>I have four groups of subjects (patient male, patient female,
>control male,
>control female) and one contrast image of each subject
>represnting the main
>effect. Is it possible to calculate the interaction of patient
> and gender
>in a random effects ANOVA. What would be the appropriate way
>to set the
>contrast?
I'm not sure which version of SPM you are using, but basically you want to
enter the contrast images for each subject into a second-level
between-subject ANOVA with 4 conditions (plus a constant if you want). Then,
if you enter them in the order: PM, PF, CM, CF, the contrast for the patient
X gender (X effect) interaction is simply [1 -1 -1 1].
Rik
|