Independence verification for reference signal under neck of human body in EEG recordings

In our previous work [8] we proposed two methods to identify the reference electrode signal under the key assumption that the reference signal is independent from EEG sources. Although this assumption is shown to be possibly true for intracranial EEG with a scalp reference. However, this assumption may not be true for scalp sources with a scalp reference electrode. As such, this limits the application of our methods to the vast majority of scalp EEG with cephalic reference. In this paper, we conduct analysis for five reference electrode locations of one subject: left thumb, left arm, left shoulder, left chest and left back. Since these locations are far away from the head of the subject, the real reference signal from these locations should be independent from all the scalp sources. As such, the second method in [8] should be able to be applied to identify the reference signal well. Our simulation results demonstrate that the corrected EEG after removing the obtained reference signal involves much more brain neural activities for anyone of these five reference locations than the corrected EEG after removing the average reference (AR), and the dominated periodical ECG artifacts in original EEG are completely removed out. This fact actually provides a strong evidence to support the above assumption of independence. The results in this paper suggest that rather rod reference sites such as hand, shoulder, chest, back, etc., indeed turn out to be much better than commonly suggested.