Mark W. Becker¹ (becker54@msu.edu), Morgan Dodd¹, Tristan Janisse¹; ¹Michigan State University

Research suggests that attention can be biased toward frequent target locations or features through implicit statistical learning. However, most prior work has examined only a single biasing contingency. We investigated whether individuals can simultaneously learn and implement two distinct contingencies: one based on location and another on color. To do so participants searched a clock-face array of 16 Landolt Cs for the C which had a break on the left or right, among Cs with breaks on the top and bottom. Each quadrant had four stimuli - one of each of four colors. The target appeared in the dominant quadrant 50% of the time, and in each of the other quadrants 16.67% of the time. Similarly, targets appeared in the dominant color 50% of the time, and in each of the other colors 16.67% of the time. Results indicate that both contingencies were learned implicitly and exerted additive effects on attentional allocation. Next, we examined whether shifting the bias for one feature to the volitional system via endogenous cues affected the implicit learning and implementation of statistical contingencies for the other feature. We found that moving one feature’s bias to the top-down volitional system did not block implicit learning of contingencies for the other feature. However, the impact of that learning on attentional allocation depended on the validity of the volitional cues: statistical contingencies biased attention when volitional cues were neutral or valid, but not when invalid. This pattern suggests that the implicit and volitional systems operate hierarchically, with the top-down volitional system gating the items over which implicit statistical learning can exert its influence.