> In this example (output of "3dinfo -verbose stats.ED+orig'[126]'" after
> analyzing one subject's data from what Mike Beauchamp provided for the
> tutorials), are the degrees of freedom 1194, and the number of pairs of
> observations 1195 (since the stat at this subbrick was a repeated measures
> t-test, df=n-1)?
No, not really. Since this is a regression model, the number of DFs for a t-test is
total number of data points - total number of regressors (including all covariates such as drifting effect - polynomials - regressors and head motion parameters)
Gang