Dear all,
I am wondering how can pmods be analyzed in 2nd level. Normally, we
contrast effects in first level and use con images in 2nd level.
But pmods have [in first level] "1" for increase and "-1" for decrease
in activity.
Shall I use both of these con images in the 2nd level (i.e., factor
"increase/decrease")?
Or shall I just use une contrast image (contrast "1") which contains
both sides of activity?
(In fact, when I open a con image with xjView, I can see both sides of
activity. Therefore I would think any con image from a pmod regressor
contains both sides. Trying to use the "increase/decrease" factor
produced a useless model, with all betas not uniquely estimatable).
If my question is not clear please let me know.
|