Dear Jeff,
The order of the columns was changed with one of the SPM updates, but the principle for setting up contrasts is still the same,
for the main effect time go with
[zeros(1,2) 1 -1 0 .5 -.5 0 .5 -.5 0 zeros(1,44)
zeros(1,2) 0 1 -1 0 .5 -.5 0 .5 -.5 zeros(1,44)]
and for the interaction group x time
[zeros(1,2) zeros(1,3) 1 -1 0 -1 1 0 zeros(1,44)
zeros(1,2) zeros(1,3) 0 1 -1 0 -1 1 zeros(1,44)]
(please double-check yourself). For main effect group you should create an average s8wp1 file for your different subjects = (image from time point 1 + time point 2 + time point 3)/3 and forward this into a two-sample t-test to test for differences between group 1 and 2 to have the correct error term for that contrast. Donald McLaren has written quite a lot on that issue on the list.
To make sure you're looking at the same selection of voxels in the two-sample t-test I'd go with the mask.nii obtained from the Flexible factorial and specify this as an explicit mask for the two-sample t-test instead of the threshold masking \ absolute threshold.
Best
Helmut
|