Abstract
Bottom-up input and top-down expectation of objects are usually intertwined to contribute to the occurrence of object-based attention. The current study demonstrates the dissociation of influences of bottom-up object input and top-down expectation on the same-object effect by using the double outlined rectangles of Egly, Driver, and Rafal (1994). One of the two rectangles was first cued, and the participants were asked to detect a target shown on the target frame. The same-object effect refers to the result of faster responses when the cue and target appears on the same object than on different objects. We manipulated the presence probability of (1) rectangles and an empty space, or (2) rectangles and a new object in the final target frame. The new object is an outlined boomerang whose orientation varies with the cue location in a way to either reduce or enhance same-object effects when the double rectangles are in the target frame. When the double rectangles are absent in the target frame, no same-object effects are observed, regardless of the probability of rectangles presence. The dependence of same-object effects on the presence of the attended object indicates the effect of bottom-up object input. However, the ever occurrence of the boomerang, even with a probability of appearance as low as 20%, influences the same-object effects when in fact it is the rectangles that appear in the target frame. That is, the expectation of the presence of the boomerang has influence on the same-object effects while the rectangles are present. Although bottom-up input usually interacts with top-down expectation, their dissocation revealed in this study sheds light on their respective roles in object-based attention.
Supported by the National Science Council of Taiwan, NSC95-2752-H-002-007-PAE and 95-2413-H-002-020