Abstract
Purpose: When observers attend to a subset of items in a dynamic display, they can establish an attentional set that keeps some classes of unexpected stimuli out of visual awareness while allowing others in. This set appears to be based on the feature dimensions differentiating the attended subset of items from other items. These studies investigate whether similar effects can be found in the detection of unexpected objects even when the difference between the attended items and distractor items is categorical, rather than featural. Method: Observers viewed 5 trials in which 4 letters and 4 numbers moved independently on random paths and bounced off the edges of a display window. On each trial, observers counted the bounces made by either the letters or the numbers. On a critical trial, an unexpected event (UE) occurred: either an ‘E' or a ‘3’ (a mirror reversed version of the ‘E’) entered the display from the right, traveled in a linear path behind a fixation point, and exited to the left, remaining visible for 5 s. Observers were then asked to report whether anything had appeared that had not been present in the previous trials. Results: For those observers who were counting bounces made by letters, 71% reported seeing the UE if it was an ‘E’, but only 38% did if it was a ‘3’. For those observers counting the bounces made by the numbers, 60% reported seeing the UE if it was a ‘3’, but only 37% did if it was an ‘E’. Conclusion: These findings are consistent with our previous work showing that when the difference between attended items and distractors is based on a feature dimension, observers fail to detect unexpected objects that are similar on that dimension to the distractor items. However, observers do detect unexpected objects that are similar to the attended items. The current results raise the intriguing possibility that attentional sets can be defined on the basis of higher-level categories, which mediate the detection of unexpected objects on the basis of category membership.