Abstract
When viewing a pattern of broadband spatial content (e.g., a natural scene), visual performance for oriented content is best for oblique orientations and worse for horizontal than vertical (Essock et al., 2003; Hansen et al., 2003; Hansen and Essock 2004; 2005; 2006). When narrowband content such as a resolution-acuity target is viewed, most observers show an oblique effect (worst performance at oblique orientations) and if more content is added to the image, the anisotropy is a horizontal effect (Hansen and Essock, 2006). We've suggested that the change in the visual anisotropy results from anisotropic orientation-tuned gain-control (strongest at horizontal), whose effect is pronounced when viewing broad-band stimuli (creating more activity in units contributing to the gain control pool) and insignificant with narrowband stimuli. Such anisotropic suppression could be a direct consequence of a corresponding bias of cortical neurophysiology (e.g., Li et al., 2003) indicating more neurons tuned to horizontal than other orientations. Here we measure (1) the timing of the transition of the oblique effect to the horizontal effect as the mask-test SOA is varied, and (2) measure the orientation tuning of the threshold elevation at four test orientations (0°, 45°, 90° and 135°). When SOA between a 1/f broadband noise mask (15° orientation bandwidth) and a test (8 cpd Gabor) is varied, an oblique effect is seen at large SOAs but becomes a horizontal effect when the broadband content is present at about the same time as the test probe (SOAs around 50 msec). When orientation tuning is probed by a mask (flanking 5° broadband wedges at varied test-mask orientation differences), elevation falloff is Gaussian (40° half-height) equally at all four test orientations, but with differential strength (peak thresholds at horizontal). Together, these results further support and extend the suggestion that the horizontal effect stems from orientation-tuned gain-control.