added to treatment as usual produced clinically meaningful short-term reductions in depressive symptoms and anxiety compared with treatment as usual alone (Aalbers et al., 2017). Figure 1 shows what an effect of that kind means as the overlap of two distributions, and the demonstrations that follow let a reader manipulate the underlying quantities.

The message across the depression reviews, and across the wider Cochrane literature examined below, is twofold. The average effect is genuine and positive, placing music therapy among credible psychosocial treatments; but the confidence it can bear is bounded by the size and quality of the underlying trials, a limitation the reviewers themselves press more firmly than any external critic.

Clinical Applications

The breadth of music therapy's evidence base is unusual among the creative arts therapies, because music engages so many systems that its clinical reach is correspondingly wide. In autism, Geretsegger and colleagues' Cochrane review found that music therapy, compared with placebo or standard care, improved social interaction, communication, and social-emotional reciprocity in children with autism spectrum disorder, giving a rigorous footing to a long-standing clinical use (Geretsegger et al., 2014). In dementia, van der Steen and colleagues' review concluded that music-based interventions probably reduce depressive symptoms and improve behavioural problems, and may improve emotional wellbeing and anxiety, in people with dementia in institutional care, a notable finding in a population for whom pharmacological options are limited and hazardous (van der Steen et al., 2018).

The reach extends into physical medicine. For people with coronary heart disease, Bradt and colleagues found that listening to music reduced anxiety and had small beneficial effects on heart rate and blood pressure, a use of receptive music as a low-cost adjunct in cardiac care (Bradt et al., 2013). For cancer patients, a further Bradt review reported that music interventions may improve anxiety, pain, fatigue, and quality of life, distinguishing throughout between therapist-delivered music therapy and simpler music-medicine listening (Bradt et al., 2016). At the very start of life, Standley's meta-analysis of music therapy for premature infants found reliable benefits on physiological and developmental measures in the neonatal intensive care unit, from feeding to oxygen saturation (Standley, 2002). And in stroke, Särkämö and colleagues showed in a landmark trial that daily music listening in the early recovery period enhanced verbal memory and focused attention and improved mood, compared with listening to audiobooks or to nothing, evidence that even receptive music can shape cognitive recovery (Särkämö et al., 2008). The pattern across these conditions, a consistent, modest, positive effect, is what the third demonstration below pools into a single estimate.

Mechanisms of Change

How music therapy works, as against whether it works, is where the field's connection to cognitive neuroscience is sharpest, and it turns on the two foundations set out above. Through the affective channel, music's capacity to engage the limbic and reward systems means that a therapist can use it to up-regulate or down-regulate emotion directly, a mechanism grounded in the neuroimaging evidence that music-evoked emotion recruits the amygdala, accumbens, and orbitofrontal cortex (Koelsch, 2014). Through the motor channel, auditory-motor entrainment lets a rhythmic cue organise movement in time, so that a patient whose gait is slow or unsteady can borrow the metronome's regularity, the mechanism Thaut and colleagues formalised as the basis of Rhythmic Auditory Stimulation (Thaut et al., 2015).

A third, quantitative principle binds these together: dose. Analysing music therapy for people with serious mental disorders, Gold and colleagues found a dose-response relationship, benefit increased with the number of sessions, with the clearest gains appearing once a threshold number of sessions was reached (Gold et al., 2009). The finding matters twice over. Clinically, it implies that an adequate course, not a token exposure, is what delivers the effect. Methodologically, it warns that a trial delivering too few sessions is testing an under-dosed treatment and may report a null that reflects the design rather than the therapy. Beyond the clinic, de Witte and colleagues' meta-analysis of music therapy and music medicine for stress found reliable reductions in both physiological arousal, such as heart rate and cortisol, and psychological stress, connecting the therapy's effects to the measurable stress-response systems and providing a mechanistic bridge from a session to the body (de Witte et al., 2022).

