Nick, List:
Still no luck with this data. See attached as an example.
To get to this point I had run:
recon-all -all
This ran to the error: mri_watershed Error: GLOBAL region of the brain
empty !
Then I followed your recommendation below
recon-all -s <subjid> -skullstrip -no-wsgcaatlas, which had no errors and
then
recon-all -s <subjid> -autorecon2 -autorecon3
This produced the result attached.
I then deleted this directory and re-imported all the dicom data for this
subject. Looked at the images, which look fine.
Then I ran recon-all -s <subjid> -autorecon1
This fails with the mri_watershed error.
Then I ran recon-all -s <subjid> -skullstrip -no-wsgcaatlas.
This runs without error but when I look at the brainmask.mgz it looks just
like the attached image.
Is there some way to set the thresholds differently? The T1 image looks
fine to me although perhaps the intensity is slightly less than usual, but
it doesn't seem particularly noisy.
Darren