Dear CONN experts,

I have a question about multiple comparisons correction when using different atlases and ROIs.

I have imported different ROI atlases (ex. AAL3, default CONN atlas, task activations) to do separate gPPI analyses, but I am now wondering if all of these atlases are used as targets in the second-level seed-to-voxel analysis? I am currently trying to use only the AAL3 atlas and have selected my AAL3 sources, but when I look at the resultsROI_Condition00*.mat files, it looks like all of my ROIs were used as targets (300+).

I'm concerned that the seed-level FDR correction is being diluted by the large number of non-AAL3 target ROIs.

My questions:

• Is there a way to specify which ROI atlas should be used as targets for seed-to-voxel analysis, or does CONN automatically use all imported atlases as targets?


•  If CONN uses all imported ROIs as targets by default, can I apply FDR correction only to a subset of target ROIs (e.g., only AAL3-AAL3 connections)?


• Would it be better to remove the other atlases entirely and re-run the analysis with only AAL3, or is there a way to control this at the second-level analysis stage?

Any guidance would be greatly appreciated!

Thank you,

Gloria