Attention capture while switching search strategies: Evidence for a breakdown in top-down attentional control

Abstract

Whereas capture experiments typically repeat a single task many times, real world cognition is characterized by frequent switching. Lien, Ruthruff, and Johnston (2010) reported that the attentional control system can rapidly and fully switch between different search settings (e.g., red to green), with no carryover and no intertrial priming. The present study examined whether this impressive flexibility is possible even when the switch is not between different features along the same dimension, but between mutually incompatible search modes. On each trial, participants were prompted to find and identify the letter that was in a specific colour (feature search mode) or was uniquely coloured (singleton search mode). Within each block, search mode was either pure or mixed; the mixed blocks contained a fixed AABB search sequence (singleton-singleton-feature-feature) in Experiment 1 and a random sequence in Experiment 2. The target display was preceded by a noninformative cue display containing a nontarget colour singleton. In pure feature search blocks, these irrelevant singleton cues were generally unable to capture attention, replicating previous findings of “contingent capture”. In mixed blocks, however, irrelevant colour singletons captured attention on feature search trials. This breakdown indicates a limitation in the sharpness of attentional control following mode switches, which might be common in the real world.

Keywords:

• Attention capture
• Visual search strategy
• Cognitive control and switching

We thank Artem Belopolsky, Charles Folk, Andy Leber, and one anonymous reviewer for comments on earlier versions of the manuscript.

We thank Artem Belopolsky, Charles Folk, Andy Leber, and one anonymous reviewer for comments on earlier versions of the manuscript.

Notes

¹ We also compared the cue validity effect for the feature-search trials in the mixed and pure blocks after eliminating the trials where the cue colour matched the target colour of the preceding singleton-search trials (i.e., “same” trials). The cue validity effect was similar between both blocks in Experiment 1 (9 ms for both blocks), t < 1.0, but was larger in the mixed than pure blocks (33 ms vs. 7 ms, respectively) in Experiment 2, t(79) = −2.54, p = .0132. Thus, the observed increased capture in the feature-search switch trials within the mixed block in Experiment 2 cannot be attributed to the fact that the cue colour in the feature-search trials sometimes matched the target colour on the preceding singleton-search trials. It is even less likely to explain the increased capture observed on feature-search repetition trials in the mixed blocks for both Experiments 1 and 2.