Dear AFNI experts,

I have followed the website below to do seed functional connectivity.
[afni.nimh.nih.gov]

I tried two different regression models: One is to include the regressor of experimental condition, the detrended time series of seed region and their interaction regressor (seed X conditions); and the other is to include everything except the detrended time series of seed region. One main difference in the results of the interaction regressors between two models is that the seed region showed almost all negative connectivity in the first model whereas the seed region showed both positive and negative connectivity in the second model. And, this negative connectivity is observed across six seed regions in frontal and occipital regions. I had hard time to believe that all six regions could have anti-phase activity with the rest of the whole head especially when they all showed positive activation in the main effect of the experimental conditions. How should I implement the optimal PPI model for my data?

Another question I had is that what the principle is to choose the parameter of the polort in detrend and 3dDeconvolve.
My experiment is block design and had four blocks, each of which is about 60-second long.

Your suggestions will be appreciated! Thank you.

Veda