Object attention in extended objects has few effects on accuracy

Abstract

Object attention theory claims that attention is allocated to objects, or spatial regions delimited by object boundaries. One influential object attention paradigm measures the effects of cuing one region of an extended object demarked by an outline on the reaction times to onset detection at the cued location, another location within the object, and a distance-matched control location in separate object (Egly, Driver, Rafal, 1994). However, reaction times in such paradigms may be affected by criterion or bias shifts as well as changes in discriminability. This study measures full psychometric functions for orientation discrimination in a cuing paradigm to examine object attention effects on location cuing in zero and high external noise. Two vertical (or horizontal) rectangular objects marked by outlined edges encompassed two locations each (i.e., top and bottom). One end of one object was peripherally pre-cued, and observers reported the orientation of a Gabor patch in one of the four locations indicated by a report cue; the precued location was tested on 5/8 of the trials, and each of the others on 1/8 of trials. Gabors of varied contrasts appeared either in high external noise (masked) or in no noise. We also tested a four-location design without the rectangles. There was a large benefit for the precued location in the object attention experiment, similar to that found in the four-location design (see Dosher & Lu, 2000; Lu & Dosher, 2000). The within object uncued position showed very small accuracy benefits relative to the between object control in zero noise and no reliable benefit in high external noise, although one of four subjects showed an effect. Any effects of object attention on accuracy are small in relation to spatial cuing effects. Most reported object attention experiments use simple response times, where voluntary changes in criterion may be important.

Footnotes

 National Eye Institute, & UC, Irvine.