Dear SPMers,
Iām analyzing task-based fMRI from an emotional perception task in patients & controls. The design has 6 conditions (5 emotions plus neutral which serves as a control condition) & two groups (healthy & patients). Main effects of group has been analyzed using a two-sample t-test.
To examine interaction effects between the emotional conditions & group I used a flexible factorial design (subjects / groups / conditions). I ran different analysis:
(1) A flexible factorial design with all emotional conditions > control (=5 conditions). Can I use the following f-contrast to examine interaction effects:
-1 -0.5 0 0.5 1 1 0.5 0 -0.5 1
(2) Because the control condition itself might be Affected by the disorder I ran another flexible factorial design using the 5 conditions & the baseline condition (versus implicit baseline) (=6 conditions). However, in this design I end up with an even number of conditions ā using the same contrast approach as above would result in:
-1 -0.5 0 0 0.5 1 1 0.5 0 0 -0.5 -1
ļ However, this would not result in uniquely specified conditions ā right?
(3) Is it possible to build a contrast in the design described in (2) that accounts for the control condition? In (1) I contrasted emotion > control on the first level, which should lead to more specific findings, since it controls for the variance between subjects on the first level. However, if the control conditions is different between healthy controls & participants this will bias my results.
Thanks once more & best regards,
Ben
|