Abstract

Remington and Folk(2001) reported a dissociation between attention and selection by showing that not every dimension of attended stimulus was selected. In their experiments, the interference from attended foils (nontarget items) occurred in the task-relevant dimension but not in the task-irrelevant dimension when the response to a target conflicted with the responses to each foil dimension. These results suggested that only the task-relevant dimension, but not the task-irrelevant dimension, of foils was selected when the foils were attended. The current study investigated whether the selection of the dimensions depended on the allocation of spatial attention. To manipulate the degree to which spatial attention is allocated, we used different degrees of perceptual load and cue predictability. When the perceptual load of task stimuli was low, the task-irrelevant dimensions of attended foils were not selected as in the Remington and Folk's experiments. When the perceptual load was high, however, either the task-relevant or task-irrelevant dimensions of the attended foils were not selected. Cue predictability also affected the dissociation. When the high informative cues appeared at the foil locations, only the task-relevant dimension was selected. When the high informative cues appeared at the locations of the target or the neutral, however, neither the task-relevant nor the task-irrelevant dimension of foil was selected. These results suggest that the selection of dimensions of foil can depend on the top-down mechanism and/or the allocation of spatial attention.