I am using MVPA to identify how resting-state fMRI spatial pattern is differently associated with longitudinal change in behavioral measure in group 1 and 2 (controlling for baseline measures, age, motion).
My initial model was this:
[group_1, group_2, age, motion, measure_baseline, measure_change_group_1, measure_change_group_2] = [0 0 0 0 0 -1 1]
However, breaking down measure_baseline into group_1 and group_2 like below yields different result in the MVPA:
[group_1, group_2, age, motion, measure_baseline_group_1, measure_baseline_group_2, measure_change_group_1, measure_change_group_2] = [0 0 0 0 0 0 -1 1].
I wonder whether I should also break down other covariates such as age, motion into each group like:
[group_1, group_2, age_1, age_2, motion_1, motion_2, measure_baseline_group_1, measure_baseline_group_2, measure_change_group_1, measure_change_group_2] = [0 0 0 0 0 0 0 0 -1 1],
or just have measure_change broken down and keep all other variables combined like:
[All_subjects, age, motion, measure_baseline, measure_change_group_1, measure_change_group_2]
= [0 0 0 0 1 -1]
Any suggestions would be greatly appreciated. Thank you