Hello.

What is the default response model if I don't use the option -stim_times?
Without using -stim_times to specify the response models, sometimes I got a regular shape of the IRF (e.g. the center voxel in the attached figure) but more often not. I was wondering whether the irregular IRF shapes were caused by the data itself or partly by the specified response model. It is not clear to me how to use the option -stim_times in 3dDeconvolv and how big difference it can make. What is the guideline to choose a response model (e.g. GAM, TENT, POLY etc.), and how does it matters? Another minor question is, with the stimulus functions given by -stim_file and the value of TR by -TR_times, why 3dDeconvolv still needs to specify input stimulus timing in tname in the option -stim_times. It seems to be redundant information, otherwise I may misunderstand the purpose of this tname file.

There is another relevant question. If the shape of IRF is random in the region of interest, how do we decide the values for -stim_minlag and -stim_maxlag in the formal process?


Zhuang