Howdy -
I'd like to do a t-test on a contrast that I didn't completely specify in the 3dDeconvolves. Specifically, I have GLTs for conditions HSS, LSS, and NOV. I'd like to test (.5*HSS + .5*LSS) vs. NOV. I had the foresight to do some linear combination contrasts of this type in 3dDeconvolve, but sadly my prescience was not total.
I'd like to test this contrast without having to re-process all the data, so I tried to do the magic in 3dttest++, like so (cruft removed and abbreviated for clarity):
3dttest++ -prefix "${STATS_DIR}/$outFile"
-paired
-AMinusB
-setA setA
01 ".5*$data_dir/stats.tp101.amblock.averaged+tlrc.HEAD[${contrast1a}#0_Coef] + .5*$data_dir/stats.tp101.amblock.averaged+tlrc.HEAD[${contrast1b}#0_Coef]"
...
-setB setB
01 "$data_dir/stats.tp101.amblock.averaged+tlrc.HEAD[${contrast2}#0_Coef]"
but no go. Is there an easy way to do this, or do I have to re-run all the 3dDeconvolves again to create GLTs for .5*HSS + .5*LSS, and all similar things?