Hi everyone,

From FreeSurfer manual and discussion forum, I found that volume and LGI can be correlated using mri_glmfit command with -pvr flag.

For that, my fsgd file has age, gender as covariates and the file is as following:

GroupDescriptorFile 1

Class Male

Class Female

Variables Age

Input        zerek1 Male 21

Input zerek2 Male 25

etc.

My concern is:

To run mri_glmfit command, I am not sure how to define my contrast matrix if I am interested in removing the effect of age and sex.

1st option I tried is: 0 0.5 0.5 0 1

Here, I think first 0 regresses out age, 0.5 0.5 should be averaging over males and females and last 0 1 correlates between volume and LGI. But this looks like it is not giving me correct correlations maps, although there is no error.

2nd option I tried is: 0 0 0 0 1

This looks like its working and maps look correct to me but I want to make sure this is the correct contrast matrix and I am not sure how this matrix is regressing out age and gender or 1st option is correct contrast matrix?

Thank you so much !

Best,

Sahil