Abstract
Previous research has shown that surround masking of a sinusoidal grating can decrease performance on a discrimination task. It has also been shown that the similarity of the mask to the target influences the degree to which performance is affected by the mask. One hypothesis states that perceptual segregation caused by differences in center and surround allows for more independent processing of the two components, thus reducing masking. In order to test this hypothesis, observers performed a standard texture segregation task (determining whether a texture rectangle is vertical or horizontal), using relative differences in phase, orientation, spatial frequency, contrast, and mean luminance between the rectangle and its surround. The results of this task provided psychometric functions of texture segregation for each dimension. These data were then used to create several levels of equally-salient illusory contours (across different dimensions) between a target center and a masking surround. If texture segregation is indeed the key for reducing center-surround masking, orientation discrimination performance should be equivalent for similar levels of contour salience, no matter what dimension is being varied. We found strong positive relationships between texture segregation performance and orientation discrimination performance. The relationships were virtually identical across the different dimensions.