I am really struggling to model the following analysis scenario and any help would be greatly appreciated.
18 participants each scanned in 3 separate sessions: session 1 baseline; session 2 placebo; session 3 after treatment.
In each session (single block-design fMRI scan) a participant is required to perform two different tasks. I wish to model the effects of the treatment (none in session 1, 'sham' in session2 and real in session 3) on any of the tasks (EV1 and EV2) or the interaction between them (EV3).
I have performed lower-level analyses for each participant and for each session, using 3 EVs as described above (EV1 = task1, EV2 = task 2 and EV3 = the interaction of tasks 1 and 2).
I know I effectively have repeated-measures within a participant across sessions for each of my within session conditions (task1 / task2 / interaction).
Can I simply run independent higher level repeated-measures analyses for each condition (i.e. using the COPEs describing Task1 or Task 2 or their interaction) to see whether there is an effect of treatment on any of them? - or should I be combining them into a single analysis?
Also .. is my interaction modelling at the lower level correct?
Thanks
James
|