Abstract
Recent work to be reviewed indicates that a brief presentation of a prime scene (as little as 200 ms) activates a functional representation of its layout. The representation is functional because it facilitates (speeds) subsequent processing of depth relations within the scene. Evidence indicates that the representations are somewhat abstract and are activated in parallel across the visual field, with relatively little capacity limitation. The representations appear to be intermediate-level, and separate from later stages of processing in which spatial relations are resolved.
The present experiments show that the positive effects of scene primes can be nullified by irrelevant objects. The general method was to present a scene prime for 200 to 800 ms, followed by a target with two laterally separated spatial probes marking locations. The prime scene was either the same scene as the target (sans markers), or the same scene with two irrelevant objects added. The objects were a task-irrelevant color, and were the same orientation as or inverted relative to the larger upright or inverted prime. In control conditions, there was the same amount of preparation time but a fixation stimulus replaced the scene prime. Reaction time to indicate which spatial marker (left or right) is closer to camera point was the main measure (accuracy was uniformly high).
Consistent with previous results, the spatial responses were faster with scene primes relative to control, with scene primes becoming fully functional after only 200 ms of presentation. When irrelevant objects were added to the prime, the facilitory effect was reduced. When the irrelevant objects were opposite in orientation to the larger scene, the effect was reduced further, to levels of performance equal to when only general scene-background information is presented. The negative effects of misoriented objects imply they interfered with the activation of an appropriate reference frame.