"Perceptual salience, not affective meaning, drives facial expression detection"

Humans are assumed to prioritize facial expressions because of the evolutionary significance of their affective content. Here, we show that this assumption does not survive rigorous control of perceptual structure. Across three studies integrating preregistered visual search experiments, geometric modeling of facial expressions, and eye-tracking data, we tested whether detection advantages for happy and angry faces reflect low-level visual salience rather than emotion. To this end, we combined frequentist interval testing with a predefined smallest effect size of interest (SESOI) and Bayesian region-of-practical-equivalence (ROPE) analyses. Behavioral and oculomotor measures reliably predicted emotion category when expressions were presented at full, perceptually unbalanced intensity. By contrast, when expression intensity was matched, both interval testing and Bayesian ROPE analyses supported the absence of a practically meaningful effect. In addition, detection speed and accuracy were best explained by geometric deviation from a neutral face, independent of emotional category. These findings indicate that apparent emotion advantages in visual search arise from perceptual salience rather than privileged access to emotion, challenging evolutionary accounts of attentional prioritization.