Results of the dynamic face localizer task and repetition suppression parameter estimates extracted from the FFA. A, The spatial extent of the key ROIs of interest, the FFA, and the OFA, presented as degree of spatial overlap between participants in normalized space. We generated these ROIs with the group-constrained subject-specific (GcSS) approach (Julian et al., 2012) for the contrast of faces > other categories at the level of each individual participant. For a table listing the full ROI list, see Extended Data Figure 3-6. B, Repetition suppression parameter estimates in the FFA for different-face and repeated-face pairs for Black and White male faces. Upper panel: to present repetition suppression effects, we subtracted neural activity in response to single faces (baseline) from the neural response to the different conditions (differ, repeat), separately for each race (Black, White). Negative values indicate neural suppression compared with baseline. Across all figures, individual dots represent neural suppression for unique participants. Each figure also visualizes the mean of each condition (as a red dot), the median (solid horizontal line), and first and third quartiles (boxplot). Lower panel: mean effect size (the difference in suppression effect) and the bootstrapped 95% confidence intervals for the comparison between Black and White targets in each condition. The results demonstrate more release from suppression (i.e., less negative values) for different White targets compared with different Black targets, but no difference in suppression for repeating Black and White targets. This pattern suggests that the FFA forms unique representations for different individuals, but only for members of one’s own racial group. We report the results of the parallel analysis in the OFA and visual cortex in Figure 4. Extended Data Figure Figures 3-1, 3-2, 3-3, 3-4 demonstrate the robustness of the effect to different analytic strategies. In Extended Data Figure 3-5, we report the results of the main analysis with female faces.