[Eeglablist] ICA Strategies

[Eli Mikkelsen]

Hello everyone,
I’m attempting to use ICA for artifact rejection.  I’m wondering, what would 
be a typical number of components to exclude?  How many components are 
usually rejected because of eye-related noise and how many from muscle noise 
elsewhere on the head?  I’m using the criteria designated in the manual for 
determining brain-related activity; namely dipole-like scalp maps, 
appropriate frequency spectra, and regular erp images. Are there any other 
strategies that might be helpful?

much thanks,

Eli