Well, the thing is, though, they shouldn't if there is a direct correlation between their measure of "plasticity" and cortical thickening. The latter has been shown to absolutely correlate with high IQ, and no need to chop the tails off of the distribution, either.
What I am wondering, though, is if they wanted to show a bimodal distribution with increasing frequency with increasing IQ, and the tail interfered with that hypothesis significantly because... they weren't really measuring a different developmental phenomenon the way they are proposing. This plasticity theory is the teacher's pet right now, but honestly, there is plenty of evidence to suggest that there are plenty of things wrong with it, too. Enrichment alone doesn't make NT kids PG, no matter how much we'd like for it to be so.
Could be it's a measurement problem, and therefore an artifact, and that the inclusion of the tails makes it crystal clear that the correlation doesn't hold up.