 |
 |
Dear SPM users,
I have a question re longitudinal fMRI modelling in SPM.
I'm constructing a 3 x 2 x 2 Repeated measures model for the within factors task (3 levels) and time-point (2 levels: t1 and t2), and the between factor group(2 levels: controls and patients). I am interested in the interaction of factors time and group, which I specify in the flexible factorial design batch. In this model, I'm also interested to include 2 within-subject covariates; duration of illness and medication exposure for the patients group and here is where i have trouble at.
I am planning to create one column for the duration of illness covariate (corresponding to both time-points) and one for medication (for both time-points). Covariates for the group controls will be modeled as zeroes. I am unsure whether I should center those covariates either at the overall mean, at a different value, or no centering at all?
I'm also unsure whether it is valid to include these covariates in a repeated measures ANCOVA model in SPM, in the first place?
Anyone experienced in this area? Any kind of advice is appreciated.
Many thanks
Vasilis