Dear Dennis,
Thanks a lot for the answer! What do you mean with 'generally'? I refer to the high-pass filter as implemented in spm_filter.m, where the cutoff frequency is given a priori, either as default (128) or as specified by the user (via the HParam parameter). I believe that the formula used in spm_filter.m to calculate the order of the filter:
n = fix(2*(k*K(s).RT)/K(s).HParam + 1);
means that if the scanning time (number_of_volumes*TR) is a multiple of HParam, then the signal corresponding to HParam will be removed too, contrary to what the SPM manual suggests. I have also checked that running analyses with different cutoffs. I am wondering whether anyone has had any thoughts on that?
Best,
Wiktor Olszowy
|