afni.nimh.nih.gov

Hi Woogul,

This is why it is good to scale your data (to a mean of 100) and your
regressors (usually to a peak of 1, unless the stimulus durations vary).

In your case, I would expect the regressors were large.

What is the output of this (assuming X.xmat.1D is the regression matrix):

3dTstat -max -prefix - X.xmat.1D\'

- rick