Abstract
Face recognition research has recently begun to examine identity learning by exposing participants to multiple, naturally varying images of a pre-experimentally unfamiliar person. Both behavioural and event-related potential (ERP) studies have shown that such exposure to within-person variability leads to the acquisition of stable representations which enable recognition of the learnt identities even from previously unseen images. While there is emerging evidence that identity information is initially harder to perceive in other-race faces, e.g., when different pictures of unfamiliar faces have to be matched for identity, it remains largely unknown whether similar difficulties also arise during identity learning, i.e., when participants become increasingly familiar with a face. Here, we present evidence from both behavioural and ERP experiments investigating learning of own- and other-race facial identities in Caucasian and Asian participants. While Caucasian participants showed better learning of own- than other-race identities in behavioural tasks, a similar own-race learning advantage was absent in Asian participants with substantial other-race contact. ERP results from Caucasian participants further revealed a more negative N250 component for the learnt own-race face images relative to images of novel identities, while the corresponding effect was reduced for other-race faces. The present results thus highlight the challenges to learn other-race identities. At the same time, our findings from Asian participants suggest that such difficulties can be overcome by extensive other-race contact. Implications for our understanding of the neural and cognitive processes involved in own- and other-race face recognition will be discussed.