Abstract
Extensive research has revealed wide-ranging human abilities to impose attentional sets, which selectively prioritize information in the visual scene based on simple featural information (e.g., color or shape). However, an understanding of when and how attentional sets are implemented remains elusive. While past work has often assumed that attentional set is determined by explicit experimenter instruction or by the observers' intentional strategies, ecologically speaking, it is also likely that attentional set is guided by cognitive control mechanisms that are sensitive to predictive relationships between the environment and particular tasks. Previous work has shown that implicitly learned environmental contexts can guide attention to specific locations or can facilitate object identification (Chun & Jiang, 1998, 1999). Can a learned context influence the implementation of attentional set? Subjects searched for a T, which was one of 8 colors on any given trial, among Ls, which were also each one of 8 colors. The search items were presented in one of 24 unique spatial configurations, which were repeated throughout the course of the experiment. On predictive trials, the target color was perfectly correlated with the repeated configuration; the target location randomly changed within the configuration from repetition to repetition. On non-predictive trials, the target color was chosen randomly with each repetition of a given configuration. Results show that, after 5–10 exposures to a predictive configuration, subjects detected the targets in these configurations more rapidly than targets in non-predictive configurations, although subjects were never explicitly aware of the relationships between spatial layout and target color. These results show that implicit control mechanisms can trigger feature-based attentional sets in a manner that capitalizes on learned regularities in the visual world.
Supported by NIH Grant R01 EY0 14193 to MMC