The paper proposes a generative model for inferring directional EEG connectivity. The approach is sound and the manuscript is well written. The Reviewers agree that the charaterization of the proposed method is well supported by both simulated and real data. I would consider a minor concern the issue that there is no real exploitation of the concurrent fMRI/EEG acquisition since the analysis is designed as two independent steps of ROI estimate (fMRI) and connectivity inference (EEG). We may consider this as an open challenge in the research agenda rather than a serious pitfall of the proposed method. As pointed out by one Reviewer I would consider a more meaningful concern the lack of code.