I wanted to try out an informed basis set for my model since it says that "it allows for variations in subject-to-subject and voxel-to-voxel responses" (although that sounds like something one would always want to do) . I'm not sure how it allows for the voxel variation since neither the variation in response start nor in duration would allow that.
But, anyway, how do you create a contrast vector for a single subject (Ch. 30 in the manual shows how to do second level analysis once you already have all the contrast images from the first level analysis). I have 4 conditions, and no matter what I enter as a contrast vector, I get the exact same results for activation.
Without the temporal derivatives, one of my contrasts for a single was (for a t-test): 1 1 1 -3 0 0 0 0 0 0 (the first 4 for the conditions in my setup). With the temporal derivatives I tried an F-test with 2 columns added after each condition, e.g. 1 0 0 1 0 0... and tried other values as well, but it shows the same activation no matter what I put in there.
Thanks in advance for any hints.
|