Abstract
Spatial selection allows for preferential processing of certain visual information. Yet, the demands for spatial selection often differ between concurrent tasks. Here, we investigated how spatial selection in a visual search task affected processing of background scenes. We examined two types of spatial cues: explicit goal-driven cues and implicitly learned probability cues. In search, participants searched for a T among Hs; underneath the search array were four scenes, one per quadrant. Memory for the scenes was periodically tested. In two experiments (total N=48), endogenous cues indicated the quadrant most likely to contain the target T, engaging goal-driven attention. In two other experiments (total N=48), location probability learning was used to induce spatial attention: the target occurred more often in one visual quadrant than others. Both manipulations successfully induced an attentional bias toward the cued quadrant in the search task. To examine whether this bias spread to the secondary scene task, we compared memory for scenes placed in the quadrant cued for visual search versus scenes in the other quadrants. When spatial attention was endogenously cued, participants remembered scenes in the cued search quadrant significantly better than scenes in other quadrants. When using location probability cuing, participants could be aware or unaware of the target location probability. We separated participants into aware and unaware groups based on self-report. When participants were aware of the search target's location probability (N=22), the memory data were comparable with endogenous cuing data. In contrast, unaware participants (N=26) did not show transfer of the search bias to the memory task. These data suggest that goal-driven attention and incidentally learned attention engage different mechanisms. Goal-driven, but not implicit, attention transfers between concurrent tasks.
Meeting abstract presented at VSS 2017