Current Directions

The field's current trajectory is set by the same two demands reshaping psychotherapy research generally: stronger outcome evidence and specified mechanism. On the outcome side, the maturing of the Cochrane programme, from the first depression review to the strengthened update and the dementia and autism reviews, has moved music therapy from a practice justified by case reports toward one with quantified, if modest, effects, and each review converts its estimate into an explicit call for the larger, better-controlled trials the evidence still lacks (Aalbers et al., 2017; van der Steen et al., 2018). A parallel effort is working to standardise what counts as a dose and a protocol, so that trials of nominally the same treatment can be compared and combined.

On the mechanism side, the growth of music neuroscience is the most consequential development, because it offers to connect music therapy to processes, limbic engagement, reward prediction, auditory-motor coupling, stress-axis regulation, that can be measured with the tools of experimental psychology and neuroscience (Koelsch, 2014; de Witte et al., 2022). Neurologic Music Therapy, the standardised, neuroscience-grounded family of techniques for sensorimotor, speech, and cognitive rehabilitation, is the clearest expression of this turn, replacing the field's older and largely metaphorical vocabulary with interventions specified at the level of the brain systems they engage (Thaut et al., 2015). Both fronts advance against the same backdrop of a broad but uneven evidence base, examined next.

Criticisms and Open Questions

The candid assessment of music therapy's evidence is that it is broad but uneven, and the most useful critiques come from within the field. The recurring limitation, stated in nearly every Cochrane review, is that the primary trials are small, heterogeneous in their interventions and outcomes, and often at risk of bias, so that even a positive pooled estimate must be read with low to moderate certainty (Maratos et al., 2008; van der Steen et al., 2018). Control conditions are a particular difficulty: comparing music therapy with no treatment cannot separate the specific effect of the music from the general effects of attention, activity, pleasure, and expectation, which is why the strongest trials use active comparators such as audiobooks (Särkämö et al., 2008).

Two deeper problems attach to the practice as well as the evidence. The first is definitional heterogeneity, and it runs along a fault line the field itself draws: therapist-delivered music therapy, in which a trained clinician uses a musical relationship toward individualised goals, is a different intervention from music medicine, in which patients simply listen to pre-recorded music, yet reviews must often pool them, blurring what is being tested (Bradt et al., 2016). The second is the allegiance and expectancy bias that inflates the psychotherapy literature generally, to which a pleasurable, engaging, relationship-based treatment is if anything especially exposed. None of this shows that music therapy does not work; it shows that the honest current claim is a real, moderate, condition-specific benefit, held with appropriate uncertainty and awaiting the larger and better-controlled trials its own advocates call for.

Worked Example

The first demonstration makes the dose-response relationship concrete. Modelling benefit as a saturating function of the number of sessions, d(n) = 0.80 × n / (n + 16), captures the qualitative finding that effect grows with dose and then levels off. At the default of 20 sessions the predicted effect is about 0.44, a small-to-moderate benefit; at 10 sessions it is only about 0.31, and at 50 sessions about 0.61. The lesson is in the shape rather than the exact numbers: the gain from the fifth session is large and the gain from the forty-fifth is small, so a trial that stops at a handful of sessions samples only the steep early part of the curve and can report a disappointing average for a treatment that would have worked at an adequate dose. The constants are illustrative, chosen to reproduce the reported pattern, not fitted to a specific dataset.

The second demonstration turns entrainment into something a reader can drive. Over a fixed four-second window it marks an auditory pulse at a chosen tempo and places a footfall on each beat, so that the walker's cadence is set by the metronome. At 100 beats per minute the pulse repeats every 600 milliseconds and seven footfalls fall inside the window; raising the tempo shortens the interval and quickens the gait, lowering it slows the walk. This is the lever Rhythmic Auditory Stimulation uses: because the motor system locks to the beat, a therapist can draw an unsteady or slow gait toward a faster, more regular cadence simply by setting the tempo of the cue.

