The effects of sequential attention shifts within visual working memory

Abstract

Previous studies have shown conflicting data as to whether it is possible to sequentially shift spatial attention among visual working memory (VWM) representations. The present study investigated this issue by asynchronously presenting attentional cues during the retention interval of a change detection task. In particular, we focused on two types of sequential attention shifts: (1) orienting attention to one location, and then withdrawing attention from it, and (2) switching the focus of attention from one location to another. In Experiment 1, a withdrawal cue was presented after a spatial retro-cue to measure the effect of withdrawing attention. The withdrawal cue significantly reduced the cost of invalid spatial cues, but surprisingly, did not attenuate the benefit of valid spatial cues. This indicates that the withdrawal cue only triggered the activation of facilitative components but not inhibitory components of attention. In Experiment 2, two spatial retro-cues were presented successively to examine the effect of switching the focus of attention. We observed equivalent benefits of the first and second spatial cues, suggesting that participants were able to reorient attention from one location to another within VWM, and the reallocation of attention did not attenuate memory at the first-cued location. In Experiment 3, we found that reducing the validity of the preceding spatial cue did lead to a significant reduction in its benefit. However, performance was still better at first-cued locations than at uncued and neutral locations, indicating that the first cue benefit might have been preserved both partially under automatic control and partially under voluntary control. Our findings revealed new properties of dynamic attentional control in VWM maintenance.

Keywords: attention; retro-cue; selective maintenance; sequential cueing; visual working memory.

FIGURE 1

Example trials for the four cue type conditions in Experiment 1. (A) The “single shift + early probe” (SE) condition. (B) The “single shift + late probe” (SL) condition. (C) The withdrawal condition. (D) The neutral condition.

FIGURE 2

Schematic picture of the stimulus sequence between memory array offset and probe onset for each cue type condition in Experiment 1.

FIGURE 3

Mean d′ for each cue type condition in Experiment 1. Error bars represent standard errors.

FIGURE 4

Trial sequence for the “double spatial cue” (DS) condition in Experiment 2 and 3.

FIGURE 5

Mean d′ for each cue type condition in Experiment 2. Error bars represent standard errors.

FIGURE 6

Mean d′ for each cue type condition in Experiment 3. Error bars represent standard errors.