Thanks, this works and it recognizes it as a legacy analysis. Some subjects work
fine this way. But for some other subjects, it exits with
an error saying aseg.mgz not found.

mris_make_surfaces: could not read segmentation volume /usr/home/rhdesai/From_old_apu_home3/freesurfer.subjects/335/mri/aseg.mgz

The command was

recon-all -subjid ${s} -legacy -autorecon2-pial -autorecon3

(I ran it first with only autorecon3, but it appears that there the inflated surface has different number of nodes than wm/smoothwm surfaces for these subjects for
some reason, so it exits with errors. To make the final surfaces again, autorecon2-pial
was used.)

thanks,
Rutvik



On Mon, Oct 6, 2008 at 7:07 PM, Bruce Fischl <fischl@nmr.mgh.harvard.edu> wrote:
hmm, Doug or Nick would be better at this, but try:

touch $SUBJECTS_DIR/$subjects/scripts/recon-all.status


then rerun the subject. Hopefully it will then recognize that it is a legacy analysis.

cheers,
Bruce



On Mon, 6 Oct 2008, R D wrote:

Hello,

I have some surfaces created by freesufer in 2006. Spherical registration
was not done at that time, which needs to be done. I ran recon-all with
-legacy, but it crashed due to disk space problems. Now it doesn't run
either with or without -legacy option. When -legacy is specified, it says
"this is not a legacy analysis. It once was, but it has been fixed." Without
-legacy, it asks for some files such as filled.mgz and aseg.mgz that are not
present. Is there a way to 'reset' the analysis (e.g., change a flag
somewhere) so it can start over with -legacy?
thanks,
Rutvik