I’m planning to run an analysis using mri_glmfit in the command line to look at relationships between cortical gray matter volume and indices of adolescent substance use. Do I need to include brain/head size as a covariate in these analyses even though they are performed in the conformed space of fsaverage and not in native space? 

I suspect the answer is yes, but in looking through the archive many of the questions pertained specifically to subcortical analyses or those performed in QDEC so I wanted to clarify for cortical volumetric analyses. Thank you for your help!

Best,