Hi Nicole, It looks like recon-all -autorecon2 had trouble segmenting the gray/white matter. you can try using 3duniformize to make the gray,white matter more uniform, before doing the freesurfes recon-all. you can also try skullstrippong in afni with 3dskullstrip first, before freesurfer processing. I have attached script files for the uniformize and skullstrip. if this does not fix it, I can come to help you, or you can e-mail the problem to the freesurfer people at freesurfer@nmr.mgh.harvard.edu . They are very helpful and will response relatively quickly. Jon
----- Original Message ----- From: "Nicole T Nowak-Saenz" nowaksae@uwm.edu To: "Jon Alan Wieser" wieser@uwm.edu Sent: Thursday, March 7, 2013 12:35:57 PM Subject: Freesurfer
Hi Jon, I think you mentioned working with Freesurfer. Have you ever seen anything like this? I'm hoping the problems with segmentation, and white/pial lines are due to the recon-all somehow leaving out half the brain (e.g., temporal lobes). To me it looks like half the brain is missing, yet some skull remains. I have a feeling this will require more than adjusting a watershed parameter. Any clue how these errors occured or what to do about them?
I should mention I had to use a flag which was provided in the failed recon-all log to correct an error like this one: DICOM WARNING: file 00020002.dcm has Rescale tags.
If this is something you know something about and have a few minutes to address, I would appreciate it. If not, I completely understand.
Nicole