Abstract
Visibility of a target in a backward masking experiment depends on the SOA between the target and mask in one of two ways: either the target becomes more visible as the SOA increases (Type A masking) or the target is least visible at some intermediate SOA value (Type B masking). One explanation of the difference between these types of masking has been that type A masking occurs when the strength of the mask is greater than the target (e.g., intensity or duration). An alternative explanation has been that some types of targets and/or masks inherently introduce type A or B masking effects. We ran a backward masking experiment that rejects both of these explanations. The mask (51 ms) was either an outline square or a sparse mask of four dots arranged at the corners of the square. The target (16 ms) was either an outline rectangle (slightly taller than wide) or five dots (arranged like on a die) a similar shape. Three distracting stimuli (targets shaped as squares rather than rectangles) were also presented (all four possible targets were spatially arranged on the corners of a large square). The observer's task was to identify the location of the rectangle target. We found type A masking if the target and mask were different shapes (one made with dots and the other an outline square). We found type B masking if the target and mask were of similar types (both outline squares or dots). The findings are inconsistent with the strength-based explanation of type A and B masking because type B masking was sometimes stronger than type A masking. The findings are inconsistent with the idea that some targets/masks inherently produce type A or B masking because both targets and both masks could produce either type A or B masking. We suggest that the results are related to the visual percepts produced when the target and mask integrate at short SOAs. Depending on the spatial properties of the target and mask, the integration can help or hinder the search for the target.
Supported by the National Science Foundation under Grant No. 0108905.