Hi,
I performed a two sample t-test on my dataset and when I choose p unc< 0.001 it shows min q<0.25 (differs by analysis ,ALFF,fALFF,ReHo etc). Now I used also CONN for same datasets and in CONN there are thresholds as "FDR-Analysis Level","FDR-Seed Level". First one takes whole brain and calculates FDR and the second one takes only ROI [i]-*and corresponding result when that ROI chosen as a seed*[/i] -for FDR calculation. Is there anything like that in AFNI? Moreover, I performed clustsim with 3dttest++ and it shows 3 cluster's alpha >0.1 (can consist of pure noise), <0.1 (I couldnt classify that because at GUI it says <<0.1 ) and 0.03 which can be considered as a meaningful. Now is my two-sample t-test result significant?
I know the FDR calculations of AFNI by the way. I read the topics about FDR in this message board. Hence, I am absolutely sure that I cant reach the q value of 0.05 which was my aim. (~55000 voxels of mask,)
Thanks,
Abdullah