Michael,
I think that the index 1 in "CalmGo[1]" is associated with the the second basis function (temporal derivative) for SPMG3.
> is the variance around each of the lower level betas taken into account for estimating
> the beta and variance of the contrast?
A simple and straightforward answer is YES.
> what is the best way to take into account as much variance in the estimates as possible to eventually run a paired
> t-test with 3dMEMA?
Obtain the contrast at the individual level, and then run 3dMEMA with one-sample (not a paired) test.
Gang