Hi Afni Gurus,

I followed Gang's directions to run a simple correlation analysis using the time series from a seed voxel (http://afni.nimh.nih.gov/sscc/gangc/SimCorrAna.html), and then used each subjects' Z-transformed correlation map in a one-sample t-test (3dttest++).

The individual correlation maps looked good for each subject (similar values, similar regions), but the output of the t-test had outrageously significant T-values (if I normally threshold around p = .001 or .0001, I had to get down to 1 x 10^-13 to have anything meaningful). Is this similar to what you might expect? How does one pick a meaningful threshold in this case?

Thanks!