Date of Award
Doctor of Philosophy (PhD)
Professor Kathryn M. Murphy
Most studies of orientation detection and discrimination have focussed on the ability to detect oriented stimuli at low contrast or to discriminate small differences in stimulus tilt. In this thesis, I present a new orientation-in-noise stimulus that is high contrast and broadband in spatial frequency and is designed to assess orientation discrimination. Thresholds were measured as the minimum amount of oriented signal required to accurately discriminate the underlying orientation. In Chapter 2, the nature of orientation discrimination was explored using two versions of this novel stimulus. Discrimination thresholds were better for a Local stimulus, where the orientation signal was spatially limited, than for a Global stimulus, where the orientation signal extented across the entire stimulus. Performance improved with increasing stimulus area and reached a limit of about 11% orientation signal. In Chapter 3, the temporal aspects of orientation discrimination were measured with the orientation-in-noise stimulus. Discrimination thresholds were better with a dynamic stimulus, containing multiple independent samples over time, than with a static stimulus presented for an equal duration. Furthermore, increasing the presentation time resulted in even better performance with the dynamic stimuli. These psychophysical results, alnog with results from a computational model, suggest that human orientation discrimination is medicated by pooling the local responses of low-level neural mechanisms over space and time, and is limited by two stages of intrinsic neural noise. The effects of an oriented context on discrimination in noise thresholds are presented in Chapter 4. Discrimination threshold increase two-fold when an oriented center is surrounded by a parallel texture. Thresholds with an orthogonal surround are the same as thresholds with no surround, and slightly better than thresholds with an unoriented noise surround. The effects of a surround on orientation discrimination are reduced when the center size is increased, a gap is increased between the center and the surround, or the surround signal strength is reduced. These results are captured by a computational model based on orientation-specific inhibition and pooling. When the center and surround components are presented either dichoptically or in depth, the effect of a parallel surround on orientation discrimination is not eliminated (Chapter 5). These results suggest that context effects are mediated by early cortical mechanisms. Together, the behavioural and computational results provide novel insights into the nature of orientation perception, and lay the foundation for future research probing the neural mechanisms of orientation discrimination.
Anderson, Nicole D., "Understanding the Mechanisms of Visual Orientation Discrimination" (2003). Open Access Dissertations and Theses. Paper 864.