We ran a functional language localizer analysis in FSL and we want to apply the mask from this analysis to our SMRI analyses using the mri_glmfit command with the --mask option.  Basically, we want to examine whether GMV within these language localizer regions is correlated with performance on various tasks. I previously used mri_convert to convert the .nii file to .mgh, and tried to run the mri_glmfit --mask command but received an error (ERROR: dimension mismatch 1 between y and mask) which I believe is linked to the fact that the previously run mris_preproc files are 4D, while our functional mask is still only 3D.  When I use tksurfer and overlay the 3D mask file, it looks fine.  

So the question is how can we convert our 3D functional mask into a format that we can use for the mri_glmfit command (Nverteces x 1 x 1 x Nsubjects)?  
--