Greetings AFNI community.
I have used 3dFourier with a high pass filter (to remove low frequency oscillations in the MR time series) and noted the shift in the scale of the resultant MR values.
I know that this alone should not affect calculation of percent signal change, but has anyone tested this empirically. Namely, does the re-scaling of the MR signal following the use of 3dFourier impact the use of either Gang Chen's method (normalize prior to 3dDeconvolve) or Doug Ward's method (implemented on the output of 3dDeconvolve) in a less than optimal manner?
philippe