One frequently studied biomarker for health and disease conditions is the P3 component extracted from scalp recorded electroencephalography (EEG). The spatial origin of this significant neural signal is known to be distributed, typically involving large regions of the cerebral cortex as well as subcortical structures. Unlike the temporal characterization of the P3 by amplitude or latency measures from event-related potentials (ERPs), the spatial characterization of the P3 component is relatively rare, typically qualitative, and often reported as differences between populations (group differences between healthy controls and clinical groups). Here we introduce a novel approach to quantitatively characterize the spatial origin of the P3 component by (1) applying second-order blind identification (SOBI) to continuous, high-density EEG data to extract the P3 component, (2) modeling the underlying generators of the SOBI P3 component as a set of equivalent current dipoles (ECDs) in Talairach space using BESA; (3) using the application Talairach Client to determine the hits associated with the anatomical structures at three level of resolution (lobe, gyrus, and cell type). We show that the hits information provided by Talairach Client can enable a quantitative characterization of the spatial configuration of the network underlying the P3 component (P3N) via two quantities: cross-individual reliability (or consistency) of a given brain structure as a part of the P3N, and within-individual contribution of a given brain structure to the whole P3N network. We suggest that this method may be used to further differentiate individuals in the absence of differences in P3 amplitude or latency, or when scientific questions or practical application cannot be supported by a yes-no answer regarding the source of a P3 component. Finally, application of our method to a group of 13 participants revealed that frontal structures, particularly BA10, play a special role in the function of a global cortical network underlying novelty processing.
bioRxiv Subject Collection: Neuroscience