Dear FSL community,
I have a question regarding a within-subject analysis in FEAT, and I couldn't find any answers on the wiki.
I conducted a pilot experiment with three conditions (A, B, and C). It's a simple block design. For conditions A and B, there are 4 runs each. My main research question concerns the difference between A and B (and there's no problem with that). However, I am also interested in the contrast A-C and B-C, but there are only 2 runs for condition C.
After having run the first-level FEAT analyses for all ten runs, I simply set up the contrasts A-C and B-C in the same way as A-B (using the first-level cope images as input), like this:
A B C
C1 A-B 1 -1 0
C2 A-C 1 0 -1
C3 B-C 0 1 -1
My question is whether it's legitimate to set up such higher-level contrasts (A-C and B-C) even though the number of runs contributing to the conditions is not equal. Does FEAT account for this? Or do I have to take any additional steps?
Any help would be much appreciated!
(As a side note, due to the nature of the experiment I cannot have all conditions alternated within runs; I know that would be a more elegant design but it's unfortunately not possible.)
Best regards,
Ingo
|