A recurring issue with my recons is that the pial surface excludes the hippocampal formation and associated MTL grey matter structures.  A typical image is attached.

Aseg has clearly captured the location of these structures correctly, as indicated by the coloring in freeview, 

so I'm wondering: does freesurfer not use the aseg results when constructing the pial surface?  If not, can it be made to do so? Is there any workaround other than manual correction of each pial surface?

Interestingly, I have never seen freesurfer exclude the ventral midbrain (or thalamus), though it seems to be in an analogous situation.

Thanks,

Aaron