Question about interpolation and big artifact removal

[juliapottkaemper]

Hello,

I have been using this pipeline for the last weeks to tweak it for our study purposes (fully automated EEG preprocessing pipeline) and I had some remaining questions about it. Some are just for clarification, if I understood it all correctly. I hope that you can help me out. If I need to post this on another forum, please let me know.

Here are some questions:

1. I have multiple resting-state EEG sessions (eyes closed and eyes open) per participant. I already discovered on forums that it is not possible to handle mulitple resting state sessions, so I adapted my script to run the pipeline twice fully (once for eyes closed and eyes open). Please let me know, if this is still the case, or if I could do it any other way.

2. Is it correct that there will be no interpolation performed of previously marked 'bad' channels (based on pyprep that I ran before mne-bids-pipeline)? Is there any wat to either way perform interpolation? pyprep works fine to exclude bad channels, so I would like to leave this step in as mne-bids does not have any option to exclude channels in the pipeline (as I read).

3. I implemented the 'extended_infomax' ICA method, as the other methods did not seem to get rid of eye-blink artifacts (I played with several parameters but this is what worked best). However, I am still struggling removing large amplitude artifact, which should have been removed with the option 'autoreject_local'. I already tried it with the more conservative option 'autoreject_global', but it does not seem to work at all to remove epochs that contain artifacts. Now I am doubting whether I am looking at the right files to begin with, but I guess these are the right files:
   sub-02/eeg/sub-02_task-restEyesClosed_proc-clean_epo.fif
   sub-02/eeg/sub-02_task-restEyesClosed_proc-clean_raw.fif

Do I understand it correctly that these files include the cleaned epoch file (where epochs that are 'bad' should be greyish?) and the raw time series without epochs? If I compare the clean file with the input file, it does not seem to have dropped epochs, however, in the report.html file I see that some epochs were dropped. So maybe it is an issue of viewing the file (I used mne.io.read_raw_fif() and mne.read_epochs())? Please correct me if I am wrong.

Thanks already for advice and suggestions. If I can provide you with any details, please let me know.

I attached my custom_py file.
custom_config_eyesclosed.json

Kind regards,
Julia