Hi Gang,
Thanks for the response. It was quite helpful. To make absolutely sure that we are doing things right, I have two follow-up questions. The first is a clarification.
1) For the GLT matrix suggestion by Kyle (first seven rows), extracting t statistics for each row (each contrast), thresholding the SPM and then taking the union would answer the question, "For this voxel, does Condition 1 differ from any other single condition?" I think this is what you are suggesting above, as a conjunction analysis.
Whereas testing the F statistic for that GLT would answer the question, "For this voxel, does any condition differ from any other condition?" As such, the F stat will for the first-7 rows of kyles GLT can be significant if even A does not significantly differ from anything
2) When calculating the F statistic for the GLT, do the regressors have to be orthogonal? If we wanted to ask, "For this voxel, does any condition differ from any other condition?" would it be best to use the following GLT matrix?:
1 1 1 1 0 0 0 0
1 1 -1 -1 0 0 0 0
1 -1 0 0 0 0 0 0
0 0 1 -1 0 0 0 0
0 0 0 0 1 1 -1 -1
0 0 0 0 1 -1 0 0
0 0 0 0 0 0 1 -1
instead of the first 7 lines of Kyle's GLT matrix?
Thanks for your input,
-Chris