Dear Christian,
thank you very much for your quick answer. To clarify, our experimental condition is the within-subject factor (=time). The intervention was brief, so all participants were tested once before treatment and once after treatment, and the order was counterbalanced (some received treatment first, others at the second time point). We modulated by the factor treatment (within-subject) and order (between-subject). We do not have a control group (i.e., one group was measured twice without treatment). Does this change anything?
You mention difference images as a possible solution. Are there any advantages or disadvantages of longitudinal models over contrast images? I thought that longitudinal models are more sensitive to changes and the intercept of the subjects is taken into account.
Thanks a lot
Susanne
|