Capture by colour: Evidence for dimension-specific singleton capture

Abstract

Previous work on attentional capture has shown the attentional system to be quite flexible in the stimulus properties it can be set to respond to. Several different attentional Bmodes^have been identified. Feature search mode allows attention to be set for specific features of a target (e.g., red). Singleton detection mode sets attention to respond to any discrepant item (Bsingleton^) in the display. Relational search sets attention for the relative properties of the target in relation to the distractors (e.g., redder, larger). Recently, a new attentional mode was proposed that sets attention to respond to any singleton within a particular feature dimension (e.g., colour; Folk & Anderson, 2010). We tested this proposal against the predictions of previously established attentional modes. In a spatial cueing paradigm, participants searched for a colour target that was randomly either red or green. The nature of the attentional control setting was probed by presenting an irrelevant singleton cue prior to the target display and assessing whether it attracted attention. In all experiments, the cues were red, green, blue, or a white stimulus rapidly rotated (motion cue). The results of three experiments support the existence of a Bcolour singleton set,^finding that all colour cues captured attention strongly, while motion cues captured attention only weakly or not at all. Notably, we also found that capture by motion cues in search for colour targets was moderated by their frequency; rare motion cues captured attention (weakly), while frequent motion cues did not.

Top-down knowledge modulates onset capture in a feedforward manner

“…2A). Still, observers did not adopt a set for all salient items (singleton search mode; Bacon & Egeth, 1994), which would have resulted in equally high selection rates for all differently coloured distractors (Folk & Anderson, 2010;Harris et al, 2015). Instead, attention was apparently simultaneously tuned to the target colour (grey) and the illusory colour (blue).…”

“…If observers bias attention to all singletons in the red premask condition (or all colour singletons; Harris et al, 2015), then all onset distractors (or all differently coloured onset distractors) should attract the gaze more strongly, compared with the grey premask condition. On the other hand, if observers bias attention narrowly to the two possible target colours, then only the two target-similar (grey and blue) distractors should strongly attract the gaze in the red premask condition, and selection rates of the target-dissimilar distractors (green and premask coloured) should not differ across the red and grey premask condition.…”

“…Presenting a grey stimulus in the same position as a coloured premask stimulus can produce a colour after-effect, so that a grey target stimulus initially appears to be greenish blue before it fades to grey (due to chromatic adaptation of the Bred^cones; see Lucasson, 1993 andTheeuwes, 2004, footnote 1). It is possible that this subjective impression of a transient colour change impaired top-down tuning to the target, either because observers tuned attention to multiple different colours or any salient colour difference (e.g., singleton search mode; Bacon & Egeth, 1994;Harris, Becker & Remington, 2015), or by impairing filtering of taskirrelevant distractors (Folk & Remington, 1998). Of note, in some studies, the target-dissimilar onset distractor could also be blue or green (Born et al, 2011), and thus was potentially similar to the perceived target.…”

Uncertainty as a determinant of attentional control settings

“…Thus, when the target, which has a distinct feature, is presented with other items having a homogenous single feature, a singleton search mode is adopted to minimize effort with a low task performance cost because it requires less effort than feature search modes. For example, when no color singleton distractor was presented, all color cues were found to capture attention, indicating a singleton search mode (e.g., Bacon & Egeth, 1994;Harris, Becker, & Remington, 2015).…”

“…For example, in Dombrowe et al's experiment, in which participants were to search for two targets presented in a target display simultaneously, target search accuracy was lower when the two targets were defined by different colors than when they were defined by a single color. This impairment is possibly due to the difficulty in maintaining multiple attentional control settings in working memory simultaneously (e.g., Biderman et al, 2017;Harris et al, 2015) or less computational efficiency when multiple attentional control settings are established than when a single attentional setting adopted (Harris et al, 2015). Thus, if the attentional control setting specifies values to ignore and other values to select when the target is defined with multiple colors and the distractor feature is certain, it is possible to maximize search performance with minimal effort.…”