Learned attention proactively modifies sensitivity to visual features by enhancing targets and suppressing distractors

Selective attention is shaped by statistical properties of the environment. For example, visual search is faster when targets or distractors appear more often in a particular color relative to less frequent colors. Here we tested whether these learned selection effects operate by proactively modifying perceptual sensitivity to specific feature values. Participants (N=200) performed a visual search task in which either targets (Exp. 1) or nontargets (Exp. 2) appeared more often in a particular color (the ‘rich’ color, 75% of trials). On a rare subset of trials, a single stimulus was briefly flashed at a random location to probe participants’ perceptual sensitivity to the different colors. The probe appeared either in a neutral gray or in one of the search colors. Across both experiments, participants searched faster when targets (Exp. 1) or nontargets (Exp. 2) appeared in the high-probability colors, indicating that they learned to select and ignore specific feature values. Critically, on probe trials, participants in Exp. 1 identified probes that matched the target-rich colors more accurately than other probes, but participants in Exp. 2 detected probes that matched the nontarget-rich color less accurately than other probes. These results demonstrate that experience-driven attention can proactively affect feature sensitivity, enhancing expected relevant features and suppressing expected irrelevant ones. This is in sharp contrast to goal-driven attention, where explicitly cued nontarget features are not proactively suppressed (Addleman & Störmer, 2022), suggesting that learning may be a particularly effective route to feature-based ignoring.