Identify the reference signal of scalp EEG recording

In [10] we proposed two methods to identify the reference signal under the key assumption that the reference signal is independent from EEG sources. This assumption may be true for intracranial EEG with a scalp reference. However, it 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 or non-cephalic reference such as earlobe. In this paper, we select the earlobe of one subject as the reference site, propose one special experimental method for the subject by taping the reference electrode about once every second and record scalp EEG for some time period. In this way we try to artificially produce the reference signal which may be independent from scalp sources so that the two methods in [10] can be successfully applied. As a result, for upcoming scalp EEG without taping the reference electrode, we propose two simulation methods to automatically identify the reference signal without taping in real time. Our simulation results show that the corrected scalp EEG after identifying the reference signal and removing it from the original scalp EEG by means of each of two simulation methods not only completely removes obvious common artifacts, but also involves much more brain neural activities than the corrected EEG after removing average reference (AR) from the original scalp EEG. Moreover, the identified reference signal during no-taping is indeed correlated with scalp sources.