Object-substitution masking disrupts feature processing for color and tilt

Background: Object-substitution masking (OSM) can help elucidate visual processing dynamics. In OSM, a target-and-distractor stimulus array is briefly presented. The target is cued by flankers, which onset with the stimulus array but offset after a delay. The flankers’ sustained presence disrupts target processing, “masking” it from perception (Enns and Di Lollo, 2001). However, what level(s) of processing OSM disrupts is unclear. Gellatly et al (2008) showed evidence that OSM masks color and tilt as features, dependent on target-flanker feature similarity. Bouvier and Treisman (2010), however, showed evidence that OSM disrupts integration, but not feature processing, of color and tilt. Neither study parameterized stimuli/flankers to low-level feature channels, potential loci for OSM. Aims: We investigated OSM for color and tilt, and the effect of target-flanker feature similarity, with stimuli/flankers parameterized to V1 color and tilt/spatial frequency channels. Methods: An array of six Gabors of mixed colors (“reddish” or “greenish”) and tilts (+22.5° or –22.5°) was displayed for 16-50 ms. Subjects reported the color or tilt of the target, which was cued by four flanker Gabors. In “offset-same” trials (control) the flankers offset with the stimulus array; in “offset-delayed” trials, 250 ms after. All Gabors were co-modulated in luminance and cone-opponent color (±L/[L+M]), except luminance-only “neutral color” flankers. Flankers’ color and tilt could be “same” or “opposite” with respect to the target, or “mixed,” or “neutral.” Six subjects each completed four experiments: 1) Report color; flankers mixed or neutral color/same tilt; 2) Report tilt; flankers same color/mixed or neutral tilt; 3) Report color; flankers mixed color/same or opposite tilt; 4) Report tilt; flankers mixed or opposite color/mixed tilt. Results: For each experiment, accuracy was decreased in offset-delayed trials versus offset-same trials, indicating masking. These results replicate feature OSM for color and tilt, but not the effect of target-flanker feature similarity.