The trial randomly assigned 60 children to each of three groups: embodied music, non-embodied music or usual activity. The two active programs lasted 12 weeks and each included three 45-minute sessions per week. Their musical exposure, session duration, frequency and instructor contact were matched, but the embodied program added structured physical engagement. Each embodied session combined warm-ups, beat synchronization, rhythm-to-movement mapping, inhibitory-control games, rule switching, paired or group synchronization and a short cool-down. Children clapped, stepped, used body percussion, stopped or reversed actions in response to cues and adjusted to changes in rhythm, tempo, instrument or rules. The program was designed as perception-action and executive-control training rather than dance instruction.

The non-embodied control offered guided listening, singing, pitch and melody discrimination, rhythm identification, musical storytelling and discussion. Whole-body locomotion, sustained sensorimotor synchronization and movement-based rule switching were minimized, allowing the comparison to test whether structured bodily engagement added benefits beyond musical participation itself. Children in the usual-activity group continued their normal school schedules without an additional music program. Outcome assessors were blinded to allocation, while instructors and fidelity observers could recognize the visibly different sessions. Assessments occurred at baseline, week 6, week 12 and a three-month follow-up; data from 166 children, or 92.2% of the randomized sample, were available at the final follow-up, and mixed-effects analyses retained all 180 participants using their available observations.

Executive function, combining inhibition, working memory and cognitive flexibility, showed a significant group-by-time interaction: χ²(6) = 78.60, p < 0.001. From baseline to the end of the intervention, the model estimated an increase of 10.6 points in the embodied group, compared with 4.2 points in the non-embodied music group and 0.6 points in controls. The embodied group’s advantage over non-embodied music was 6.40 points, with a 95% confidence interval of 3.80 to 9.00, while its advantage over control was 10.00 points, with a 95% confidence interval of 7.40 to 12.60. The post-intervention standardized difference between embodied music and control was d = 1.36. Within the embodied condition, inhibition rose from 50.4 to 61.8, cognitive flexibility from 50.3 to 61.1 and working memory from 49.9 to 58.7 on T-scored indices. These component changes were reported descriptively rather than tested individually, so only the composite supported the planned inferential comparison. The large composite estimate requires independent replication before broad generalization.

Parent-reported emotion regulation also improved more after embodied music, with a group-by-time interaction of χ²(6) = 51.40, p < 0.001. The embodied group gained 8.8 points from baseline to week 12, versus 4.5 points in the non-embodied group and 0.7 points in controls. Its change exceeded that of non-embodied music by 4.30 points, with a 95% confidence interval of 2.00 to 6.60, and control by 8.10 points, with a 95% confidence interval of 5.70 to 10.50. The standardized embodied-versus-control difference at the post-intervention assessment was d = 1.20. Movement-rhythm synchronization, a secondary outcome, produced the largest group-by-time effect: χ²(6) = 129.30, p < 0.001. The embodied group increased by 15.0 points, compared with 4.1 points in the non-embodied group and 0.3 points in controls, producing a d of 1.88 relative to control. The unusually large synchronization effect is especially notable because this was the outcome most closely tied to the activities practiced.

More distal outcomes showed a weaker pattern. Psychological well-being, measured with the KIDSCREEN-10 health-related quality-of-life index, had a significant group-by-time interaction, χ²(6) = 27.80, p < 0.001. Embodied music improved by 6.1 points from baseline to week 12, versus 2.6 points in the non-embodied group and 0.5 points in control. Its advantage over non-embodied music was 3.50 points, with a 95% confidence interval of 1.10 to 5.90, and the standardized difference from control was d = 0.68. Resting RMSSD, a time-domain index of heart-rate variability, also showed a group-by-time interaction, χ²(6) = 18.60, p = 0.005. It rose by 7.1 milliseconds in the embodied group, 2.6 milliseconds in the non-embodied group and 0.5 milliseconds in controls, with an embodied-versus-control difference of d = 0.62. Because RMSSD is sensitive to breathing, posture, activity, recording conditions and processing choices, the investigators treated it as a convergent physiological observation rather than evidence for one specific regulatory mechanism.

