If you are going to use the FSGD framework, you either have to have all covariates be DOSS or all be DODS, you can't have a mixture. If you want a mixture, then you need to create your own design matrix and then contrast matrix from that. Pass the design matrix to mri_glmfit with --X (and don't include the fsgd)
On 11/5/19 5:28 PM, Georgina Angelopoulou wrote:
External Email - Use Caution
I am sorry, please let me rephrase. I want to compare patients vs controls in Surface area and grey matter volume. I have already checked age slope and it is not significant for either GMV or SA. However eTIV slope is significant in both cases (we do check for eTiv slope first as we did for age, right?) So in that case that we have two nuisance variables (age and eTiv), eTIV is significant but age isn't, how are we going to formulate the contrast? Would 1 -1 0 0 0 0 be correct?
Thanks once again, Georgina
Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer