Dear SPM list
I write to ask for comments on an FIR model. My experiment uses a
sparse event-related design, presenting a stimulus every 14 seconds.
The TR is 3sec, 294TRs. So, the effective peristimulus sampling
frequency is 1Hz.
Is there a *correct* binsize and window for this experiment?
Specifically, is it preferable that the binsize equals either the
effective peristimulus sampling rate, or the TR? I presume that a
reasonable window for this experiment is 14 secs; is that correct?
I note that in the SPM5 manual, for example, the worked example has a
TR of 2 sec, an effective peristimulus sampling frequency of 2Hz but
uses 2 sec bins.
My apologies if this is trivial, or is documented elsewhere.
Thanks in advance;
Regards,
Simon Glynn
Fellow, Epilepsy
University of Pennsylvania
|