Dear SPMers,
we incorporated an emotional subsequent memory paradigm in two groups of patients. My question regards the main and interaction effects in this design. The design has 2 (EMOTION: emotional / neutral) x 2 (MEMORY: subsequently remembered / subsequently not-remembered) x 2 (GROUP: patients a / patients b) design. My experience from previous fmri studies is that building the contrasts for interaction effects on the first level and using two-sample t-tests on the second level leads to more robust results, therefore I built the following conditions on the first level analysis:
1. All stimuli > implicit baseline
2. Emotional > non-emotional
3. Remembered > non-remembered
4. Emotional remembered > emotional non-remembered
5. Non-emotional remembered > non-emotional non-remembered
I addressed the main effects as follows:
1. EMOTION: 1-sample t-test for ‘Emotional > non-emotional’
2. MEMORY : 1-sample t-test for ‘Remembered > non-remembered’
3. GROUP: 2-sample t-test ‘All stimuli > implicit baseline’
To address GROUP x EMOTION & GROUP x MEMORY effects:
1. GROUP x EMOTION: 2-sample t-test for ‘Emotional > non-emotional’
2. GROUP x MEMORY: 2-sample t-test for `Remembered > non-remembered’
To address GROUP x EMOTION x MEMORY effects:
1. 2-sample t-test for ‘emotional-remembered > emotional non-remembered’
2. 2-sample t-test for ‘non-emotional remembered > non-emotional non-remembered’
Is this correct?
Thanks in advance & best regards
Ben
|