Abstract
Human beings are experts at processing faces, with a striking impairment for other-race (OR) faces, the so-called other-race effect (ORE). Despite the robustness of the ORE at the behavioural level, Event Related Potential (ERP) studies have failed to consistently show sensitivity to race for the early category-specific N170 component. The observed discrepancy between electrophysiological findings is currently unresolved. Such inconsistency could be related to the heterogeneity of the task constraints, as well as to the lack of control for evident differences in low-level properties across races. In addition, most of these studies have only adopted one group of observers (usually, Western Caucasians). These limitations question whether previous findings can be generalized to the entire population and leave debate on the sensitivity of the N170 to race wide open. Surprisingly, no ERP studies have attempted to investigate the ORE with a well established marker of visual expertise, such as the Face Inversion Effect (FIE).
Here we exploited the FIE to investigate whether the N170 is sensitive to race. We recorded electrophysiological signals of 15 Western Caucasian and 15 East Asian observers while presented with Western Caucasian, East Asian and African American faces in two orientations (i.e., upright and inverted). To control for potential confounds in the ERP signal relating to differences in low-level properties across the race of the faces, we normalized amplitude-spectra and contrast across all images. Given the heterogeneous nature of the distributions, statistical analyses were performed using both classic statistical tests and robust approaches (i.e., percentile bootstrap). No differences on the N170 were observed in the upright conditions. Critically, inverting faces elicited larger amplitudes on the visual category of greater expertise (i.e., SR faces). These observations indicate a finer-grained neural tuning for SR faces at early stages of face processing.