search

nipreps / fmriprep

fMRIPrep is a robust and easy-to-use pipeline for preprocessing of diverse fMRI data. The transparent workflow dispenses of manual intervention, thereby ensuring the reproducibility of the results.
https://fmriprep.org
Apache License 2.0
637 stars 295 forks source link

Option for custom cutoff frequency in high pass filter #1961

Open zkhan12 opened 4 years ago

zkhan12 commented 4 years ago

Hey, Currently the cosine basis for the high pass filter cutoff is set at 0.008 Hz. There could be an option to specify a user defined cutoff for this filter.

jdkent commented 4 years ago

(I asked zkhan12 to open this issue/pull request). The use case for this would be for variable trial(block) lengths for task fmri. The length of the block determines what a reasonable high pass filter would be. The choice of high pass filter can be "wrong" if the frequency of the block is lower than the high pass filter, removing the activation of the block.

If there was more structure around task events in the bids specification, then perhaps the frequency of blocks could be determined and a high-pass filter could be chosen algorithmically.

ref: mit mindhive temporal filtering

MaurizioSicorello commented 3 months ago

If I searched correctly, changing the 128s default of the high-pass filter is still not supported by fmriprep. I would love for this feature to be implemented, as it would be really useful to directly be able to (a) use the drift regressor and (b) compCor regressors.