ABSTRACT Using a public database, the research combined event-related potentials (ERP), time-frequency analysis (TFR) and machine learning to compare brain activities during Confusion and Unconfusion states. The findings revealed that processing ambiguous emotions decreased LPP amplitude compared to the Unconfusion state, supporting LPP as a biomarker for emotional ambiguity. Additionally, fear-related ambiguity showed greater beta wave power than happiness, indicating heightened brain sensitivity to threats. Machine learning analysis further demonstrated that fear-related features were more critical in distinguishing emotional ambiguity than happiness, corroborating non-phase-locked signal results. These results show that emotion types, particularly fear, significantly modulate brain activity in processing emotional ambiguity, highlighting increased sensitivity and adaptability to fear-related ambiguity. This research enhances our understanding of cognitive processing in emotional ambiguity and offers insights into the neurobiological basis and therapeutic strategies for emotional disorders.
Read full abstract