Abstract
We examined whether segmentation can alleviate the effect of a superimposed mask. We measured contrast increment thresholds for a horizontal static test grating (2 cpd, 30% contrast) under three conditions: 1) the test grating presented alone; 2) with a similar-sized superimposed mask grating of orthogonal orientation, creating a plaid; 3) with a superimposed mask having twice the diameter of the test. We hypothesized that this large mask condition might facilitate segmentation of the test. Masks had fixed contrast of 30%. The stimuli were presented at fixation and observers were asked to detect an increment of variable contrast on the test grating in a 2AFC task.
For two out of four observers, thresholds were elevated by a factor of two in both superimposed mask conditions relative to the grating-alone condition, consistent with cross-orientation masking. For the other two observers, however, only the smaller mask produced this elevation, while the large mask reduced thresholds to the level of the grating-alone condition. This suggests the test grating could be effectively segmented from the compound stimulus. The segmentation effect was more robust with drifting stimuli: all four observers showed no threshold elevation in the large-mask condition, while their thresholds were elevated twofold in the small-mask (plaid) condition. Peripheral presentation, in particular, appeared to enhance segmentation. By contrast, brief foveal presentations precluded segmentation, inducing the full masking effect regardless of spatial extent. Thus it appears that visual organization has a profound impact on early processes that determine contrast sensitivity.
This work was supported by funds from Smith Kettlewell Eye Research Institute