Dear experts, 

We are working with a gPPI model and we have doubts about the ROIs which we are working with. 
We have done previously an analysis with CAT12 with the aim to extract the ROIs in which our patients have atrophy compared to healthy controls. 
After that, we have selected also functional ROIs from a network atlas.
Can we study with a gPPI model in CONN the connectivity from atrophy ROIs to functional ROIs? Is it correct? 
If it is possible, how we have to do the contrasts in the first and second level analyses?

Thank you in advance, 

Best regards

Naiara