Jerry,
There was a problem in the recon-all script which you can fix. Edit recon-all, in the mri_fill section (look for --- Fill ---), on the line where it says 'if($UseAseg)', change 'aseg.mgz' to 'aseg.auto.mgz'.
The problem is that aseg.mgz contains the CC segmentations, and mri_fill is not aware of those segmentation labels (yet), so it doesnt find the CC. The aseg.auto.mgz is not CC-segmented, and mri_fill is able to find the CC. This fix will appears in the next release (v4.0.3).
Nick
On Mon, 2008-02-04 at 12:18 -0500, Jerry Yeou-Wei Chen wrote:
Hello,
We ran into the following error during autorecon2 with one subject:
mri_fill: could not find any points where lh and rh wm are nbrs
It has been suggested to another user that the "-noaseg" flag be used. (http://www.mail-archive.com/freesurfer@nmr.mgh.harvard.edu/msg06560.html) However, we would like to ask how this might affect the subsequent steps and how to take this into account in subsequent group analysis. Also, we would like to clarify if this would prevent the subcortical volumetric calculations.
Thanks!
- Jerry
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer