I am reading up on directed connectivity estimates (e.g. Granger causality, DTF, dDTF) in an attempt to find a method that allows me to appropriately convert channel-wise connectivity estimates into an ROI representation of connectivity estimates. In the fMRI community, I know it is common practice to form some sort of ROI-representation of the data (e.g. averaging voxel activity w/in each ROI) PRIOR to deriving connectivity estimates. I would like to avoid this for my data (using ECoG datasets), and instead would like to appropriately convert the connectivity estimates that are derived from single channels into ROI-ROI connectivity summaries. So far, I have found very little, and would appreciate any tips/resources/methods for dealing with this problem.