Hi Gabe
this wasn't really a bug per-se, just induced some behavior that people didn't like. You can't skip the mri_ca_normalize step or the aseg won't be accurate. The aseg patch I believe just starts autorecon2-cp after the aseg hs been created, but Nick or Zeke can correct me if I'm wrong.
cheers Bruce
On Mon, 6 Oct 2014, Marx, Gabe wrote:
Hello Freesurfer experts,
I had a question regarding the v5.1 control point mri_ca_normalize bug. I read the release notes and know that this bug can be worked around by adding the –nocanorm flag to my recon-all however I have become worried about the ramifications of skipping mri_ca_normalize in my pipeline. Would someone be able to give me a better description as to what mri_ca_normalize is doing and what I am sacrificing by taking it out of my pipeline? Furthermore, in regards to the patch for the recon-all script to fix this bug, what is the patch doing exactly? If I had some data in which I used the –nocanorm flag and other data in which I used the patch would I still be able to make valid analysis if I merged them? Would there be significant inconsistencies?
Thanks!
Best,
Gabe