Dear SPM experts,
I am reading about the pre-processing steps for DCM of resting state analysis and I come across the regression of discrete cosine basis functions (to be included in the GLM, 0.0078-0.1Hz). For what I have understood this is a way to model the resting state frequency fluctuations and filter out unwanted frequencies (band-pass filtering). In some papers, this step is followed by HP filtering of the data (to remove low frequencies, <0.0078Hz). Could you please clarify the need of this HP filtering? Is the modelling using discrete cosine functions not enough to model/clean the data?
Thank you very much,
Kind regards,
Bianca
|