Hi Freesurfer experts, 

I am trying to convert the aseg back into native space of the anatomical T1 image. I followed the instructions as in the webpage:

specifically this line of code:

cd $SUBJECTS_DIR/<subjid>/mri
mri_label2vol --seg aseg.mgz --temp rawavg.mgz --o aseg-in-rawavg.mgz --regheader aseg.mgz

However, the result is very odd. The aseg boundaries appear correct in the subcortical regions. however, at the gray mater boundaries, there are single voxels incorrectly labeled as different subcortical-regions. As a result, in the viewer, the gray matter boundaries look multicolored. 

This is problematic, as I would like to extract a mask of specific ROIs (e.g. right amygdala) but there are many other voxels labeled '54' in addition to the right amygdala. 

See attached images. I would greatly appreciate your assistance!! 

Thank you, 

Michelle