You might have introduced a new defect, or just made an existing one worse. If there are topo defects, you can't just ignore them. You don't necessarily have to remove all the dura. Typically there is a "peninsula" where the main WM attaches to the the "island" of dura. All you have to do is cut that peninsula.
On 10/15/2021 9:21 AM, M Pope wrote:
External Email - Use Caution External Email - Use CautionHi all,
I ran recon-all successsfully on a subject. I then did some manual edits on both the brainmask (removal of dura) and wm and then ran recon again for the edits. This usually works fine, however for one subject the recon (after edits) fails because of time limit on my process. It gets stuck on the fix toplogical errors stage, as there is a defect with 36000+ vertices included. I looked at the lh.inflated.nofix and lh.orig.nofix and indeed you can see there is a lot of dura/skull being included in the surface on the lh. However, recon did work just fine the first time where these did happen, and the final surfaces look fine (the manual edits I did were few!) Gcut also does not work because it removed too much of the temporal areas and the cerebellum.
Are there any suggestions on how to fix this without having to go in and manually remove all the dura (it is a lot!). I attach the recon.all log so you can see how the first recon ran (without errors) and then after edits where it fails. Could I perhaps use the -nofix flag?
Thanks, Marianna
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://secure-web.cisco.com/1PQXVpHHHzmaj81WDM7Hq3HYRz6L5gMf9eXQ3fjM9XnO1F3...