Giving a toddler puzzles does something narrower than the marketing claims: in a University of Chicago study, children who played with them between two and four had better spatial skills by four and a half, a specific ability, tied to later success in STEM

More puzzle play, and more complex puzzles, went with stronger spatial performance. It remains possible that the sort of home that produces puzzle play also produces spatial skill through some other route entirely. The lead author said as much at the time, noting that further work would be needed to show whether puzzle play actually causes the gains. It is closer to “young children who play with puzzles tend to have better spatial skills, for reasons not yet pinned down.” Spatial skill, they argued, is a talent that standard testing tends to miss.