Psychophysical analysis of three separate anisotropic patterns in vision : putting visual reference frames in conflict to study the class 1 and 2 oblique effects and the horizontal effect.

Abstract

This dissertation describes and investigates a debate about differences in the visual system’s anisotropic sensitivity to orientation as well as where and how the brain encodes visual orientation information. Three different visual perceptual patterns are discussed: the Class 1 and 2 oblique effects and the horizontal effect. The oblique effect is the ability to perceive cardinal orientations more easily than oblique orientations, and is found using narrowband stimuli. In the horizontal effect, oblique orientations are perceived most easily, horizontal orientations are perceived worst, and vertical orientations fall in-between, and is found using broadband stimuli. While the majority of authors refer to the oblique effect as a whole, there is a case that it can be divided into Class 1 and Class 2 depending on what the task is measuring: low-level properties of the visual system or higher-level categorization and memory tasks. Moreover, when determining which reference frame the visual system uses to define the oblique effect, there is a disagreement as to whether it is the retinal or gravitational reference frame. To add evidence that there are two classes of the oblique effect and that they are defined by different reference frames (hence the disagreement), retinal, gravitational, and patterncentric coordinates were put in conflict with one another. The effects of phenomenological coordinates were also tested to determine whether top-down cognition played a role in either the Class 1 or Class 2 effect. That the Class 1 oblique effect remained defined by retinal coordinates no matter the test. The Class 2 oblique effect was defined by gravitational coordinates during head tilt with no available patterncentric information. However, when patterncentric coordinates were available the Class 2 effect pattern was defined by these coordinates. Phenomenological coordinates seemed to make no difference to either effect. These results add to current evidence that the oblique effect should be separated into two classes: Class 1, which shows the low-level orientation characteristics of the visual system and is defined by retinal coordinates, and Class 2, which is defined by patterncentric coordinates, and gravitational coordinates when there is no patterncentric input. There have been no studies to date to address the coordinates used by the visual system to define the horizontal effect. The horizontal effect is thought to be important for scene information, so the hypothesis was that it would follow patterncentric coordinates. Again, I tested retinal, gravitational, patterncentric, and phenomenological reference frames. During head tilt (no patterncentric information available), the standard horizontal effect did not remain tied to gravitational coordinates, nor did it transfer completely to retinal coordinates. The pattern did change completely with patterncentric coordinates during scene tilt, indicating that the horizontal effect is important for scene information. However, during scene tilt alone, retinal and…