Effects of music training on executive functions in preschool children aged 3–6 yearssystematic review and meta-analysis

Lu Yanan, Lin Shi, Ahmad Faudzi Musib•View original

OverviewBalancedshimmer voice

If you've ever watched a preschooler try to wait their turn, remember a simple rule, or switch from one game to another without melting down, you've seen executive functions at work. These are the brain's air traffic control systems for goal-directed behavior. Psychologists usually talk about three key components. Inhibitory control is pausing or suppressing a tempting response. Working memory involves holding and juggling information over a short stretch. And cognitive flexibility is shifting perspectives and adapting when the rules change. During the preschool years, these functions are still wiring up. By around age four, many kids start to handle basic inhibition tasks better, and flexibility really comes online between the ages of three and five, continuing to improve after that. Under the hood, these skills correspond to maturing prefrontal circuits. Lateral prefrontal regions carry much of the working memory load, while medial areas help with flexible switching, and the inferior frontal cortex supports inhibition. That neural scaffolding is still under construction, which is exactly why early childhood represents such a powerful window for development. Here's the big educational question: can we nudge that development with experiences that are structured, regular, and fun? Music has been a prime candidate. Think about what making music demands. You focus on sound, keep time, sequence actions, store little chunks of melody in mind, and ignore distractions to stay with the beat. It's attention, memory, and control, all braided together. Some studies have linked music activities to better executive functions in young kids, but the findings have been mixed. Different ages, different tasks, and different programs. One persistent gap has been dose. Do the duration, the weekly frequency, and how long each session runs actually shape the gains? That's where a recent systematic review and meta-analysis by Lu, Shi, and Musib comes in. They set a tight scope—preschoolers ages three to six—and a clear question: do group-based music activities improve inhibitory control, working memory, and cognitive flexibility, and how does training dose modulate those effects? They were thorough. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines that Page and colleagues popularized, they registered a protocol in the International Prospective Register of Systematic Reviews, known as PROSPERO, so the plan was public from the start. They cast a wide net across four major databases—Web of Science, PubMed, Scopus, and the China National Knowledge Infrastructure—searching from each database's inception up through early September of 2023. They used English and Chinese terms for music activities and preschool ages, paired with executive function keywords, and even traced references in earlier reviews to catch overlooked studies. Eligibility was well-defined and practical. Kids had to be three to six years old and typically developing. The intervention had to be group music—singing, playing, composing, practicing rhythm or pitch—delivered with some structure. Controls could be passive or active but not music-based. At least one executive function outcome had to be measured: inhibition, working memory, or flexibility. They included randomized and nonrandomized comparative designs, but they excluded specialized instrumental or vocal training that looked more like conservatory preparation than preschool circle time. On the back end, they collected pre- and post-test means and standard deviations—sometimes even extracting values from figures with a tool called WebPlotDigitizer—cataloguing who was studied and how, and assessing study quality with the Cochrane risk of bias framework outlined by Higgins and Green. Blinding is notoriously difficult in educational interventions, and this field is no exception; only a handful of studies managed to blind outcome assessment. Some used solid randomization, while others relied on more casual assignments, and reporting on allocation concealment was thin. The positive note: selective reporting did not appear rampant, and almost all studies reported complete outcome data. For those interested in statistics, here's what they did. They used a random effects model, essentially assuming the true effect might differ across studies due to variations in programs and samples. They calculated standardized mean differences to compare across different tasks and scales. Following guidance from Larry Morris on handling pre and post designs, the change score was the post-test minus the pre-test, and the pre-test variability anchored the standardization. They quantified heterogeneity with the I-squared statistic and treated 25, 50, and 75 percent as low, moderate, and high. They also checked for publication bias using Egger's test. Moderator analyses were planned in advance, looking at how many weeks the program ran, how many sessions per week, and how long each session lasted. Out of nine hundred forty-two records, fifty-four made it to full-text screening. Twelve studies cleared the bar for the review, and ten had the necessary data for the meta-analysis. In total, eight hundred seven children were represented—four hundred forty-nine in music programs and three hundred fifty-eight in controls—drawn from studies in the United States, Germany, China, the United Kingdom, and Canada. Two studies, including work by Bolduc and by Moreno, contributed qualitatively but were not included in the pooled statistics. So what changed for kids who participated in music? Across the three executive function domains, the headline is a consistent, small to moderate advantage for music groups over controls. Inhibitory control showed the clearest improvement, with a standardized mean difference of 0.38, and the confidence interval was solidly above zero. Working memory wasn't far behind at 0.35, again clearly positive. Cognitive flexibility had a smaller but still significant effect at 0.23. If you like Cohen's benchmarks, those first two hover in the small to moderate range, with the third falling in the small but reliable zone. There was some variability across studies—heterogeneity was substantial for inhibition, moderate for working memory, and lower for flexibility—which isn't surprising given the differences in music