> I want to know if I should use stim-times to produce IRFs and run the Anova
> with those.
How did you model the hemodynamic response functions? With a fixed-shape IRF or a set of basis functions?
> Or should I create a bucket file from 3dDeconvolve and run the anova on the
> Tstat and/or the Coef/Beta files? Or is both ways fine?
The currently available programs only take beta values for group analysis. I'm working on a new program that will take both beta and the corresponding t-statistic (preferably from 3dREMLfit) for group analysis.
Gang