Hello - I have a question regarding the correct 2nd level analysis to run in SPM. Data was collected from 12 subjects. Each subject underwent four sessions of task-based fMRI. During each of the four sessions, each subject was under the influence of a different treatment (Treatments A, B, C, and D). During each treatment, the effectiveness of the therapy was evaluated (Score), and the presence or absence of a side effect was evaluated (Side Effect).

The questions i want to answer are:

Are there regions of activation due to task (for treatments A-D) that correlate with the effectiveness of the therapy (Score)?
Are there differences in the functional activation maps during side-effect inducing versus non-side-effect inducing treatment?
Combining 1. and 2.: is there an activation pattern that correlates with both good treatment response and the absence of side effect?

My planned approach:

for 1: run a within-subject ANOVA for the four conditions, with the Score for each condition as the covariate
for 2: again, within-subject ANOVA for the four conditions, with a binary covariate (1 for side effect, 0 for no side effect)
for 3: not really sure

Am I on the right track? Thanks very much

-Will