Thanks Doug!

I am manually segmenting the fusiform gyrus into anterior and posterior portions and save the segmentation. Then use mri_segstats to obtain the volumes. I have two problems.

(1) segmented volumes (sum of anterior + posterior fusiform), total volume from mri_segstats (#1007 ctx-lh-fusiform or #2007 ctx-rh-fusiform), total volume from ".stats" files (rh.aparc.stats, lh.aparc.stats) do not match one another.

(2) in tkmedit, "save segmentation as" does not save my segmentation job. In a directory that I created, I see COR image files but I cannot view them in tkmedit and I cannot extract any volume using mri_segstats (the results are all 0s). However, two things worked for me: (a) if I just use "save segmentation" in the exisiting aparc+aseg.mgz, I can load it and see my segmentation job and get a volume using mri_segstats from aparc+aseg.mgz; (b) although "save segmentation as" did not work for the fusiform, but it worked for the hippocampus, which I tried as a test.

Thanks for your advice in advance!

Regards,
Yunjo


Douglas N Greve wrote:
Use the surface-based stats file (?h.aparc.stats) for cortical structures, use aseg.stats for subcortical gray matter, and wmparc.stats for the white matter.

dogu

Yunjo Lee wrote:
Hi Doug again,

I have a quick question related to my previous one.
Then which one is more accurate, for example, mri_segstats from aparc+aseg.mgz or aseg.stats, rh.aparc.stats? I got a bit confused by which one you meant by "stats files".

Thanks,
Yunjo


Doug Greve wrote:
 
The stats files are computed with partial volume correction.

doug

On Thu, 15 Jul 2010, Yunjo Lee wrote:

   
Hi,

I am a new user of FreeSurfer and trying to figure out gray matter segmentation. A volume that I extracted from aparc+aseg.mgz using mri_segstats did not match the volume reported in /stats/ text files (for example, aseg.stats, rh.aparc.stats). Even without specifying a mask in mri_segstats, there's always a discrepancy of 100 voxels or something. Where do the numbers in /stats/ files come from? Aren't they from aparc+aseg.mgz?

Thanks!

Yunjo



     


_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer