> Just for clarification, if one would want to use the full shape information for the group analysis.
> you would take the -iresp outputs and run 3dDeconvolve again for every subject? (because
> unfortunately i dont think -iresp comes with a b and t- value).
> Or how about using the -fitts time series?

If you estimate the hemodynamic response with basis functions such as TENT or CSPLIN at the individual level, you can directly take the beta values for group analysis. If you use SPMG2/3, you may use the output from -iresp (beta values based on the reconstructed response from the SPM basis functions) for group analysis. In the case of TENT or CSPLIN, the output from -iresp is basically the same as those beta values associated with the basis functions.

Gang