Just set the CDR-related slope components in the contrast matrix to 0. If you truly believe that the CDR has no effect, then you can create a new fsgd without CDR as a variable. This will improve both DOF and efficiency.
doug
Martin Kavec wrote:
Hi,
I would like to study cortical thickness in a cohort of patient with respect to MMSE and CDR. I prepared a fsgd file where I have both MMSE and CDR listed:
GroupDescriptorFile 1 Title Mytitle Class Male Class Female Variable Age Variable MMSE Variable CDR Class Ctrl Class MCI Input subject1 Male 70 27 0.5 MCI Input subject2 Female 65 28 0 Ctrl
My design matrix should thus have 4*(3+1)=16 columns/regressors.
I am puzzeled a bit, how would I set the contrast matrix to study only correlation of thickness with MMSE, while ignoring CDR. Or should I have a separate fsgd file?
Thanks,
Martin _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer