Holger Wiese¹, Linda H. Lidborg¹, A. Mike Burton²; ¹Durham University, ²University of York

Humans know literally thousands of faces. Accordingly, the question arises why we so rarely mix them up. As a potential answer, the representation of a familiar face, once activated, may inhibit potentially competing representations, therefore focusing activation on the best match. In contrast to this suggestion, we recently demonstrated that two face representations can be activated simultaneously, and without apparent cost, when the stimuli are presented at different locations. Here, we tested whether simultaneous activation or inhibition is observed when two faces are presented sequentially, but in the same location. We used event-related brain potentials in a repetition priming paradigm and presented two prime stimuli in sequence, each shown for 200 ms, followed by a fixation cross (600 ms) and the target (1,000 ms). While we never repeated specific images, either the first, the second, or neither of the two primes could show the same identity as the target. In Experiment 1, using only familiar faces, we found more negative amplitudes for repetitions of the second prime relative to non-repetitions at occipito-temporal channels from approximately 300 ms after target presentation. No corresponding priming effect was observed for the repetition of the first prime, suggesting inhibition by the second stimulus. In Experiment 2, primes could be either familiar or unfamiliar, while all targets were familiar faces. Again, no priming was observed for the first prime. Critically, amplitudes to familiar first-prime repetitions were less negative when the second prime was familiar relative to unfamiliar, suggesting stronger interference in the former relative to the latter condition. We conclude that inhibition of competing familiar face representations does occur, potentially focusing activation on a single best match. The comparison to our previous results further suggests that such inhibition builds up over time and is effective with a 200 ms stimulus asynchrony.

Acknowledgements: Funded by Economic and Social Research Council (ESRC, UK) grant ES/X014657/1 to HW and AMB.