Hi Rahul

Sorry, meant to respond to your previous post. The defects can change with different versions and don’t necessarily reflect things being “better” or “worse”.  Sometimes even a single voxel difference in the wm segmentation can create a giant defect. Can you investigate a bit and see what differences are causing the defect? If the defects are too large we won’t be able to properly correct them. Even if recon-all finishes it may not be accurate if e.g. large chunks of skull were left around be the brain extraction

Cheers

Bruce

From: freesurfer-bounces@nmr.mgh.harvard.edu <freesurfer-bounces@nmr.mgh.harvard.edu> On Behalf Of Paul, Rahul
Sent: Wednesday, September 22, 2021 10:35 AM
To: freesurfer@nmr.mgh.harvard.edu
Subject: [Freesurfer] Freesurfer version

Hi freesurfer experts,

I was using FreeSurfer v6 for my research. Recently, I started move to FreeSurfer v7.2.

I encountered the following issue:

Some patients, that had extra large error (xL error), were running fine (recon-all) in freesurfer 6 (it took time, but the process finished, and the segmentation was fine too) but in freesurfer 7, same patient recon-all encountered the extra large error (e.g., CORRECTING DEFECT 35 (vertices=9718, convex hull=1801, v0=40087 Command terminated by signal 11) and after couple of hours recon-all exited (recon-all -s 8375 exited with ERRORS ). I am not sure why recon-all is existing for freesurfer 7.2 and how can I stop that. 

Thank you

Rahul