[Freesurfer] statistical question

Hi there, This is a bit outside the pure technical questions. My subject pool has both boys and girls, with the boys having larger intracranial volume. My results when controlling for gender and intracranial volume however are not the same (obviously, one is a dichotomous variable, the other continuous). Looking at the results, i prefer the latter over the former, more consistent with the literature, and more internally consistent (i.e., brain regions associated with intenalizing sub-scales and the internalizing total scale). But it could be argued that I am removing important variance - maybe stressed out kids have smaller heads?

Any recommendation?

thanks