[Eeglablist] Number of independent components to removed in EEG preprocessing before different analysis method

[周云晖]

Dear colleagues, 



I have searched the entire eeglablist archieve about the number of independent components (ICs) to remove during preprocessing. Many have mentioned that there are, usually, 5-20 ICs that are clearly originated from neural activity. However, there seems to be less agreement on how many number of ICs to remove in preprocessing.


Here I wonder whether the number of ICs to remove dependents on the goal of experiment, i.e., whether the plan is to do ERP, connectivity, time-frequency, EEG decoding, microstate, or some other analysis. Are there certain type of analysis that may require "cleaner data" than others? Are there any common practice about number of ICs to remove in each field of analysis method?


Thanks for any discussion.





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20181105/32701d77/attachment.html>