classes and assessment tasks. When they checked for publication bias, the tests did not raise red flags for two outcomes and were borderline for flexibility, which argues against a simple story of "only positive studies get published" driving the results. Now we get to the part that educators and program designers really want to know: dose. Longer wasn't always better for everything, but a pattern emerges if you look at the domains individually. Start with inhibition. Programs that ran for twelve weeks or more produced larger gains relative to controls, around half a standard deviation on average. Meeting more often helped as well—three or more sessions per week increased the effect into a similar range. There was a sweet spot for session length: twenty to thirty minutes per session was where the gains were most consistent. Stretching past forty minutes did not add benefit for inhibition; in fact, those longer sessions showed effects that were imprecise and not clearly different from zero. If you've watched a room full of four-year-olds after half an hour on the rug, that ceiling makes intuitive sense. Working memory told a slightly different story. Again, programs that lasted twelve weeks or longer outperformed shorter ones, showing a clear boost. Session length in the twenty to thirty minute window also correlated with stronger gains. However, frequency flipped the expectation a bit. Fewer than three sessions a week actually showed a larger effect, while three or more sessions were positive but smaller. That might sound odd, but think about consolidation and variety. Working memory could benefit from spaced, focused practice rather than daily repetition. It's merely a hypothesis, not a conclusion, but it fits the data pattern. Cognitive flexibility, the shape-shifting skill, seemed to require a different approach. Longer programs were beneficial, and meeting at least three times a week nudged the effect in the right direction, but what really stood out was session length. When classes ran for over forty minutes, the flexibility boost became what most would consider a large effect. That's a considerable leap, based on a smaller subset of studies, so treat it with caution. Still, it suggests that to practice rule switching, perspective taking, and set shifting in music—consider switching between rhythmic patterns, call-and-response with changing rules, or moving from one instrumental role to another—you may need more continuous time during a session for the brain to stretch and adapt. A quick pause for interpretation. These moderator analyses slice the same set of studies into subgroups, with the exact activities varying widely—rhythm drills, pitch games, movement with music, or combinations of all three. The tasks used to measure outcomes weren't uniform either, and for each domain, the analysts selected the most common task available to keep comparisons cleaner. All that contributes to the heterogeneity. So, treat the dose guidance as patterns and not prescriptions set in stone. Zooming back out, the overall picture is encouraging but not a blanket endorsement. Small to moderate improvements in key executive skills, delivered through an activity kids enjoy, are meaningful at scale—especially during the period when those prefrontal circuits are most malleable. At the same time, the evidence base has soft spots. Blinding participants to a music class isn't feasible, but blinding those who score the tests sometimes is, and only four studies managed that. Randomization and allocation concealment were inconsistently described, with some studies depending on quasi-random assignments. Such choices can influence the results. On the plus side, selective reporting did not appear to be the main driver, and almost every study reported complete outcome data. It's also important to note what was not included. Highly specialized instrumental lessons were excluded from this review, focusing instead on group music activities that naturally fit into preschools—such as singing, clapping games, simple instruments, and playful composition. That matters for generalizability. If you're a teacher designing a curriculum, the takeaway is actionable: create a consistent three-month arc; meet two to three times each week; aim for twenty to thirty minutes to target inhibition and working memory; and if you want to challenge flexibility, consider including periodic longer sessions where kids can really explore different musical roles. Methodologically, there are several behind-the-scenes choices that strengthened this synthesis. Using change scores standardized by pre-test variability levels the field across disparate measures. Random effects modeling acknowledges that "music class" isn't a single thing and allows the average effect to vary across contexts. Checking for publication bias—finding no clear indication in two of the three domains—boosts confidence that we aren't just observing a mirage formed solely from the most positive studies. Where does this leave us? With a grounded, optimistic answer to the initial question. Yes, regular, structured music activities can provide preschoolers with a measurable boost in the very skills that help them focus, remember, and adapt. The gains aren't magical, but they're real, and they align with what we know about the developing brain. They also depend on how the program is designed. Dose matters, and it impacts different skills in various ways. Two closing thoughts about what comes next, framed by what Lu, Shi, and Musib actually advocate. First, better trials. This means tighter randomization, clearer reporting, and, whenever possible, blinding the assessors. It also involves standardizing core elements of the music curriculum so that future meta-analyses aren't trying to average apples, oranges, and the occasional tambourine. Second, mechanism. A few well-designed studies pairing music classes with brain measures—event-related potentials to track attention control or functional magnetic resonance imaging to observe shifts in prefrontal engagement—could reveal whether the behavioral gains rely on the neural systems we suspect. In the meantime, you don't have to wait for a lab to approve a drum circle. The science so far suggests: sing the songs, keep the beat, and play the switching game where the rules change with a clap. Do it regularly. Keep most sessions crisp, and once in a while, allow them to run longer. You're not just making music; you're helping to build the mind's conductor.

