Re: [Freesurfer] correlation in mri glmfit controlling for 2 categorical variables and 1 continuous variable