 |
 |
Dear SPMers, I'd like to identify regions commonly activated by two contrasts, and compare these regions across 2 groups of subjects. Based on Rebecca Elliott's recent e-mail regarding conjunction analyses, my thought is that I should do a multiple regression analysis with group and condition as regressors, where groups A and B get assigned values 1 and -1 and conditions a and b get values 1 and -1. I assume that a conjunction of the following two contrasts (see below) would give me regions which are activated by both conditions (but not more by one than the other), without taking into consideration the group. group condition mu 0 1 1 0 -1 1 On the other hand, I assume, a conjunction of these two contrasts (see below) would give me regions which are activated by both groups, without taking into consideration the condition. 1 0 1 -1 0 1 Neither of these contrasts is exactly what I'm looking for, because they don't explicitly model the conditions and groups at the same time. What I think I'd like is a way to look at the conjunction of 2 conjunctions: 1 1 1 1 -1 1 -- regions activated by conditions a and b in group A and -1 1 1 -1 -1 1 -- regions activated by conditions a and b in group B. Any thoughts? Thanks very much, Silvia Bunge Neurosciences Program Stanford University