Hi Dustin,
While those parameters go into the baseline, it does
seem possible to do this. The main difficulty is that
the labels are not actually known to be referenced as
symbolic GLTs using -gltsym.
However you can construct the GLTs the old fasioned
way, using -glt. To get the proper regressors order,
either add -bout to see labels in the output bucket,
or look at the X.xmat.1D header to figure it out.
Given that, try to create each full GLT matrix that
you want. It is tedious, but it should at least work.
- rick