Hello Dr. Alfonso,

I have a query regarding the interpretation of second-level analysis results. I have behavioural variables as predictors and the functional connectivity as the outcome variable for the second-level analysis. Could you explain how the interpretation of second-level results would vary if I choose multivariate regression instead of bivariate regression in the first-level analysis tab? Could you provide a specific example where multivariate regression would be a better approach?

 I am using the Jülich Brain Atlas, where each region is parcellated into multiple subregions (e.g., the Insula is parcellated into 32 subregions). In such a case, would multivariate regression be more useful? 

If I enter 150 such regions from the Julich Brain Atlas in sources/seeds/ROI, will it control for the effect of all 149 seeds and show unique connectivity for just one region? Is entering 150 regions as ROIs of interest a good idea, given that the p-value may not survive after controlling for the number of comparisons?

Thank you for your time

Shruti