External Email - Use Caution
Thank you Matt for letting me know that.
John
For what it’s worth, we had problems with FreeSurfer 7.X with surface placement
accuracy on HCP-Style data (i.e. 0.8-0.7mm isotropic T1w and T2w) and have
continued to recommend using FreeSurfer 6 for the HCP Pipelines for now. We
usually quote 2.6mm average cortical thickness for young adults, but it does
depend on age.
Matt.
From: <freesurfer-boun...(a)nmr.mgh.harvard.edu> on behalf of Xiaojiang Yang
<xjyan...(a)gmail.com>
Reply-To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu>
Date: Friday, March 12, 2021 at 10:29 AM
To: Freesurfer support list <freesurfer(a)nmr.mgh.harvard.edu>
Subject: [Freesurfer] Cortical thickness on version 6 vs 7
External Email - Use Caution
Dear Freesurfer developers,
I have a cohort of subjects that are already run under FS 6. I recently
upgraded Freesurfer from 6.0 to 7.1.1, and now I also have run these subjects
under FS 7.
When comparing the cortical thickness results obtained from FS6 and FS7, I
found that FS 7 results are much smaller than the FS6 ones. On average, about
13%-21% decrease in the subject's average cortical thickness. Specifically, on
a cohort of images from GE scanners, the average cortical thickness from FS6 is
2.5mm, but from FS7 is less than 2.0mm.
According to some published papers, I think 2.5mm as the average cortical
thickness is more accurate than 2.0mm.
Have you encountered this issue or question before? Do you think I should keep
using FS6? Please give me your advice. Thank you!
John.