Can people suppress salient visual distractors without foreknowledge of their colors?

Several lines of evidence suggest that observers can suppress salient visual distractors to prevent them from capturing attention. Currently, it is unclear whether such proactive suppression is possible when the defining feature of the distractor varies unpredictably across trials. Using probe-recall rates and oculomotor data, Gaspelin and Luck (2018, JEPHPP) showed that suppression is not possible when target and distractor features swap unpredictably across trials. These results indicate that suppression may be tied to the visual feature that defines the distractor (feature suppression). However, we previously reported that salient distractors elicit an event-related potential (ERP) component associated with suppression (the distractor positivity, PD) even when a salient distractor is randomly intermixed with a less-salient distractor that requires no suppression (Gaspar et al., 2016, PNAS). Here, we tested the feature-suppression hypothesis more directly by varying the color of the distractor while maintaining its high salience. In one experiment, participants viewed displays containing eight or nine green circles, a green diamond (shape-singleton target), and, on distractor-present trials (50%), a nongreen circle (color-singleton distractor). Critically, the distractor color was varied randomly across trials (magenta, red, orange, blue, and cyan) to prevent feature-based suppression. Both target and distractor elicited contralateral positivities over the posterior scalp in the time range of the P1/N1 components (Ppc; 100—200 ms). The target elicited a subsequent N2pc (index of attentional selection), while the distractor elicited a subsequent PD. These findings indicate that salience-based suppression can occur without foreknowledge of the distractor’s color and are thus inconsistent with the feature-suppression hypothesis. The Ppc results are also inconsistent with suppressive interpretations of the pre-N2pc positivity (i.e., it does not appear to be an early PD).