Sleep problems changed less than the cognitive, emotional and physiological outcomes. The overall group-by-time interaction was significant, χ²(6) = 15.20, p = 0.019, and the embodied group’s reduction from baseline to week 12 exceeded the control-group reduction by 2.90 points, with a 95% confidence interval of −4.80 to −1.00. However, the embodied-versus-non-embodied contrast was not significant after adjustment for the planned pairwise comparisons. Its value was −1.70 points, with a 95% confidence interval of −3.60 to 0.20. The descriptive standardized difference between embodied music and control at week 12 was d = 0.52. The pattern broadly tracked distance from the intervention: synchronization and executive control showed the largest differences, while well-being, RMSSD and sleep showed smaller ones. The researchers cautioned that this ordering does not establish a layered developmental pathway, particularly because effects can also depend on measurement reliability, scale dispersion and how closely an assessment resembles the trained behavior.

Most gains were still visible three months after the sessions stopped, though they had attenuated from their post-intervention peaks. Relative to baseline, the embodied group retained increases of 9.3 points in executive function, 7.9 points in emotion regulation and 13.1 points in synchronization at follow-up. The results are consistent with partial short-term maintenance, but the study did not compare continued training with booster sessions or withdrawal of practice. Attendance was similar in the two active groups, averaging 91.8% for embodied music and 92.4% for non-embodied music, and mean fidelity across active sessions was 94.6%. Independent observers rated 86 of 432 sessions, with an intraclass correlation of 0.92. Four children in the embodied group and one in the non-embodied group reported transient soreness or fatigue that resolved without treatment; no serious adverse events, medically treated injuries or harm-related withdrawals occurred.

Exploratory change-score correlations were consistent with the idea that movement-rhythm synchronization may participate in broader self-regulation. Synchronization change correlated with executive-function change at r = 0.48, emotion-regulation change at r = 0.36 and well-being change at r = 0.27. An indirect-effect model yielded a standardized serial indirect effect of 0.022, with a bootstrap 95% confidence interval of 0.009 to 0.043, linking embodied-versus-non-embodied assignment through synchronization, executive function and emotion regulation to later well-being. These results cannot demonstrate mediation, however, because the proposed mediators were not experimentally manipulated and their changes substantially overlapped in time. A residual association with well-being also remained. Other unmeasured factors, including enjoyment, physical activity, peer interaction, competence and belonging, could contribute to the observed relationships, so the statistical pathway should not be interpreted as a demonstrated causal sequence.

The study has important constraints. It included typically developing children from six public primary schools in one Chinese city, limiting generalizability to other ages, clinical populations and educational systems. Although the embodied package combined several ingredients—movement, rhythm, inhibition, rule switching and social coordination—the trial cannot isolate the effect of any single component. The program was not prospectively registered, complete double blinding was not feasible, and the power calculation did not explicitly model school-by-condition clustering. Questionnaire-based well-being and sleep results may also be affected by reporting bias. Mixed-effects models assumed missing data were missing at random, an assumption that cannot be verified from the observed data alone, and alternative missing-not-at-random scenarios were not tested. Subject to those limitations, the randomized active control provides evidence that adding structured bodily engagement to school music can strengthen gains in several proximal self-regulatory outcomes in children, while the unusually large effects on synchronization and executive function remain in need of independent confirmation.

Subject of Research: Randomized trial of movement-based music training and self-regulation outcomes in children aged 8–11 years

Article Title: Effects of an embodied music intervention on executive function, emotion regulation and psychophysiological well-being in children: the role of movement-rhythm synchronization and self-regulation

Article References: Shu, Q. (2026). Effects of an embodied music intervention on executive function, emotion regulation and psychophysiological well-being in children: the role of movement-rhythm synchronization and self-regulation. Frontiers in Psychology, 17, Article 1988460. https://doi.org/10.3389/fpsyg.2026.1988460

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1988460

Keywords: embodied music, children, randomized controlled trial, executive function, emotion regulation, movement-rhythm synchronization, self-regulation, psychological well-being, heart-rate variability, school intervention, Effects, embodied