Dear all,
I have two GLMs: The 1st one includes the outcome as the category regressor and the prediction error (PE) as the parametric modulator; the 2nd is similar to the 1st one, except separating the outcome as outcome_win and outcome_lose and the corresponding PE as positive and negative PE(PE+; PE-).
For the 1st GLM, I set 1 to get PE results; for the 2nd GM, I set 1/2 1/2 can also get a PE result.
I'm wondering why the two kinds of results can be quite different. If I want to get the PE result, which one I should use? Thanks in advance.
Best,
Sasa
|