Investigating the Processing of Faces from Different Races and Artificially Created Minimal Groups using ERP and fMRI

In a series of six ERP and fMRI experiments, we investigated the time course, brain activation and functional connectivity changes in processing faces from different races and/or from different artificially created minimal groups (Tajfel, 1970). We found that the race of faces can be detected as early as ∼100 ms after face onset as reflected by greater occipito-temporal P100 to Black than White (same-race) faces (Experiments 1, 2, 4) and greater frontal N100 to White than Black faces (Experiments 2, 4). We observed Black-White differences in the frontal N200 (Experiments 1, 2, 4), which were associated with implicit racial associations (Experiment 1, He, Johnson, Dovidio, & McCarthy, 2009). In an fMRI study (Experiment 3, He, Johnson & McCarthy, under review), we observed greater occipital face area (OFA) activation to Black than White (same-race) faces. The OFA activation difference and the occipito-temporal P100 difference may provide both localization and timing information for this race differentiation process. In Experiment 2 (He et al., under review) and 4 (He, Johnson & McCarthy, in prep1), removal of skin color differences or all physical differences between Black and White faces did not eliminate the differentiation of race in early face processing stages. Merely labeling racially ambiguous faces as Black or White was sufficient to create the same early ERP differences as using real Black and White faces (Experiment 4). In Experiment 5 (He, Johnson & McCarthy, in prep2), we observed early P100 and N100 differences in processing faces of artificially created ingroup and outgroup members. In Experiment 6 (He et al., in prep2) when both race and minimal group information were presented, the occipito-temporal P100 responded to race but not to minimal group, while the frontal N100 responded to minimal groups but not to race. Different types of group information may be processed in parallel ways without interacting, but only race information generated later ERP differences (e.g., frontal N200). Finally, we propose a neural model to explain the timing and the potential neural pathways in processing faces from different races/groups by integrating the findings from our ERP and fMRI studies.