Lateralized costs of divided attention to faces.

We used a dual-task paradigm to study the relationship between lateralized face perception and attention by measuring the costs of dividing attention between faces viewed in opposite visual fields. Observers performed judgments of either the sex, orientation, or color of one (single-task, cued) or both (dual-task, uncued) faces in a tachistoscopically viewed pair. We observed dual-task costs (i.e., decreased accuracy for dual- relative to single-task performance) for categorical judgments of face sex (female/male) and orientation (upright/inverted), both of which necessitated visual processing of face-specific information. We did not observe costs for judgments of face color (red-tinted/greyscale), which could be performed without processing face-specific visual information per se. We also observed an unexpected "feature contrast" effect of category-incongruency for judgments of face sex, such that observers showed no dual-task cost when faces in a pair belonged to opposite categories (e.g., female/male) as compared with the same category (e.g., female/female). Finally, dual-task costs for judgments of face sex and orientation (but not color) showed a left visual field (LVF) advantage: dual-task costs were greater in the right visual field (RVF) than in the LVF. We interpret this LVF cost advantage for judgments of face sex and orientation as indicative of the type of visual processing needed to perform face-based judgments. Our results show, for the first time, that the LVF advantage in face perception is directly related to capacity limits induced by divided attention to faces.