The third demonstration shows why a field of small trials across many conditions must be pooled. Given illustrative per-condition effect sizes, each with a standard error, the inverse-variance method weights every condition by its precision and returns a combined estimate whose interval narrows as conditions are added. With all five illustrative conditions included, depression, dementia, autism, stroke, and cardiac anxiety, the pooled effect is about 0.44 with a 95% confidence interval from roughly 0.28 to 0.60; depression alone gives about 0.55 but with a far wider interval, and adding the precise dementia estimate pulls the pool to about 0.46 while tightening it. Each condition added sharpens the overall picture, which is exactly why a broad but thinly-evidenced field depends on synthesis, and equally why a synthesis of biased trials narrows the interval around a number that may itself be too high.

Discussion

Music therapy occupies an unusual position in clinical science: a practice whose founding premise, that organised sound can heal, has grown more rather than less credible as neuroscience has mapped music onto the brain's emotion, reward, and motor systems, even as its outcome evidence, though broad, has remained thinner in any single condition than that premise deserves. The history divides into the expressive and the empirical. For its first decades the field was carried by the clinical imagination of Gaston, Nordoff and Robbins, and Bonny, justified by case and conviction; over the last three decades it has submitted itself to systematic review and, latterly, to mechanistic dissection, with results that are encouraging in aggregate and humbling in the fragility of any one trial. Table 1 sets the two great channels through which the therapy is thought to act side by side.

Read as a whole, music therapy is a case study in a practice catching up with its own theory, and in some measure overtaking it. Its expressive founders were right about music's power earlier than the science could confirm, and the current task is neither to celebrate the intuition nor to dismiss the uneven evidence, but to run the larger trials and mechanism studies that would let the therapy's real, moderate, and condition-specific benefits be stated with the confidence its clearest advocates already feel. Until then the honest summary is the one the reviewers themselves give: a broadly useful treatment, moderately supported, with the strongest cases for depression, dementia, and rhythmic rehabilitation, and much still to prove.

Glossary

Active music therapy.

Methods in which the client makes music, by singing, playing, or improvising with the therapist, so that the musical interaction itself is the treatment; contrasted with receptive methods.

Auditory-motor entrainment.

The spontaneous synchronisation of movement to a regular auditory pulse, arising from reciprocal connections between the auditory and motor systems; the basis of rhythmic gait rehabilitation.

Bonny Method.

Helen Bonny's receptive method, Guided Imagery and Music, in which a relaxed client listens to sequenced classical programmes and follows the imagery and feeling that arise.

Creative arts therapies.

The family of treatments that use an artistic medium as the means of psychological change, comprising music, art, dance/movement, and drama therapy.

Creative music therapy.

The improvisational approach developed by Nordoff and Robbins, in which therapist and client make music together in real time and the evolving musical relationship carries the therapeutic work.

Dose-response relationship.

The finding that benefit grows with the number of sessions, up to a saturating ceiling; clinically an argument for adequate dose, methodologically a warning that under-dosed trials can miss a real effect.

Effect size.

A standardized measure of a treatment's benefit, here the difference between treated and control means in standard deviations, allowing results to be pooled across studies in meta-analysis.

Music medicine.

The use of pre-recorded music for listening, typically delivered by medical staff rather than a therapist; distinguished from music therapy proper, which involves a trained clinician and a therapeutic relationship.

Music-evoked emotion.

The affective response to music, shown by neuroimaging to engage the amygdala, nucleus accumbens, and orbitofrontal cortex, the same limbic and reward structures that underlie emotion generally.

Neurologic Music Therapy.

A standardised, neuroscience-grounded family of techniques for sensorimotor, speech, and cognitive rehabilitation, specified at the level of the brain systems each engages.

Receptive music therapy.

Methods in which the client listens to music the therapist selects or performs, often to induce a state, evoke imagery, or support relaxation; contrasted with active methods.

Rhythmic Auditory Stimulation.

A neurologic music therapy technique that uses a steady auditory beat to entrain and regulate gait, applied in rehabilitation after stroke and in Parkinson's disease.

Standardized mean difference.

An effect size computed as the difference between two group means divided by their pooled standard deviation, the common currency in which music-therapy trials are pooled.

Systematic review.

