Abstract
In this study, we examined the extent of inter-trial learning in the context of the distractor previewing effect. The distractor previewing effect is an inter-trial effect observed during odd-ball target-discrimination searches, and refers to the finding that target discrimination performance is slower if, on an immediately preceding trial, no target was present and all items shared a defining feature with the current target, relative to when all distractors shared a defining feature with the current distractors. Participants were presented with runs of four consecutive trial pairs in which target and distractor features were fixed. Runs consisted of four consecutive target-color previewed (TP) or distractor-color previewed (DP) trial pairs. Runs could contain a repetition of the color segmentation from a preceding run (e.g. two consecutive runs in which target-present displays consisted of green targets among red distractors) or a switch of color segmentation. Our goal was to evaluate learning observed within TP or DP runs, as well as the extent to which learning generalized across runs (e.g. two consecutive DP runs vs. one DP run followed by a TP run). As participants viewed a run of trials, RTs decreased both for DP and TP runs, and in both cases RTs reached an asymptote after three trial pairs. On the subsequent run, RTs increased dramatically if the target-distractor color segmentation was switched, but this decrement in performance was not modulated by the repetition of the preview condition. In sum, our results suggest that inhibitory biases against specific features are strengthened over time after repeated experience with the same feature assignments (green as distractor color). In contrast, participants seemed unable to learn repeated feature-based inter-trial contingencies (e.g. within the current run of trials, if green is a distractor on the current trial, red will be the target color on the following trial).