Hi Alfonso and CONN experts,

Following up on this question from a few years ago, it seems that primarily the gPPI modelling in the CONN toolbox (and all other gPPI toolboxes) is still directional, do you know if there is any way yet to circumvent this and get a symmetric ROIxROI correlation matrix for task-based connectivity in CONN or using another tool?

If not, do you know if it is acceptable to calculate the average based on RegionA - RegionB and RegionB - RegionA in order to obtain a approximate measure of this? Or can one pick only the strongest connection as is done in DWI?

I was also wondering why it is important to use bivariate regression as opposed to bivariate correlation for gPPI models (as is recommended in the manual)? I am asking this as I noticed that I get the same matrices from selecting gPPI and bivariate correlation and examining the matrices found in /results/firstlevel/SBC_01/results...ROI_Condition001.mat and if I use the graph theory toolbox, set my threshold to 0, select correlation coefficients, then select "export data" and view results.Z. Does this mean that the graph theory option calculates bivariate correlations even in the case of gPPI and is this still a valid measure of connectivity in this case or is it also not recommended to use graph theory analyses on gPPI data? Lastly, I also noticed in the results.Z matrices the values are all positive values even if I select two-way in the gui, is there a reason for these correlations being expressed as only positive in the graph theory portion of the toolbox? 

My ultimate aim is to produce large network-based correlation matrices (ideally symmetric) for specific task condition contrasts that I can use post-hoc to examine large-scale connectivity associated with specific functions. I know I can do this for each ROI individually using the  the "import values" option at the second-level covariate and could build my own matrix this way but was wondering if there was another way (such as using the graph theory toolbox) to explore large-scale networks within the gPPI framework.

Thank you so much for your time and thank you for making such a great multipurpose connectivity toolbox!

Isabella