Abstract
We previously reported (VSS ‘11) that people could constrain attention to two distinct colors and that both colors biased attention in parallel. Here we investigated whether this ability generalizes to selection by orientation. Participants’ eyes were tracked while they searched displays of 24 oriented bars for one with a small break on its left side. Each bar was set to be one of four equally spaced orientations (± 22.5° from vertical and horizontal). There were four cuing conditions (a single bar oriented +22.5° from horizontal, a single bar oriented -22.5° from vertical, both of these bars, or all four orientations). Participants were informed that the target, if present, would always appear in a bar of the cued orientation(s). Cuing conditions were blocked and block order was counter-balanced. Within cuing conditions there was also an effective set size manipulation such that the cued items were 6 or 12 of the items in the array. We analyzed reaction time, the number of fixations on each orientation, and the transitions between fixations. Results show that attention can be constrained based on an orientation cue, but it is not very effective. In addition, when the cue was +22.5° from horizontal, the distractors that were -22.5° from horizontal were fixated far more often than the distractors that were -22.5° from vertical (even though both types of distractors were 45° from the cue). A similar asymmetry appeared for the -22.5° from vertical cue. This asymmetry in distractor effectiveness suggests that the biasing of attention was based on a categorical description ("close to horizontal" or "close to vertical") rather than the precise cue orientation. Finally, we found no evidence that the two orientations could bias attention in parallel. These results are quite different from the color results and suggest that not all features are created equal.
Meeting abstract presented at VSS 2012