[Eeglablist] Automatic EEG artifact removal pipeline for resting-state data (based on Pernet et. al 2021)

Gil Avila, Cristina cristina.gil at tum.de
Wed Dec 15 07:56:34 PST 2021


Dear EEGLab community,
I am very much interested in applying the automatic pipeline proposed in Pernet et. al, 2021 (https://urldefense.proofpoint.com/v2/url?u=https-3A__doi.org_10.3389_fnins.2020.610388&d=DwIFAw&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=26evuRPI7ihxkgneJMve7S7krYk1Lb33N5FoPyomPauMOSmRe_vKSpYDUwlq6nT6&s=R25GPlR_WvmJZikcg26xwwhhiLNDp4GEBgLVxGSiYPM&e= ) to resting state data. I am working with a 29 mobile dry electrode EEG system (CGX Quick-32r, Cogniomics) and plan to use the pipeline for large-scale dataset cleaning. I would like to contribute to the community by providing an easy solution for automatically cleaning resting-state data. Therefore, I would like to know your suggestions for adapting this pipeline to resting state in a setting with a lower number of electrodes.
So far, I tried the Pernet pipeline with default settings in 5-min eyes-closed recordings. However, when applying clean_raw_data() the number channels that were rejected is often higher that what I would expect based on visual inspection (on average 4 ± 2 channels out 29 are rejected).
I also tried to not reject any channels and be stricter on the independent components that I keep after applying ICLabel. That is, instead of rejecting components whose probability of being 'eye' or 'muscle' is higher than 0.8, keeping only components whose probability of being 'brain' is higher than > 0.8. After visual inspection, data looked mostly clean (except for one channel that went off in one recording) and I avoided loosing information by rejecting whole channels.
I was also wondering what do you think of the later approach. Do you have any other suggestions?
Best regards,
Cristina
--
Cristina Gil Ávila - PhD candidate
Department of Neurology
Technische Universität München
Munich, Germany
cristina.gil at tum.de<mailto:cristina.gil at tum.de>
painlabmunich.de<https://urldefense.proofpoint.com/v2/url?u=https-3A__www.painlabmunich.de_&d=DwIFAw&c=-35OiAkTchMrZOngvJPOeA&r=kB5f6DjXkuOQpM1bq5OFA9kKiQyNm1p6x6e36h3EglE&m=26evuRPI7ihxkgneJMve7S7krYk1Lb33N5FoPyomPauMOSmRe_vKSpYDUwlq6nT6&s=zsg2xxG2I-jLbcA3bEBdyD1K888FjrrodINLofpV3L8&e= >



More information about the eeglablist mailing list