In two minds: Multiple attentional control settings in visual search

Many everyday tasks require that we search for multiple targets at once. Nevertheless, the manner in which attention allocation is governed in such tasks remains unclear. The aim of this thesis was to explore whether more than one attentional control setting can be maintained simultaneously. Firstly, I asked whether the capture of attention is contingent on attentional control settings for two target colours (Chapter 2). Participants completed spatial cueing tasks in which they searched for either of two possible target colours (e.g. red and green), varied unpredictably. Prior to the target, a spatially non-predictive colour singleton cue appeared at one of the possible target locations. In all five experiments, cues that matched either target colour captured attention and produced significant cueing effects, while irrelevantly coloured cues did not. This was the case even when the cue colour was not linearly separable from the target colours in colour space (Experiments 3 and 4), demonstrating that setting for multiple colours is not limited by linear separability. Moreover, the presence of capture by target-coloured cues did not depend on the match between the cue colour and the previous or subsequent target colour, which suggests that guidance by the two target colours remained relatively stable across the experiment. Finally, the results could not be explained by a single inhibitory set for irrelevant distractor colours (Experiment 4 and 5). I concluded that the data were most consistent with the maintenance of multiple attentional control settings for colour. In Chapter 3, these findings were extended to explore whether two different attentional control settings could be directed to separate spatial locations. Participants searched RSVP streams for multiple colour-location conjunction targets (e.g. green letters on the left and red letters on the right). Target coloured distractors that did not match either conjunction (i.e. red on the left or green on the right) nevertheless captured attention and interfered with performance. This occurred even when the targets were previewed early in the trial (Experiment 2), indicating that participants were unable to isolate feature sets to specific spatial regions. However, previewing the irrelevant conjunction distractors reduced their interference, an effect that was due to top-down rather than bottom-up factors (Experiments 3 and 4). The results suggest that it is possible to inhibit irrelevant conjunctions, thereby improving search performance for conjunction targets. Together, the findings from this thesis demonstrate that attention can be governed by multiple attentional control settings applied independently and in parallel at a featural level. By combining facilitative sets with top- down inhibition, selection of complex conjunction targets can be made more efficient