how did you manually create the brain.mgz files? These must be intensity normalized (with mri_normalize) first, otherwise nothing will work. The white matter should be close to or equal to 110 everywhere in the brain.mgz volume. On Thu, 13 Apr 2006, Jenni Pacheco wrote:
Hi Tom,
What kind of scans are you working with? How were they collected? Have you averaged more than one acquisition?
Jenni
From: Tom Bayer [mailto:tabayer@gmail.com] Sent: Thursday, April 13, 2006 7:41 PM To: Jenni Pacheco Subject: Re: [Freesurfer] aseg.auto.mgz
autorecon2 ran all the way through. I first noticed a problem because the wm.mgz file is only a narrow band of voxels and its distorted. The brain.mgz file is missing some areas of white matter. aseg.auto.mgz is missing a hemisphere, and aseg.mgz is very similar in shape to wm.mgz, which is an irregularly shaped blob that occupies about 1/4 of the coronal section, and is roughly centered in the image.
I am fairly certain that the failure occured in the step generating aseg.auto.mgz or one of the steps just prior, because nu_noneck.mgz looks good.
I created brainmask.mgz and brainmask.auto.mgz using manual thresholding and they are exactly to my specifications. Strangely, the output file brain.mgz still has problems. It looks exactly as it did before manual thresholding.
on a previous post to the list, it was suggested that default thresholding took over because brainmask.mgz and brainmask.auto.mgz do not match. I assure you that they do match. I cop[ied the mask that I manually created so that both files match exactly.
-tom
On 4/13/06, Jenni Pacheco jpacheco@nmr.mgh.harvard.edu wrote:
Hi Tom,
We're going to need a little more info. Did autorecon2 finish? Did it fail part way through? When you say everything has problems, what kind of problems?
go here to see the types of information we need in order to best help you: https://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
thanks,
Jenni Pacheco, M.A. Massachusetts General Hospital Athinoula A. Martinos Center for Biomedical Imaging Bldg. 149, 13th Street, CNY 2301 Charlestown, MA 02129 jpacheco@nmr.mgh.harvard.edu
On Thu, 13 Apr 2006, Tom Bayer wrote:
I'm having problems with the output of recon-all -autorecon2. proceeding down the recon-all dev table, I checked the outputs of each step. everything looks good until aseg.auto.mgz everything after has problems. What kind of manual intervention do I need to do for this?
-Tom