Hi,

They are the same. Typically for functional connectivity analyses I recommend using bandpass as part of the denoising step (e.g. using RegBP) but in some cases (e.g. when using conn_module('PREP',...) to preprocess your data for non-connectivity analyses outside of CONN and you are planning to skip denoising) using functional_bandpass allows you to still bandpass filter your data if needed as part of your preprocessing pipeline (in those cases denoising may be placed as one of the latest steps, e.g. it could go before or after spatial smoothing, both should produce the same result). In all cases to specify a low-pass filter simply set the low-frequency limit to 0 and to specify a high-pass filter set the high-frequency limit to inf (or to the Nyquist frequency of your sample), e.g. [0 0.1] is a 0.1Hz low-pass filter, and [0.1 inf] is a 0.1Hz high-pass filter.

Hope this helps

Alfonso

Originally posted by Shengwei Zhang:

Hello, just curious what the difference between functional_bandpass (one of individual functional steps) and the later denoising bandpass (as in batch.Denoising.filter) is? If the functional_bandpass is chosen, where should it be among the individual steps (e.g. after functional_smooth) and if it's possible to make it as high/low pass instead? Thanks. Shengwei