Masked and unmasked electrophysiological repetition effects of famous faces

We investigated immediate repetition effects of sequentially presented famous face pairs. The first face (F1) was presented masked or unmasked and preceded the second face (F2) with different SOAs (84 ms vs. 500 ms). Participants judged F2 with regard to either semantic category (actor vs. singer; indirect task) or perceptual match with F1 (same vs. different; direct task). Repetition shortened RT for unmasked but not for masked F1 conditions. In event-related brain potentials (ERPs), unmasked repetition effects were influenced by task and SOA and consisted a modulation of an occipitotemporal N170, an inferior temporal N250r (200-300 ms), a central-parietal N400 (300-500 ms), and a parietal P600 (500-800 ms). An early occipital negativity (onset approximately 100 ms) was present at the 84-ms SOA but diminished in the 500-ms SOA condition, probably reflecting a fast decaying iconic memory trace. Masked repetition effects in the indirect task were limited to a significant early (100-150 ms) prefrontal/lateral frontal and central-parietal modulation, and a strong trend for a reduced N170 amplitude. This suggests that masked repetition modulated early visual processing but did not influence processes beyond approximately 200 ms that reflect the access to facial representations and semantic information for people.