The second part of this series demonstrate additional preprocessing steps. Specifically, it addresses the problem of eye artifacts which are omnipresent in EEG recordings. It also demonstartes a procedure for repairing and rejecting noise-contaminated channels and segments.

Preprocessing is an important and controversial topic in EEG research. Here, I discuss it’s necessity and present a minimal preprocessing pipeline that deals with the most common sources of noise while avoiding to distort the data. I demonstrate each step using publicly available data.

We investigated how the human brain represents sound elevation by measuring brain responses to sounds from different locations using lectroencephalography. We show that elevation can be decoded from those responses and that decoding predicts localization accuracy.