Inhibitory control is pausing or suppressing a tempting response. Working memory involves holding and juggling information over a short stretch. And cognitive flexibility is shifting perspectives and adapting when the rules change.

During the preschool years, these functions are still wiring up. By around age four, many kids start to handle basic inhibition tasks better, and flexibility really comes online between the ages of three and five, continuing to improve after that. Under the hood, these skills correspond to maturing prefrontal circuits.

Lateral prefrontal regions carry much of the working memory load, while medial areas help with flexible switching, and the inferior frontal cortex supports inhibition. That neural scaffolding is still under construction, which is exactly why early childhood represents such a powerful window for development.

Here's the big educational question: can we nudge that development with experiences that are structured, regular, and fun? Music has been a prime candidate. Think about what making music demands.

You focus on sound, keep time, sequence actions, store little chunks of melody in mind, and ignore distractions to stay with the beat. It's attention, memory, and control, all braided together. Some studies have linked music activities to better executive functions in young kids, but the findings have been mixed.

Different ages, different tasks, and different programs. One persistent gap has been dose. Do the duration, the weekly frequency, and how long each session runs actually shape the gains?

That's where a recent systematic review and meta-analysis by Lu, Shi, and Musib comes in. They set a tight scope—preschoolers ages three to six—and a clear question: do group-based music activities improve inhibitory control, working memory, and cognitive flexibility, and how does training dose modulate those effects? They were thorough.

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines that Page and colleagues popularized, they registered a protocol in the International Prospective Register of Systematic Reviews, known as PROSPERO, so the plan was public from the start. They cast a wide net across four major databases—Web of Science, PubMed, Scopus, and the China National Knowledge Infrastructure—searching from each database's inception up through early September of 2023. They used English and Chinese terms for music activities and preschool ages, paired with executive function keywords, and even traced references in earlier reviews to catch overlooked studies.

Eligibility was well-defined and practical. Kids had to be three to six years old and typically developing. The intervention had to be group music—singing, playing, composing, practicing rhythm or pitch—delivered with some structure.

Controls could be passive or active but not music-based. At least one executive function outcome had to be measured: inhibition, working memory, or flexibility. They included randomized and nonrandomized comparative designs, but they excluded specialized instrumental or vocal training that looked more like conservatory preparation than preschool circle time.

On the back end, they collected pre- and post-test means and standard deviations—sometimes even extracting values from figures with a tool called WebPlotDigitizer—cataloguing who was studied and how, and assessing study quality with the Cochrane risk of bias framework outlined by Higgins and Green. Blinding is notoriously difficult in educational interventions, and this field is no exception; only a handful of studies managed to blind outcome assessment. Some used solid randomization, while others relied on more casual assignments, and reporting on allocation concealment was thin.

The positive note: selective reporting did not appear rampant, and almost all studies reported complete outcome data.

For those interested in statistics, here's what they did. They used a random effects model, essentially assuming the true effect might differ across studies due to variations in programs and samples. They calculated standardized mean differences to compare across different tasks and scales.

Following guidance from Larry Morris on handling pre and post designs, the change score was the post-test minus the pre-test, and the pre-test variability anchored the standardization. They quantified heterogeneity with the I-squared statistic and treated 25, 50, and 75 percent as low, moderate, and high. They also checked for publication bias using Egger's test.

Moderator analyses were planned in advance, looking at how many weeks the program ran, how many sessions per week, and how long each session lasted.

Out of nine hundred forty-two records, fifty-four made it to full-text screening. Twelve studies cleared the bar for the review, and ten had the necessary data for the meta-analysis. In total, eight hundred seven children were represented—four hundred forty-nine in music programs and three hundred fifty-eight in controls—drawn from studies in the United States, Germany, China, the United Kingdom, and Canada.

Two studies, including work by Bolduc and by Moreno, contributed qualitatively but were not included in the pooled statistics.

So what changed for kids who participated in music? Across the three executive function domains, the headline is a consistent, small to moderate advantage for music groups over controls. Inhibitory control showed the clearest improvement, with a standardized mean difference of 0.38, and the confidence interval was solidly above zero.

Working memory wasn't far behind at 0.35, again clearly positive. Cognitive flexibility had a smaller but still significant effect at 0.23. If you like Cohen's benchmarks, those first two hover in the small to moderate range, with the third falling in the small but reliable zone.

There was some variability across studies—heterogeneity was substantial for inhibition, moderate for working memory, and lower for flexibility—which isn't surprising given the differences in music classes and assessment tasks. When they checked for publication bias, the tests did not raise red flags for two outcomes and were borderline for flexibility, which argues against a simple story of "only positive studies get published" driving the results.

