Abstract
Visual change detection can be quite difficult (Rensink, 1997). In general, changes are noticed more quickly and accurately if they occur in the location of the current focus of attention, or in a location that was recently attended. To investigate the types of cues that might serve to guide attention within a scene, this experiment investigated the effects of explicit and implicit semantic priming on a subsequent change detection task. Participants first attempted to read aloud a briefly presented prime word (33 or 200 ms duration, followed by a mask of 250 or 83 ms, respectively), then looked for a difference between two alternating versions of a real-world scene. Successfully read primes (of either 200 or 33 ms duration) were coded as “Explicit”, and unidentifiable primes as “Implicit.” Primes named either the object that changed (Helpful), or named another object in the picture (Misdirected). Across all subjects, Helpful primes yielded faster change detection times than did Misdirected primes. This effect was significant for both Explicit and Implicit trials, although the effect was somewhat larger in the Explicit trials (Explicit Helpful RT = 4273 ms, Misdirected = 8769 ms, Implicit Helpful = 6920 ms, Misdirected = 10023 ms). However, performing a median split on the subjects based on the number of successfully read 33 ms primes showed that the overall effects were primarily driven by those subjects who were able to read more 33 ms primes. It is proposed that those subjects closer to the prime detection threshold were able to more fully process the implicit semantic information than those subjects who were further from threshold. Having demonstrated the influence of semantic primes on change detection, this approach can be used to characterize the manner in which attention is allocated within a complex scene and to determine how objects in our visual world are semantically encoded.