I am having a bit of a problem the aseg on this subject. (See 2013.03.22.cor108.aseg.jpg attached) It is way overestimating the white matter bilaterally. The problems are about this bad for about 20 or more slices so editing manually seems like... a less than efficient proposition. Any thoughts on how to fix this programmatically?
For reference, I have a attached some other images with no segmentation and using the Desikan-Killiany atlas (aparc+aseg) and the Destrieux atlas (aparc.a2009s+aseg) which look fine.
Thanks, -Ronny