Now we get to the part that educators and program designers really want to know: dose. Longer wasn't always better for everything, but a pattern emerges if you look at the domains individually. Start with inhibition.

Programs that ran for twelve weeks or more produced larger gains relative to controls, around half a standard deviation on average. Meeting more often helped as well—three or more sessions per week increased the effect into a similar range. There was a sweet spot for session length: twenty to thirty minutes per session was where the gains were most consistent.

Stretching past forty minutes did not add benefit for inhibition; in fact, those longer sessions showed effects that were imprecise and not clearly different from zero. If you've watched a room full of four-year-olds after half an hour on the rug, that ceiling makes intuitive sense.

Working memory told a slightly different story. Again, programs that lasted twelve weeks or longer outperformed shorter ones, showing a clear boost. Session length in the twenty to thirty minute window also correlated with stronger gains.

However, frequency flipped the expectation a bit. Fewer than three sessions a week actually showed a larger effect, while three or more sessions were positive but smaller. That might sound odd, but think about consolidation and variety.

Working memory could benefit from spaced, focused practice rather than daily repetition. It's merely a hypothesis, not a conclusion, but it fits the data pattern.

Cognitive flexibility, the shape-shifting skill, seemed to require a different approach. Longer programs were beneficial, and meeting at least three times a week nudged the effect in the right direction, but what really stood out was session length. When classes ran for over forty minutes, the flexibility boost became what most would consider a large effect.

That's a considerable leap, based on a smaller subset of studies, so treat it with caution. Still, it suggests that to practice rule switching, perspective taking, and set shifting in music—consider switching between rhythmic patterns, call-and-response with changing rules, or moving from one instrumental role to another—you may need more continuous time during a session for the brain to stretch and adapt.

A quick pause for interpretation. These moderator analyses slice the same set of studies into subgroups, with the exact activities varying widely—rhythm drills, pitch games, movement with music, or combinations of all three. The tasks used to measure outcomes weren't uniform either, and for each domain, the analysts selected the most common task available to keep comparisons cleaner.

All that contributes to the heterogeneity. So, treat the dose guidance as patterns and not prescriptions set in stone.

Zooming back out, the overall picture is encouraging but not a blanket endorsement. Small to moderate improvements in key executive skills, delivered through an activity kids enjoy, are meaningful at scale—especially during the period when those prefrontal circuits are most malleable. At the same time, the evidence base has soft spots.

Blinding participants to a music class isn't feasible, but blinding those who score the tests sometimes is, and only four studies managed that. Randomization and allocation concealment were inconsistently described, with some studies depending on quasi-random assignments. Such choices can influence the results.

On the plus side, selective reporting did not appear to be the main driver, and almost every study reported complete outcome data.

It's also important to note what was not included. Highly specialized instrumental lessons were excluded from this review, focusing instead on group music activities that naturally fit into preschools—such as singing, clapping games, simple instruments, and playful composition. That matters for generalizability.

If you're a teacher designing a curriculum, the takeaway is actionable: create a consistent three-month arc; meet two to three times each week; aim for twenty to thirty minutes to target inhibition and working memory; and if you want to challenge flexibility, consider including periodic longer sessions where kids can really explore different musical roles.

Methodologically, there are several behind-the-scenes choices that strengthened this synthesis. Using change scores standardized by pre-test variability levels the field across disparate measures. Random effects modeling acknowledges that "music class" isn't a single thing and allows the average effect to vary across contexts.

Checking for publication bias—finding no clear indication in two of the three domains—boosts confidence that we aren't just observing a mirage formed solely from the most positive studies.

Where does this leave us? With a grounded, optimistic answer to the initial question. Yes, regular, structured music activities can provide preschoolers with a measurable boost in the very skills that help them focus, remember, and adapt.

The gains aren't magical, but they're real, and they align with what we know about the developing brain. They also depend on how the program is designed. Dose matters, and it impacts different skills in various ways.

Two closing thoughts about what comes next, framed by what Lu, Shi, and Musib actually advocate. First, better trials. This means tighter randomization, clearer reporting, and, whenever possible, blinding the assessors.

It also involves standardizing core elements of the music curriculum so that future meta-analyses aren't trying to average apples, oranges, and the occasional tambourine. Second, mechanism. A few well-designed studies pairing music classes with brain measures—event-related potentials to track attention control or functional magnetic resonance imaging to observe shifts in prefrontal engagement—could reveal whether the behavioral gains rely on the neural systems we suspect.

In the meantime, you don't have to wait for a lab to approve a drum circle. The science so far suggests: sing the songs, keep the beat, and play the switching game where the rules change with a clap. Do it regularly.

Keep most sessions crisp, and once in a while, allow them to run longer. You're not just making music; you're helping to build the mind's conductor.