Hi there, I'm working on a project in which I manually "erased" voxels (changed to '1') in the optic chiasm region in the wm.mgz and brainmask.mgz. This was done to eliminate those voxels from the surface reconstructions. We've since realized that this isn't necessary...our results are largely unaffected by the edit. So we've skipped that edit on most (110+) scans. For the few scans that have the optic chiasm edits, I'd like to regenerate the original surfaces but would like to retain the numerous other edits I've made to the brainmask.mgz and wm.mgz volumes outside the optic chiasm region.
I have tried to do this by "cloning" the voxels from the original brainmask and wm volumes to the edited version, thereby restoring the original voxel values in only this region. When I then rerun recon all from the autorecon2-cp option, the "restored" voxels in the wm volume are set set to "0" (so no longer the original values and not even the "erased" value of "1") before the surfaces are generated. The result is that the surfaces in the optic chiasm area remain "clipped." I've tried this on a couple different scans and after moving the wm.seg.mgz volume out of the mri/ directory.
Any idea why this is happening? Is there a solution that does not require me to edit these scans from scratch?
Thanks, Jason
freesurfer@nmr.mgh.harvard.edu