Dear all,

I would like to know if cutting a time-series with 3dTcat (e.g., removing the first 300 TRs in a 3000 TR time-series) yields a minimally different time-series if the cutting is done before vs. after using AFNI proc concerning the power spectrum.

Let me make it more clear with two cases.
1. I cut the time-series before preprocessing with 3dTcat (or I simply use the "-tcat_remove_first_trs" option in AFNI proc) to remove the first x number of TRs.
2. I preprocess the time-series with AFNI proc, and cut it afterwards with 3dTcat.

Would the resulting power spectrum, e.g., via using 3dPeriodogram, be identical?

I read that a fundamental or basic wave can serve as a reference for Fourier analysis. That wave is the longest wave available and its usually the length of the record (in our case, the length of the fMRI run – all sample points). If AFNI proc handles preprocessing like this, I could imagine that the power spectra between case 1 and 2 could differ, even if the difference is only minimal.

Let me know what you think,
Philipp



Edited 1 time(s). Last edit at 05/15/2022 01:55PM by Philipp.