Hi there,

I wonder if the fsgd file and the command lines listed below are corrected.

My sample included 4 groups with different diagnosis (g1, g2, g3, g4).

I'm interested to map the vertices that show a relationship between the FC (obtained by FS-FAST) and cortical thickeness, regressing out the diagnosis.

To this aim I used the --pvr option:

mri_glmfit --y ces.nii.gz --wls cesvar.nii.gz --surface fsaverage rh --glmdir

Thickness.wls --nii.gz --fsgd g4v1_thickness.fsgd --C intercept.mtx --C

slope.mtx --pvr $SUBJECTS_DIR/rh.thickness.10B.mgh

The fsgd file lists:

-------------------------------------------------------------

GroupDescriptorFile 1

Title Relationship FC-thick reg out dignosis

Class g1

Class g2

Class g3

Class g4

Input XX1 g1

Input XX2 g1

...

Input YY1 g2

Input XX2 g2

...

Input XY1 g3

Input XY2 g3

...

Input YX1 g4

Input YX2 g4

...

--------------------------------------------------------------

intercept.mtx 0.25 0.25 0.25 0.25 0

slope.mtx 0 0 0 0 1

thanks piero