Flexible perception of face attributes under naturalistic visual constraints

Abstract

How does human brain adapt face perception to naturalistic visual constraints? While faces are typically perceived under dynamic and uncertain conditions, most studies use static presentation and enough temporal exposure, leaving the neural mechanisms of adaptive face processing poorly understood. Here, we employed steady-state evoked potential electroencephalography (EEG) to track how perception of multiple face attributes (age, emotion, gender, and race) adjusts to two ecologically relevant scenarios: (1) faces gradually sharpening from blur to clear (mimicking approaching from a distance), and (2) faces presented with incrementally increasing exposure times (mimicking brief, flashed encounters). By progressively increasing sensory input (via spatial frequency content in Experiment 1 and exposure time in Experiment 2) in a stepwise manner, we show that emotion and race categorization emerge early, even under high blur (e.g., 4.89 cycles/image) or brief exposures (e.g., 41.7-50 ms), reflecting coarse visual processing. In contrast, age discrimination requires higher clarity (e.g., 7.31 cycles/image) but shorter exposure (e.g., 41.7 ms) if clear images are presented. In both scenarios, gender processing exhibits the strongest dependence on clarity (e.g., 10.94 cycles/image) and time (e.g., 66.7 ms). Representational similarity analyses further show that reliable response patterns for emotion and race emerge earlier in the posterior brain region relative to those for gender in both experiments. Together, these results identify a flexible temporal order for face perception that adapts to naturalistic visual constraints, bridging the gap between controlled laboratory paradigms and naturalistic social vision.