Visual, Spatial and Temporal Filtering and Hyperacuity

点击翻译按钮获取中文摘要

Visual acuity is a measure of how well the eye can resolve very tiny targets, and is limited by the size and distribution of the photoreceoptor cells in the eye. However, a slight sideways displacement break in an otherwise straight line can be detected when the displacement is as much as five times less than the normal acuity would predict. This discriminability is known as hyperacuity, and can be demonstrated by this stimulus called vernier displacement, or by displacement of a grating pattern or by bisection of a pattern. Understanding hyperacuity is important for understanding visual information processing because hyperacuity reveals the ultimate precision of visual localization in space. This project will study how spatial and temporal properties of information channels in the eye and brain determine the limits of hyperacuity. Single neurons (nerve cells) in the visual system have particular response properties that preferentially filter particular frequencies in time (cycles/sec) or space (cycles/degree, of a grating, for example). Hyperacuity targets or vernier and grating displacement will be presented to human observers, and then these targets will be "masked" by the presence of flickering illumination, or superimposed gratings, or spatiotemporal noise. The discriminations will be compared for these different cases. To determine the dynamic characteristics of the neural mechanisms involved will require measurement of displacement thresholds as a function of luminance contrast, color contrast, mean luminance, and temporal frequency. Finally, a comparison will be attempted between human hyperacuity performance and the hyperacute capabilities of single neurons in the macaque monkey's visual pathway. The goal of this work is a comprehensive understanding of the neural basis for hyperacuity. The approach combining psychophysics, physiology and mathematical modelling is rare and powerful. Understanding this wonderful performance in terms of neural mechanisms will lead to a better theoretical basis for understanding the ultimate limits of visual performance.