Hi everyone,
In some cases I get a really bad subcortical segmentation and I was wondering what the reason might be and how I could fix this. Attached is a screenshot showing the nu_noneck.mgz with the aseg.mgz overlaid.
Viewing this screenshot I would say, that during mri_remove_neck the possible brain is not correct, therefore more tissue is erased and the talairach_with_skull.lta is wrong. Though it is not written in the ReconAllDevTable I would guess, that the bad talairach_with_skull.lta is used during the following stages which affects the aseg.mgz.
1. What does the possible brain during mri_remove_neck refer to?
2. Could this simply be fixed by using a greater radius for mri_remove_neck or is most probably something else terribly wrong?
Thanks in advance,
Falk
