Artificial intelligence-generated and algorithmically personalized music interventions and their effects on brain function, behavior, and health: a critical review
- 1chool of Education, Jilin International Studies University, Changchun, Jilin, China
Abstract
This critical review synthesizes empirical studies of algorithmically personalized and artificial-intelligence-generated music interventions reporting neurophysiological, neuroimaging, or objective behavioral outcomes. Literature was identified through structured searches of PubMed, Scopus, Google Scholar, IEEE Xplore, and the ACM Digital Library through 12 August 2026, with CNKI and Wanfang searched on 22 September 2026, supplemented by citation tracking. Twenty-one eligible studies (at least 1,005 participants) were identified. Five things the term artificial intelligence covers in this literature are separated: algorithmic selection from a corpus, rule-based sonification, parametric modulation of pre-existing material, machine-learning generative composition, and closed-loop delivery contingent on the listener’s own signal. Only three studies used a machine-learning generative model, so most of the evidence concerns algorithmic personalization of existing material. Studies of closed-loop neural systems reported the most consistent neurophysiological signal, and music-based alpha neurofeedback produced higher normalized alpha power than yoked random feedback. In an open-loop design, personalized EEG-derived music was reported to attenuate anxiety induction where self-selected preferred music and silence did not, with enhanced prefrontal theta oscillations as the associated correlate. Eight studies carried a comparator capable of isolating an algorithmic component from music exposure, and all eight favored the algorithmic condition. These eight address different causal questions, use different designs, and point to no single mechanism, so their results are best read as preliminary and hypothesis-generating. Evidence quality was very low throughout. The median reported sample size was 22, no study blinded participants or measured expectancy, and no specific finding has been independently replicated. GRADE certainty was very low for every outcome domain, covering EEG oscillatory power, sleep architecture, autonomic regulation, epileptiform activity, and cognitive, behavioral and affective outcomes. Neurophysiological change is reported more often than behavioral change, and behavioral change more often than clinical benefit. The evidence supports continued investigation and does not support clinical deployment. Eight reporting and design requirements for future trials are proposed.
Related
- Converging Paths: Artificial Intelligence for Integrating Biological and Digital Biomarkers in Personalized Longevity Medicine
- Self-Generated Goals in the Age of Foundation Models: Artificial Autotelic Agents and Their Implications for Cognitive Science
- Growing up with artificial intelligence: An amplifier of inequalities and a public health issue
- Notification: Artificial Intelligence in Autism Spectrum Disorder Diagnosis: A Scoping Review of Face, Voice, and Text Analysis Methods.
- Early Detection of Red Palm Weevil Infestation: A Systematic Review of Sensing Technologies, Artificial Intelligence, and Field Readiness
BibTeX
@article{jin2026artificial,
title = {{Artificial intelligence-generated and algorithmically personalized music interventions and their effects on brain function, behavior, and health: a critical review}},
author = {Jingting Jin},
journal = {Frontiers in Human Neuroscience},
year = {2026},
month = oct,
doi = {10.3389/fnhum.2026.1967909},
url = {https://doi.org/10.3389/fnhum.2026.1967909},
note = {CC BY},
}
RIS
TY - JOUR
TI - Artificial intelligence-generated and algorithmically personalized music interventions and their effects on brain function, behavior, and health: a critical review
AU - Jingting Jin
PY - 2026
DA - 2026/10/23
JO - Frontiers in Human Neuroscience
DO - 10.3389/fnhum.2026.1967909
UR - https://doi.org/10.3389/fnhum.2026.1967909
AB - This critical review synthesizes empirical studies of algorithmically personalized and artificial-intelligence-generated music interventions reporting neurophysiological, neuroimaging, or objective behavioral outcomes. Literature was identified through structured searches of PubMed, Scopus, Google Scholar, IEEE Xplore, and the ACM Digital Library through 12 August 2026, with CNKI and Wanfang searched on 22 September 2026, supplemented by citation tracking. Twenty-one eligible studies (at least 1,005 participants) were identified. Five things the term artificial intelligence covers in this literature are separated: algorithmic selection from a corpus, rule-based sonification, parametric modulation of pre-existing material, machine-learning generative composition, and closed-loop delivery contingent on the listener’s own signal. Only three studies used a machine-learning generative model, so most of the evidence concerns algorithmic personalization of existing material. Studies of closed-loop neural systems reported the most consistent neurophysiological signal, and music-based alpha neurofeedback produced higher normalized alpha power than yoked random feedback. In an open-loop design, personalized EEG-derived music was reported to attenuate anxiety induction where self-selected preferred music and silence did not, with enhanced prefrontal theta oscillations as the associated correlate. Eight studies carried a comparator capable of isolating an algorithmic component from music exposure, and all eight favored the algorithmic condition. These eight address different causal questions, use different designs, and point to no single mechanism, so their results are best read as preliminary and hypothesis-generating. Evidence quality was very low throughout. The median reported sample size was 22, no study blinded participants or measured expectancy, and no specific finding has been independently replicated. GRADE certainty was very low for every outcome domain, covering EEG oscillatory power, sleep architecture, autonomic regulation, epileptiform activity, and cognitive, behavioral and affective outcomes. Neurophysiological change is reported more often than behavioral change, and behavioral change more often than clinical benefit. The evidence supports continued investigation and does not support clinical deployment. Eight reporting and design requirements for future trials are proposed.
SN - 1662-5161
ER -