Abstract
The surprise-attention hypothesis, stating that expectancy-discrepant stimuli can capture attention, is tested in two visual search experiments with accuracy as the dependent variable. The expectancy-discrepant stimulus was the unannounced presentation of a location precue in a new colour. As the precue was presented at the location of a nontarget, attentional capture was expected to register in a performance decrement in the critical trial. Experiment 1 revealed that attentional effects are absent after 100 ms but present after 400 ms. Experiment 2 showed that this effect is due partly to presentation duration. These results indicate that surprise capture has a late onset and requires a stable representation. In addition, the results indicate that traditional measures of attention capture and verbal reports of awareness can dissociate.
Notes
1Another way to view the two hypotheses is in terms of spatial versus nonspatial interference. The delayed-onset account assumes spatial costs (i.e., costs that are proportional to the distance between the target and the singleton), whereas the interference account assumes nonspatial costs, which depend on the mere presence of the singleton and not on its position.