Visual search selectively enhances recognition of the search target

In Experiment 1, subjects performed a probe recognition task embedded within a task in which they searched for a single target letter (either a T or an L) in a field of distractor letters (Ls or Ts, respectively). Pairs of targets were more often correctly recognized as identical than were pairs of distractors. Also, pairs of items consisting of a distractor and a hybrid composed of a T and an L were correctly recognized as different more often than were pairs of items consisting of a target and a hybrid. This result would be expected if subjects were more likely to interpret a hybrid as the target. In Experiment 2, subjects performed the same search task used previously, but now the probe task involved recognition of either a single target, a single distractor, or a single hybrid. A probe L was recognized better when it served as the target than when it served as the distractor in the concurrent search task. The hybrid was more likely to be incorrectly identified as the letter that served as the target. These results are consistent with models which assume that search instructions activate in a top-down manner the central representation of the search target, even in the absence of target stimuli. This interpretation is supported by the finding that neurons that code objects in higher-order cortical visual areas show enhanced activity when monkeys are preparing to search for the object that they code, even when it is not present.