Abstract
Previous studies have shown that the detrimental effect of negative contrast information on familiar face recognition is largely eliminated by retaining a contrast positive eye-region, creating so-called ‘contrast chimeras’ (Gilad et al., 2009). Event-related brain potentials, however, revealed that contrast chimeras elicit an enhanced N170, suggesting inefficient face detection, and no repetition priming in the N250r (Wiese et al., 2019). This latter finding was interpreted as suggesting that contrast-positive eyes, when presented in an otherwise negative face, are insufficient to activate representations relevant for recognition. Alternatively, however, this absent N250r might reflect a knock-on effect from disrupted face detection at the earlier N170 stage, interfering with subsequent stages of identity recognition. Here, we created ‘blur chimeras’, with face regions except for the eyes spatially low-pass filtered at six cycles/image. We hypothesized that this would reduce accessibility to detailed recognition-relevant information, while retaining typical face detection efficiency. Participants were presented with 240 prime/target pairs of celebrity faces, either unfiltered, fully blurred, or as blur chimeras. In half of the pairs, identity was repeated, while the other half consisted of non-repetitions. Unfiltered pairs provided a typical prime N170, and an N250r indicating clear identity repetition priming. Moreover, unlike fully blurred faces, blur chimeras elicited a similar N170 to unfiltered faces. Crucially, blur chimeras but not fully blurred faces elicited a significant N250r, demonstrating identity repetition priming. Compared to unfiltered faces, however, the blur chimera N250r was slightly delayed, suggesting less efficient processing. We conclude that detailed information from the eye-region in an otherwise low-pass filtered face is sufficient to activate familiar face representations, provided that the manipulation allows for typical face detection efficiency. Despite this, processing identity with detailed information only from the eye-region is inefficient, and additional facial details facilitate recognition.