Abstract
Saliency map models of attentional guidance predict that attention should be allocated to items in descending order of saliency (Theeuwes, 2004; Turatto et al., 2004). If the most salient item is not the target, does attention go to the second-most salient item? To examine this question, participants were tested using a visual search task involving a target and seven distractor items arranged in a circle around the periphery of the display. Stimuli were gray discs containing sinusoidal gratings that could only be identified when directly fixated. The target was a disc with a grating orientation of 45 degrees to the left or the right, and the distractors were discs with horizontal or vertical gratings. Two of the discs were either brighter or darker than the remaining discs and coincided with the target at chance level. Compared to the remaining six discs, one (first-most salient) of these two differed in luminance more than the other (second-most salient). According to saliency map-based models of visual search (Wolfe, 1994; Itti & Koch, 1999), the most salient items are attended to first, items with the second-highest salience are attended to second, and so forth. Response times to target detection were fastest when the target was in the most salient item, slower when the target was in the next-most salient item, and slowest when the target was in the remaining items. This effect held regardless of whether the salient items were brighter or darker than the other items in the display.