Dear Freesurfer team,

I am running the glm analysis on CBF maps. I have two class and I want to see the group difference regressing out the cortex thickness. I know that I can add the average cortical thickness as one covariable in FSDG file. However, I think it might worth trying to do this with GLM by using a per vertex regressor (thickness). 

My fsgd file is like below:

Class Patient

Class CTL

INPUT subj1 Patient

INPUT subj2 CTL

....

My command is below: 

mri_glmfit --y lh.fsaverage.cbf.sm10.mgz --fsgd fsgd.txt --C contrast.mtx --glmdir lh.output.glmdir --surf fsaverage lh --pvr lh.thickness.sm10.mgz

My contrast matrix is -1 1 0. 

I am wondering if what I do is correct or not. Does it make sense or I should use the mean thickness as regressor? 

Thank you for your time!

Best regards,

Vicky