Abstract
Numerous studies have shown that response latencies to visual targets are slowed by the presence of an irrelevant singleton. Some have suggested that these items impair target processing by biasing pre-attentive stages of processing. However, modulations in response latency may also be driven by disruptions that occur during post-perceptual stages of processing such as response selection. Nonetheless, if an irrelevant singleton biases pre-attentive processing, it should also disrupt the accuracy of target discrimination in a data-limited task. Recently, using a go/no-go version of a traditional additional singleton task, Theeuwes and colleagues have shown that the presence of an irrelevant singleton within a visual display produces a significant reduction in accuracy, consistent with the notion that pre-attentive processing was impaired. However, the observed d' cost in this study might also be explained by a reduction in observers' response thresholds on singleton-present trials. If the presence of the irrelevant singleton increased the likelihood of a “go” response, then false alarm rates would increase even though target discriminability was unaffected. In the present experiments, we sought to provide an additional test of the claim that irrelevant singletons disrupt in pre-attentive processing. First, we replicated the typical response latency cost that has been reported in the additional singleton paradigm. We then modified this task by using brief displays ([[lt]]80 ms.) and backward masks. Observers made unspeeded responses to a target item defined by a specific attribute. Because of the nature of this task, it is unlikely that any performance would be influenced by disruptions in post-perceptual processing. Under these conditions, we find that a salient distractor item produced no cost in target discriminability, suggesting that singleton items do not disrupt pre-attentive processing, but instead influence later, post-perceptual stages.