Hi-
I had questions about the instructions for this, as documented here:
[
afni.nimh.nih.gov]
My understanding is that the PPI model has to be a full model. But does that mean every psychological regressor needs an interaction term?
In modeling data from a probabilistic reversal learning task, for example, I have four events of interest:
1. Valid rewards leading to stays;
2. Valid punishments leading to stays;
3. Valid punishments leading to shifts;
4. Invalid punishments.
There are several other event-types of less interest:
1. Valid rewards leading to shifts;
2. Invalid rewards;
3. Non-responses
4. Outcomes followed by non-responses.
What is the full model then?
Is it a model with interaction terms between the seed region time series and regressors for the events of interest? All events? Just one or two events?
I am think in terms of the template script I saw on the web page above.
Thanks,
Jim Waltz
