Hi all,
Sorry about this long message. I'm trying to run the higher level analyses, but I'm afraid I don't have the model set up correctly. I have 3 subjects scanned twice, one with a 2D movie intervention and one with a 3D movie intervention. On each movie scan day, the subject was required to do 2 mental rotation tasks before the movie and 2 after the movie. I combined the 2 task runs before the movie calling them "pre" and 2 task runs after the movie calling them "post" for each subject and each version movie intervention. Hence, I have 3D pre, 3D post, 2D pre, 2D post, for each subject. Below is how I set them up in the glm using fixed effects; inputs are first level FEAT directories from tasks 1-4 (inputs 1+2=tasks 1+2; inputs 3+4=tasks 3+4).
EV Group EV1 EV2 Contrasts Title EV1 EV2
pre post
Input 1 1 1 0 C1 pre mean 1 0
Input 2 1 1 0 C2 post mean 0 1
Input 3 1 0 1
Input 4 1 0 1
Then I fed all the subjects' pre and post outputs into another higher level analysis using mixed effects FLAME with the cope images, combining all three subjects' pre and post. For example, for 3D version (Inputs 1, 3, 5 are the Pre, and Inputs 2, 4, 6 are the Post); similarly done for the 2D version:
EV Group EV1 EV2 Contrasts Title EV1 EV2
Pre 3D Post 3D
Input 1 1 1 0 C1 pre3D 1 0
Input 2 1 0 1 C2 post3D 0 1
Input 3 1 1 0 C3 post-pre3D -1 1
Input 4 1 0 1 C4 pre-post3D 1 -1
Input 5 1 1 0
Input 6 1 0 1
If I have done the above correctly and want to do another higher level analysis to see the contrast "post-pre3D" - "post-pre2D", how would I set that up? I tried just taking the cope images (post-pre) from the previous analysis and set up as shown below:
EV Group EV1 EV2 Contrasts Title EV1 EV2
post-pre3D post-pre2D C1 3D-2D (post-pre) 1 -1
Input 1 1 1 0
Input 2 1 0 1
When I set them up like this, the Post-stats section did not even run. Help! What would be the proper way to compare post-pre(3D)-post-pre(2D)?
Thanks,
Amy
|