Attention and priming for top-bottom reflections of object images

Purpose: Using a priming paradigm, Stankiewicz, Hummel, & Cooper (1995) found evidence for a dual representation of object shape - one part of this representation requires attention to generate, but is invariant with left-right reflection; the other part does not require attention but is left-right specific. The present experiments extended this paradigm to investigate the role of attention in priming for misoriented (i.e., upside-down) objects. Do upright image images prime both upright and inverted images (and vice versa), and how does this pattern vary as a function of attention? The answer to this question has important implications for understanding the locus of priming for ignored images: If it resides at the level of basic features, then we should observe priming in all view-consistent conditions (i.e., upright-to-upright and inverted-to-inverted); by contrast, if it resides at the interface between perception and object memory, then it should obtain only in the upright-to-upright condition. Methods: Trials consisted of paired prime and probe displays. Each prime display depicted two objects, one of which was attended, and one ignored. The probe display depicted a single object that was either (a) the object attended in the preceding prime display; (b) the object ignored in the preceding prime display; or (c) a (baseline) object that did not appear in the preceding prime display. Prime and probe objects were either upright or top-bottom reflected (inverted). The dependent variable was priming (response time for primed objects relative to baseline) as a function of attention, prime orientation (upright vs. inverted) and probe orientation. Results: Preliminary results suggest that ignored upright primes facilitate the recognition of upright probes; no other facilitation was observed for ignored primes. Conclusions: This pattern suggest that priming for ignored images resides at the interface between perception and memory.