Time-varying networks of ERPs in P300-speller paradigms based on spatially and semantically congruent audiovisual bimodality.

Objective: In the P300-speller paradigm using spatially and semantically congruent audiovisual (AV) stimuli, AV stimuli elicited significantly higher event-related potential (ERP) amplitudes than those for a single visual (V) stimulus. Nevertheless, it remains unclear whether ERPs in AV and V spelling paradigms are generated with identical brain network architecture. Approach: In this study, we constructed time-varying networks for ERPs in AV and V spelling paradigms based on adaptive directed transfer function to investigate the dynamic processes underpinning the processing of stimuli in the two spelling paradigms. Main results: Our analysis revealed that early AV integration, AV spatial and semantic information, and late AV integration enhanced activation in attention-related areas between 40–160 ms, areas associated with attention to target stimuli between 200–280 ms, and areas associated with decision-making between 320–520 ms, respectively. Left temporal areas were associated with AV spatial-semantic information-processing and late AV integration, the activation of which impacted the activation of brain areas associated with attention (P3a) and decision-making (P3b). In addition, the ability and efficiency of information transmission from brain networks in the AV spelling paradigm were significantly stronger than those in the V spelling paradigm between 40 and 160 ms, 200 and 280 ms, and 320 and 520 ms. Significance: This work provides a theoretical basis for deepening our understanding of the neural mechanisms underscoring ERPs production in the AV P300-speller paradigm as well as a theoretical reference for further optimization of the induced paradigm based on AV stimuli and improves the performance of BCI system.