Music-based interventions in adult stroke rehabilitation: a narrative review of clinical applications, proposed mechanisms, and implementation.

BACKGROUND: Music is used in stroke rehabilitation through therapist-delivered music therapy, neurologic music therapy techniques, other structured music-based interventions, and technology-assisted systems. Inconsistent terminology and overstated mechanisms complicate interpretation. OBJECTIVES: To synthesize the clinical applications, proposed mechanisms, limitations, and implementation considerations of music-based interventions in adult stroke rehabilitation. METHODS: We conducted a targeted narrative review of peer-reviewed English-language literature identified through PubMed/MEDLINE, CINAHL, and reference chaining. We prioritized adult stroke systematic reviews, controlled studies, mechanistic investigations, and implementation research through June 2026. Seminal pre-2010 studies and evidence from other neurological populations were used only to explain mechanisms or gaps and are labeled as indirect. We did not undertake exhaustive retrieval, formal risk-of-bias assessment, or quantitative pooling. RESULTS: Evidence is most consistent for rhythmic auditory stimulation for gait and selected music-supported upper-limb programs. Melodic intonation therapy may improve trained utterances and repetition in nonfluent aphasia, but transfer to everyday communication is uncertain. Listening and singing programs show potential for cognition, mood, engagement, and sleep, although protocols are heterogeneous. Post-stroke swallowing evidence is preliminary, and blood-brain barrier effects remain preclinical. Outcomes may vary with stroke phase, phenotype, cognition, hearing, comorbidity, musical reward and preference, dose, and provider expertise. CONCLUSIONS: Music-based rehabilitation comprises mechanism-informed adjuncts rather than one treatment. Clinical use should link a defined musical ingredient to a functional target, document provider and dose, incorporate preference and safety screening, and measure everyday transfer. Larger pragmatic and mechanism-specific trials with prespecified moderator analyses are needed.