A structured synthesis of all eligible studies on a question, the Cochrane review being its most rigorous form; in a field of small trials it is the principal means of estimating an average effect.

Key Researchers

Helen L. Bonny (1921-2010). Music therapist and researcher who developed the Bonny Method of Guided Imagery and Music, drawing music therapy into the study of consciousness and inner experience. Wikipedia - Wikidata

Joke Bradt. Professor of creative arts therapies at Drexel University; lead author of the Cochrane reviews of music interventions for cancer patients and for anxiety in coronary heart disease. ORCID - Faculty Page - Google Scholar

E. Thayer Gaston (1901-1970). Music educator and researcher at the University of Kansas, often called the father of American music therapy for his role in building the profession's training and research. Wikipedia - Wikidata

Christian Gold. Music-therapy researcher at NORCE and the University of Vienna; the field's leading meta-analyst, whose work established the dose-response relationship and the Cochrane depression and autism reviews. ORCID - Faculty Page - Google Scholar

Stefan Koelsch. Neuroscientist of music at the University of Bergen; author of the influential account of the brain correlates of music-evoked emotion. ORCID - Faculty Page - Wikipedia

Wendy L. Magee. Professor of music therapy at Temple University; researcher on music therapy in neurorehabilitation, disorders of consciousness, and music technology. ORCID - Faculty Page - Google Scholar

Alfredo Raglio. Music-therapy researcher at the Istituti Clinici Scientifici Maugeri and the University of Pavia; investigator of music therapy in dementia and neurological disorders. ORCID - Faculty Page

Teppo Särkämö. Cognitive neuroscientist at the University of Helsinki; lead author of the landmark trial showing that music listening enhances cognitive recovery and mood after stroke. ORCID - Faculty Page - Google Scholar

Michael H. Thaut. Music neuroscientist at the University of Toronto; founder of Neurologic Music Therapy and Rhythmic Auditory Stimulation, and author of the account of rhythmic entrainment and the motor system. ORCID - Faculty Page

Frequently Asked Questions

What is music therapy?
It is the evidence-based clinical use of music by a credentialed professional to accomplish individualised health goals within a therapeutic relationship; a recognised form of psychotherapy and one of the creative arts therapies, requiring no musical training from the client (Aalbers et al., 2017).

How is music therapy different from just listening to music?
Music therapy involves a trained therapist using music toward specific clinical goals within a relationship, whereas simply listening to pre-recorded music, often called music medicine, has no therapist and no individualised plan; reviews distinguish the two because they are different interventions with different effects (Bradt et al., 2016).

Who founded music therapy?
The modern profession took shape in the mid-twentieth century; E. Thayer Gaston helped establish its academic and professional footing, while Paul Nordoff and Clive Robbins developed creative music therapy and Helen Bonny developed Guided Imagery and Music.

Does music therapy actually work?
Cochrane reviews find genuine benefit across several conditions, most clearly for depression and for the neuropsychiatric symptoms of dementia, though many trials are small and the certainty of the evidence is often low to moderate (Aalbers et al., 2017; van der Steen et al., 2018).

What conditions is music therapy used for?
Its evidence base spans depression, dementia, autism spectrum disorder, stroke and neurological rehabilitation, anxiety in cardiac and cancer care, and the care of premature infants, among others (Geretsegger et al., 2014; Standley, 2002).

How does music therapy work in the brain?
Through two main channels: music engages the limbic and reward systems that generate emotion, allowing mood and stress to be regulated, and a steady beat entrains the motor system, allowing movement to be organised in time (Koelsch, 2014; Thaut et al., 2015).

Does the number of sessions matter?
Yes; a dose-response relationship means benefit tends to grow with the number of sessions up to a point, so an adequate course is important and trials with too few sessions may under-estimate the treatment's effect (Gold et al., 2009).

Can music help after a stroke?
Yes; a controlled trial found that daily music listening in early stroke recovery improved verbal memory and focused attention and lifted mood compared with audiobooks or no listening, and rhythmic auditory stimulation is used to rehabilitate gait (Särkämö et al., 2008).

References