Neural dynamics of racial categorization predicts racial bias in face recognition and altruism.

The classification of individuals into different racial groups provides a precondition for racial bias in cognition and behaviour, but how the brain enables spontaneous racial categorization is not fully understood. Here using multimodal brain imaging measures, including electroencephalography, functional magnetic resonance imaging and magnetoencephalography, we probe the neural dynamics of racial categorization by quantifying the repetition suppression of neural responses to faces of different individuals of each racial group (that is, Asian, black or white). We show that categorization of other-race faces engages early two-stage dynamic activities in neural networks consisting of multiple interactive brain regions. Categorization of same-race faces, however, recruits a different and simple network in a later time window. Dynamic neural activities involved in racial categorization predict racial biases in face recognition and altruistic intention. These results suggest that there are distinct neural dynamics by which the brain sorts people into different racial groups as a social ground for cognition and action.