Abstract
A growing body of evidence suggests that semantic knowledge can influence the guidance of attention. For example, when observers search for a particular object, their attention is often biased toward semantically related objects in the display. In the present study, we assessed whether learned associations among objects can similarly bias attention. Participants searched for one of four targets among pairs of novel shapes. During an initial training phase, each target always co-occurred with an associated distractor in a fixed spatial configuration. During a subsequent test phase, each target appeared equally often with the associated distractor or a different distractor. The target and distractor could also be presented in the same pair or different pairs during this phase. In our first experiment, the target and distractor were presented in the same pair during the test phase. In this case, participants were faster to detect the target when the associated distractor was present. However, while participants were slower to respond on target-absent trials, they were not slower to respond when the associated distractor was present. This suggests that attention was biased toward the associated distractor, but only when it appeared at the same location as the target. In a second experiment, the target and distractor were presented in different pairs during the test phase. In this case, participants were slower to detect the target when the associated distractor was present. Participants were also slower to respond on target-absent trials, and were even slower to respond when the associated distractor was present. This suggests that attention was biased toward the associated distractor, even when it no longer predicted the location of the target. Together, these findings indicate that learned associations among objects can bias attention, and suggest that statistical learning may contribute to the effects of semantic relationships on attention.