Hi Ingo,
You are talking about the pial surface and not the aseg.mgz segmentation values, aren't you? The surfaces don't usually get into the cerebellum. If you remove them from the brainmask.mgz but don't rerun the aseg it won't change the cerebellar volume values. There's no reason that it should work superiorly and not inferiorly . You can't extend the pial surface directly, but if the wm below the subcortical junction is darker than 110, you can put some control points in the body of the wm (make sure it's not in partial volume voxels) and that should bring the intensity of the whole region up.
Send an image of the where you think the pial surface is incorect and we'll take a look.
cheers, Bruce
On Mon, 20 Nov 2006, Wolf, Ingo wrote:
Dear all,
I have 2 question concerning edits to the pial and wm surface.
- After editing the pial surface in the ventral temporal lobe (parts of
the cerebellum were erroneously included), there were no visible changes in the pial surface after rerunning autorecon2. Is it possible that correction of the pial surface by editing/deleting voxels in the brainmask.mgz volume shows an effect in superior parts of the brain (e.g. upper parts of the dura) but not in inferior parts that are adjacent to other tissue? How do I correct errors like that? And how do I extend the pial surface anyways?
- When deleting those exessive pixels of the cerebellum that were
classified as pial surface, I obviously deleted parts of the cerebellum in the brainmask.mgz. Will this have an effect on the cerebellar volume measurements?
Thanks a lot! Best,
Ingo Wolf