External Email - Use Caution
Dear Faeze,
I would say it depends on what you want to do with the data. If you are interested in measures for which the pial surface is important and you see obvious errors in it, I would recommend editing.
If you only care about something unrelated like amygdala volume, ignore it but document this so people who use the same data for a different analysis later are aware of it.
I guess you already found this: https://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/TroubleshootingData
It takes a lot of time to edit (~15 minutes per image, plus the re-processing time and the 2nd check whether it looks good afterwards), but afaik there is no other way besides discarding the image. Some FreeSurfer Wiki pages describes alternatives for certain kinds of errors (e.g., changing watershed parameters) but we did not have much success with that in the past and after running it, you have to check all images again and maybe you will then *still* need to edit them).
We are seeing these kind of issues (dura left-over and assigned to pial GM) in about 1/3 of all our images, and so we usually edit roughly one third of them. I like the advice on the linked wiki page though (in Section 5.1): "Don't edit too much! This reduces reliability and is almost never needed."
In many papers it is not clear from the methods section how this was handled, and it would be great to hear what others on this list think and how they handle it, imo.
Tim
On June 4, 2019 at 12:07 AM Faezeh Vedaei faezevedaei@gmail.com wrote:
External Email - Use CautionHello
I just needed to get help from someone has experience working on correcting pial surface segmentation by recon-all. Based on FreeSurfer tutorial and some instructions I have found by googling I have found some clues that was helpful, but for the issue I have right now was not enough. As you know I need to make sure that there is no dura or blood flow included in pial surface (cortex), so need to do editing by removing extra voxels. For some of our subject's data, there are some spots or areas obviously has no barrier between dura and cortex, and in my eyes it seems dura in included in the cortex (or cortex is extended to the dura) which needs correction. But I was wondering if I want to remove them, that would be so time-consuming and hard to check and edit every spot and slice, as these are shown a lot on many spots. So do you recommend me to do editing on these areas or forget about? I really appreciate that if you clarify me about this process specifically dura removing correction.
Bests Faeze _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
-- Dr. Tim Schäfer Postdoc Computational Neuroimaging Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy University Hospital Frankfurt, Goethe University Frankfurt am Main, Germany