Hello Alfonso,

I am currently working with the CONN toolbox on a study involving patients with depression, where I am interested in the functional connectivity of the subcallosal cingulate cortex (SCC, ROI from the Harvard-Oxford atlas of CONN) using a seed-based connectivity approach in a pre/post-treatment design.

In addition, I am also examining:

• the functional connectivity between the SCC and the default mode network (DMN),


• as well as connectivity within the DMN itself (PCC, MPFC, LPr, LPl).

To do this, I performed a ROI-to-ROI analysis including the SCC and the four DMN regions. This is my first time working with fMRI data, and I am unsure about how to appropriately select the second-level analysis parameters in CONN.

I initially used the default option “customize (advanced family-wise error control settings)” and explored several cluster threshold choices (network-level, cluster-level with MVPA omnibus test or mass/intensity, ROI-level with MVPA omnibus test or mass/intensity, and “none”).

However, my results vary depending on the selected option: sometimes only MVPA analyses are significant, while other levels also yield significant findings. In my manuscript, I mentioned that I tested multiple levels, but I am not certain whether this is methodologically appropriate.

Could you please advise on the recommended approach for selecting and reporting these analyses? In particular, is there a preferred strategy for a design with a relatively small number of ROIs to ensure a statistically sound and coherent interpretation?

Thank you very much for your time and for developing CONN—it has been extremely helpful, even though I am still in the process of fully understanding these analyses.

Best regards,