Hello! I am currently formatting a dataset of simultaneous EEG-fMRI recordings into BIDS format and I want to ensure I’m handling the EEG data correctly. I have the data at various stages of preprocessing and have a couple of questions regarding best practices:
The “SoftwareFilters” parameter in the eeg.json file: If my EEG data has already undergone gradient artifact correction, should this be specified in the “SoftwareParameters” field?
Raw vs. preprocessed data: what is the rule of thumb for sharing simultaneous EEG-fMRI datasets? Should the main BIDS root contain the absolute raw data (no gradient artifact correction, no bandpass/notch filters) or is it acceptable to include EEG data that has already undergone basic filtering (MRI artifact removal, 1 Hz highness, 70Hz lowpass, 60 Hz notch) as the primary dataset?
Any guidance would be greatly appreciated thank you!
Let’s say that if the filtering was done by your acquisition device and given to you like this, then it can go in raw data otherwise (a human did artefact removal or some other scripts or tool was run on the data) it is considered derivatives.
For examle, in the world of MRI, in most cases motion corrected data for fMRI would be considered derivatives but some scanner can and do give you a MoCo (motion corrected) version of your data: in this case the MoCo could be considered raw data.