Abstract
<p>Dynamic face perception depends on how local information from moving face parts is used within the broader configuration of the face. The present study used a temporal-direction discrimination task to examine the part-based neural contributions of the eyes and mouth during dynamic face processing. Across two experiments, Turkish and non-Turkish speakers judged whether muted videos of neutral faces articulating Turkish speech were presented forward or backward in time, while the eye and mouth regions were independently frequency-tagged in the EEG. Face orientation and blink presence were manipulated to assess how canonical face configuration and local dynamic cues shape part-based responses. Behaviorally, participants discriminated temporal direction above chance, with higher sensitivity for upright faces and with-blink face videos. Turkish speakers showed higher sensitivity than non-Turkish speakers specifically for upright faces, suggesting that language familiarity facilitated behavioral judgments when speech-related facial information was available in its canonical configuration. In the EEG, frequency-tagged responses were stronger for the mouth than for the eye, possibly reflecting the more sustained visual input generated by continuous articulatory mouth movements compared with the more transient motion in the eye region. Blink presence selectively enhanced eye-related responses, and face orientation modulated the differentiation between eye- and mouth-related responses. However, temporal direction did not produce a robust overall difference in frequency-tagged response strength, and language group did not modulate the tagged responses. Overall, the findings suggest that temporal-direction discrimination relies on how eye- and mouth-related cues are used within the face configuration, while frequency-tagged EEG responses reveal condition-dependent differences in the visual encoding of these dynamic face parts.</p>