A 2014 study had adults in their sixties learn quilting or digital photography for 16 hours a week, and found real memory gains that a large music-training meta-analysis could not replicate in children

The claim that music lessons make children cognitively sharper across the board has circulated for decades, and it is one of the more thoroughly tested claims in this area of psychology. A 2017 meta-analysis by Giovanni Sala and Fernand Gobet, covering 38 studies and more than 3,000 participants, found the overall effect of music training on children’s broader cognitive and academic skills was small, and that the effect shrank as study quality improved. Where studies used weaker designs, without proper control groups or randomisation, the apparent benefit was larger. Where studies were rigorously controlled, the benefit to general cognition mostly disappeared. The researchers concluded that far transfer, the idea that a specific skill like music training reliably improves unrelated cognitive abilities, was not well supported by the evidence once methodological quality was accounted for.

This is one meta-analysis, not the final word on every possible cognitive benefit of musical training, and it specifically concerns transfer to children’s broader cognitive and academic skills rather than every possible outcome music training might produce. But it is a large, quality-weighted synthesis, and its central finding complicates the popular version of the claim considerably.

A different kind of study found something that did hold up

A separate and more rigorously designed piece of research points to a real effect, but locates it somewhere more specific than “creativity” as a general category. In a 2014 paper in Psychological Science, Denise Park and colleagues ran what became known as the Synapse Project. Adults in their sixties and seventies were randomly assigned to one of several conditions for an average of 16.5 hours a week over three months: learning digital photography, learning quilting, doing both, or, in the comparison conditions, socialising in a group without learning a new skill, or doing low-demand cognitive tasks alone with no social contact.

The results were specific. Participants in the photography and quilting conditions, the ones learning a genuinely new, sustained, cognitively demanding skill, showed improved episodic memory compared with participants in the social-only or low-demand conditions. A related follow-up analysis from the same research group found the pattern held up when the researchers looked more closely at consistency of performance across the sample, not just average scores.

What the researchers think was actually doing the work

The Synapse Project researchers were specific about what they believe produced the effect, and it was not the artistic content of quilting or photography as such. Their explanation centres on sustained, effortful, productive engagement, learning something that genuinely challenges working memory, sustained attention, and problem-solving over an extended period, rather than passive consumption or low-demand repetition. Socialising with a group, while valuable for other reasons, did not produce the same memory gains on its own. Neither did solitary low-demand mental activity. What mattered was the combination: something new, something difficult enough to require real learning, sustained over months rather than days.

That detail reframes what “creative skills that are good for your brain” most plausibly means. The evidence does not clearly support treating creativity itself as a special cognitive ingredient distinct from other kinds of demanding learning. What the Synapse Project actually tested were two specific creative skills, and what its own authors credit for the result is the sustained cognitive demand those skills happened to provide, not something unique to artistic expression as a category.

Why these two findings are not actually in conflict

Read together, the music-training meta-analysis and the Synapse Project are not contradicting each other so much as pointing at different mechanisms. Sala and Gobet’s null result concerns a specific, narrower question: whether music lessons, often taken for a fixed weekly hour alongside a child’s ordinary school routine, reliably transfer to broad cognitive skills, in a population, children, where confounding factors such as family background and existing academic engagement are notoriously difficult to separate from the effect of the lessons themselves. The Synapse Project tested something more intensive and more controlled: adults randomly assigned to genuinely sustained, high-demand learning, with proper comparison groups built in from the start.

The likely honest reconciliation is that the dose and rigour of engagement matter more than the specific creative activity chosen. An hour of weekly music lessons alongside an otherwise ordinary routine looks, on the evidence, unlikely to reliably transform a child’s general cognition. Sixteen hours a week of sustained, difficult, novel skill acquisition, of the kind the Synapse Project’s adult participants did, produced a measurable memory benefit relative to less demanding alternatives.

What this suggests for choosing a skill to learn

For an adult weighing whether to take up a new creative pursuit for the sake of cognitive benefit rather than simple enjoyment, the research suggests the choice of activity matters less than how demanding and sustained the engagement actually is. Photography and quilting were the specific skills tested in the Synapse Project, but the researchers’ own account of the mechanism suggests other sufficiently difficult, sustained, novel pursuits, learning an instrument seriously as an adult beginner rather than as a child taking weekly lessons, taking up pottery, learning to code, could plausibly produce a similar kind of benefit, though this specific extension has not itself been directly tested.

What this does not prove

The Synapse Project measured adults in their sixties and seventies over three months, and it is not established how far its findings extend to younger adults, to shorter or less intensive engagement, or to creative activities not directly tested in the study. The music-training meta-analysis concerned children specifically, not adults learning new skills later in life, so the two findings should not be treated as directly interchangeable evidence about the same population. Neither line of research supports the idea that any creative hobby, regardless of how much genuine cognitive demand it places on the learner, will reliably improve memory or broader cognition. What the combined evidence does support is a narrower, more useful claim than the popular one: it is sustained, effortful engagement with something genuinely new and difficult that the research links to measurable cognitive benefit, and creative skills are one well-tested way, among others, of providing exactly that kind of demand.