[Freesurfer] medial wall crossing the surface

Dear All,

I too am seeing the medial pial surfaces crossing hemispheres in some subjects.

The approach that appears to be working for me involves making copies of brainmask.mgz and aseg.mgz For the left hemisphere crossing into the right, edit brainmask.mgz and aseg.mgz to delete voxels at the crossing belonging to the right hemisphere. I then run recon-all -autorecon-pial -hemi lh -subjid subject followed by returning the brainmask.mgz and aseg.mgz back to their original copy.

This is similarly repeated for the right hemisphere crossing into the left.

Are there any issues or concerns with this approach particularly in regard to thickness or the segmentations?

Thanks,

Paul