Amodal completion in passively viewed displays: A priming study

Schulz et al. (OPAM, 2001) used priming to investigate the time course of grouping by color. Participants were shown ambiguous masked prime displays for a variable prime duration. A central column of elements in each prime appeared behind a transparent surface. The central column could group with elements on the right or the left depending on whether grouping was based on a pre- or a post-constancy representation of color. Participants reported the grouping of unambiguous target displays that matched either the pre- or the post-constancy representation of the prime. At short prime durations, responses were faster when the target matched a pre-constancy interpretation of the prime. At long prime durations, responses were faster when the target matched a post-constancy interpretation. Here we extend this line of inquiry by using priming to investigate the time course of amodal completion. Primes consisted of ambiguous displays in which a central column of half disks abutted a thick, opaque vertical stripe, giving rise - with unlimited viewing - to the percept of complete disks partially occluded by a stripe. The central column of half disks was flanked on one side by two columns of complete disks, and on the other side by two columns of half disks. This prime display was masked after various durations either long enough to permit amodal completion of the central half disks or not. Following the prime, we presented unambiguous target displays that were similar to the prime displays, with the exception that the central column contained either complete disks or half disks presented without 'occluder'. Participants reported the grouping of the unambiguous target displays. A reaction time advantage for targets that matched a post-completion interpretation increased with prime duration.