Separating early and intermediate vision EEG signatures of facial emotion recognition in a fast divided attention task to eye and mouth features.

Introduction: Face emotion recognition is a core component of social cognition. The N170 event-related potential (ERP) is a well-known neural marker of face perception. However, the sensitivity of the N170 and other early and intermediate ERP components, such as the P100 and P250, to the emotional content of facial expressions remains debated, possibly due to the influence of attentional mechanisms. We investigated whether early facial emotion perception can be tracked by modulation of these EEG signatures when participants are required to selectively process both eye gaze and mouth emotional cues within a very short time frame. Methods: EEG data were recorded while 20 participants performed an emotion recognition task involving neutral, happy, and sad facial expressions. The task required the integration of gaze and expression cues within a brief 350-ms presentation, ensuring focused attentional deployment. We tested three hypotheses: (1) the P100 does not encode emotional expressions despite being modulated by attention; (2) the N170 is sensitive to emotional facial expressions; and (3) the P250 reflects higher-level emotional processing. Results: The P100 was not sensitive to facial emotional content. In contrast, both the amplitude and latency of the N170 were modulated by emotional expression, with larger amplitudes for happy compared with neutral and sad expressions, and longer latencies for happy and sad than for neutral faces. The P250 was also modulated by emotion, with sad faces eliciting lower amplitudes than happy and neutral faces. Discussion: These findings indicate that the P100 does not encode emotional face recognition when selective attention is controlled for. In contrast, emotional facial expressions elicit distinct N170 and P250 responses, supporting the involvement of early and intermediate neural systems in emotional face processing.