Disentangling Event-Related Potential and Oscillatory Sources using Time- and Frequency-Domain Source Imaging.

Objective: To determine whether event-related potentials (ERPs) and neural oscillations arise from common or distinct neural sources by combining two EEG source localization methods. Methods: We applied eLORETA and DICS, which localize activity in the time and frequency domains respectively, to both simulated and real EEG data. Simulations assessed the spatial accuracy and specificity of each method. Real data analyses were used to examine the correspondence between ERP-related activity and oscillatory dynamics in relevant frequency bands. Results: For the P300, eLORETA and DICS yielded overlapping source localizations, indicating a shared neural origin. Although alpha desynchronization temporally aligned with the P300, its sources were spatially distinct. In the case of the auditory N1, both methods revealed bilateral activation, with higher spatial precision observed using eLORETA. For the N400, DICS revealed sources that overlapped with eLORETA's localization, suggesting a distributed semantic network engaged through amplitude asymmetry mechanisms. Conclusions: ERP components may result from both shared and distinct oscillatory sources, varying by frequency band and cognitive function. The complementary use of eLORETA and DICS enhances interpretability of source dynamics. Significance: This study underscores the value of integrating time- and frequency-domain source imaging to deepen our mechanistic understanding of stimulus-locked brain responses.