Dear all,

I am currently using the conn toolbox for gPPI analysis. I have analyzed my data until the second-level analysis and am now looking at the results.

I want to perform a cluster-FWE correction of 0.05. I am also using a voxel threshold of p > 0.001 uncorrected; however, in most conditions, I don't get any results. So I used a voxel threshold of 0.005 for example. I notice that in the results, there is still a peak p-unc that is less than 0.001 for all clusters. 

Can someone help me to understand what exactly CONN does in showing these results, and why the same cluster wouldn't show up, when I put a voxel threshhole of p > 0.001? Also, do you know any paper that is discussing this matter, especially for gPPI?

Thank you in advance!

Best Regards from Germany,

Negin