I would like to include covariates in the freesurfer repeated anova steps https://surfer.nmr.mgh.harvard.edu/fswiki/RepeatedMeasuresAnova. When i tried including 3 covariates, mri_glmfit complained about "martix is ill-conditioned". 

How do i fix it? 

Below is an example my fsgd file. I've also attached the matrix design in the email.

 
GroupDescriptorFile 1
Class PsyDis1
Class PsyDis2
Variables TP1vsTP2 Age Llesion Rlesion
Input Psy1_Pre.long.Psy1 PsyDis1 1 68 96 61 
Input Psy1_6mos.long.Psy1 PsyDis1 -1 68 96 61
Input Psy2_Pre.long.Psy2 PsyDis2 1 39 217 140
Input Psy2_6mos.long.Psy2 PsyDis2 -1 39 217 140

contrast: 0 0 1 0 0

Best, 

Paul