That's a hard problem, with no reasonable way to measure it in in a large population in sight, or even direction of the relationship taken into account. Ideally you'd take a bunch of kids and look at their brains and then see how they grew up and see whether you could find anything that altered the distribution in similar cases - but ....

Well, you see the problem? It's a sort of twiddling your thumbs style studying, rather than addressing more immediate problems that might do something at a reasonable price/timeline.