 |
 |
| Hello, Could someone please help? In a PPI analysis when creating the PPI regressor it is recommended that each condition to have the same weight. For instance, if you have conditions A and B (of interest) and condition C (some sort of control) and would like to see the PPI interaction, psychological and physiological regressor one would define a matrix like: [1 1 1; 2 1 1; 3 1 -1]. Usually it is not recommended to have a matrix like: [1 1 1; 2 1 1; 3 1 -2]. My understanding from previous posts is that with such a matrix one would assume that connectivity is twice weaker in the control condition. I also understand that is better to create 3 PPI interactions, 3 psychological variables and one seed. But I can't though understand how would I show in the PPI the potential differences in these conditions (one is "high load" (A), the other is "low" (B) and the control (C), really does not impose any cognitive demand? And I still would like to collapse A and B and contrast with C. Also, how big is the bias for the contrast [1 1 1; 2 1 1; 3 1 -2]? Thank you for clarifying this, Rachel |