Submission 633
Neural and Behavioral Correlates of Emotional Face Processing for Genuine and AI-Generated Expressions
SymposiumTalk-02
Presented by: Annika Ziereis
The proliferation of AI-generated facial expressions raises critical questions about how authenticity beliefs shape emotional face processing. This study re-analyzed single-trial EEG data from N = 40 participants from Ziereis and Schacht (2024, Scientific Reports) to examine how subjective beliefs about facial expression authenticity interact with neural processing of emotional content. While the original study focused on neural responses to genuine and artificially created emotional facial expressions (happy, angry, neutral) across different stimulus sizes (small, medium, large), the present re-analysis specifically investigated how participants’ beliefs about authenticity (real vs. fake) modulated event-related potentials (ERPs). Results revealed that early to late face-sensitive ERP components (P1, N170, EPN, LPC) were affected by the inherent stimulus properties and emotion, but the influence of subjective authenticity beliefs became increasingly pronounced at later processing stages. More consistent modulations emerged within the EPN time window, whereas the LPC exhibited more complex interactions involving both the perceived and inherent authenticity of expressions. These findings suggest that neural responses to faces are shaped not only by facial expressions but also by subjective evaluations of authenticity. This research highlights how authenticity judgments influence social and emotional perception in digital contexts and underscores the need to investigate subsequent neural and behavioral responses, such as memory and mimicry responses to artificial faces.