Abstract
The underlying mechanisms for inter-individual differences in face recognition ability are still poorly understood. We investigated whether participants performing high or low in a Famous Face Recognition Test differ in effects of face priming. Identity self-priming typically results in behavioural benefits and consistent event-related potential (ERP) modulations, even when familiar target faces are preceded by different (Schweinberger et al., 2002) or geometrically distorted (Bindemann et al., 2008) images of the same identity. This argues for robust, largely image- independent representations of familiar faces, a prerequisite for the ability to put different images of the same identity together (Jenkins et al., 2011). Here, we tested in good and poor recognizers, whether face representations can also be activated by faces of different identities, when these look similar to familiar targets. In an immediate-repetition-priming paradigm, famous target faces were either preceded by i) a different image of the famous face, ii) the face of a “lookalike”, i.e. an unfamiliar face resembling the famous target face in appearance, or iii) a different famous face. Participants performed a face familiarity task on the targets. In addition to response times and accuracies, ERPs for target faces were analyzed. First, in line with previous studies, repetition priming effects occurred when famous faces were preceded by different images of the same faces (RT, ACC, P200, N250r, N400). Second, attenuated repetition priming effects were found when famous target faces were preceded by faces of unfamiliar lookalikes (RT, ACC, N250r, N400). Third, repetition priming effects in the N250r and N400 for lookalike primes were more reliable for participants with high face recognition skills. This suggests that i) mental representations in good recognizers are characterized by a larger flexibility, and ii) that high and low performers also differ at the level of post perceptual access to semantic information about familiar persons.
Meeting abstract presented at VSS 2015