Follow-up: I then run into indexing problems when calling freeview with subfield volumes (similar to http://bit.ly/1pdEWRZ), which leads me to believe kvlQuantify is the only script that relies on the old convention. Changing the names back to CA2_3.mgz, etc., fixes THAT problem.

Rather than changing the filenames to the old convention, it's probably better to fix the script and leave the filenames alone.

On Wed, May 21, 2014 at 1:06 PM, C.P. Frost <carfro@gmail.com> wrote:

Hello all,

My site has recently upgraded to the newest version of Freesurfer, and I've run a set of subjects on recon-all using the hippocampal subfield segmentation flag. Everything went well. Then in running kvlQuantifyHippocampalSubfieldSegmentations.sh, I received error messages indicating that certain expected files didn't exist.

In particular, among output files labeled with the scheme $subj/mri/posterior_$side_$struct.mgz, the script expected ...CA2-3.mgz and ...CA4-DG.mgz, and existing files were named ...CA2_3.mgz and ...CA4_DG.mgz -- in other words, the script expected a hyphen but instead the filenames contained an underscroe.

After renaming the files appropriately, kvlQuantify... successfully generated the nonPartialVolumeStats files and everything worked fine.

I didn't see any threads with this solution, and only one or two that mentioned the problem (http://bit.ly/1vHbV5K), so I thought I'd offer it up in case it could be addressed in a future release.

Cheers,
C.P. Frost