Peter Lennie - US grants

The macaque's retina contains several distinct functional classes of ganglion cell, at least two of which project to the lateral geniculate nucleus (LGN) and terminate in different laminae. The segregation of pathways, preserved throught the LGN, becomes obscured in the striate cortex. The experiments proposed here are intended to answer the following questions: (i) which cortical neurons are driven by the different classes of input, and how does their behavior reflect the distinctive inputs; (ii) what are the visual roles of these "parallel pathways". The plan of investigation is first to clarify some properties of the distinctive classes of LGN neuron in monkey (sensitivity to spatio-chromatic contrast; sensitivity to stimulation remote from the classical receptive field; whether or not the neurons receive inputs from rods) so that the inputs to cortical neurons can be better specified. The second stage is to link cortical neurons to their different classes of inputs by the quantitative analysis of receptive field properties in cortex to reveal which classes of LGN neurons could, and which could not, provide the input to the cortical cell. Sinusoidal grating patterns whose luminance and color can be modulated in space and time will be the principal stimuli, because they are potent, they faciliate the application of Fourier techniques to characterize the underlying structure of receptive fields, and they provide results that can be compared with psychophysical measurements of human visual sensitivity. The third component of the project is to measure the "thresholds" of individual neurons (i.e., to find the stimuli that evoke weak, reliable, responses) so that the sensitivities of neurons can be compared with the sensitivities of human observers. It should then be possible to establish which neurons might, and which could not, underlie different aspects of human vusual performance. Some measurements of sensitivities of individual neurons will be undertaken in the awake, behaving, monkey to establish whether or not the properties of cells differ in important ways from those found in the anaesthetized animal.

image processing; computers; biomedical equipment purchase; television; vision;

The Center for Visual Science is a multidisciplinary center that provides advanced research facilities and specialized instruction for graduate students, postdoctoral researchers, and faculty whose work involves visual science although their disciplines are diverse. Members of the Center have their primary appointments in a variety of departments, including anatomy, brain research, electrical engineering, ophthalmology, optics, physiology, and psychology. This is an application for support of resources and facilities that would enhance the quality and quantity of research at the Center in a way that is not possible through other mechanisms of support.

The overall aim of the work proposed here is to understand how central visual pathways analyze information about the chromatic and spatial properties of objects. Three major projects are proposed. Through a combination of physiological and psychophysical experiments these attempt to elucidate the connection between a particular aspect of vision and its underlying physiology. The first project will explore the possibility that the chromatic opponency in receptive fields of P-cells of macaque LGN arises not through selection of cone classes, but through the center receiving input from a single cone, with the surround drawing indiscriminately on inputs from L and M cones. The physiological measurements will use interference fringes of appropriate color to analyze the organization of cone inputs in the receptive fields of P-cells at various eccentricities. Psychophysical experiments to estimate the density with which chromatically opponent mechanisms sample the image will explore the corollary that a paucity of midget P-cells in the periphery accounts for our impaired color vision. The second project will use silent-substitution methods to examine how gain-controlling mechanisms in individual cone pathways act together with noise to control chromatic sensitivity in P-cells. Work undertaken during the previous period of support revealed paradoxical changes in the gain of P-cell responses following changes in background chromaticity, and showed that changes in noise interact with these in regulating detectability of stimuli. Because changes in background chromaticity also led to long- lasting changes in maintained discharge of P-cells, the project will also examine how these changes are expressed by cortical neurons. The third project will explore the idea that the behavior of neurons in the early stages of visual cortex accounts for the perceptual distinctiveness of visual textures. The experiments will examine how neurons (particularly complex cells) in V1 and V2 respond to visual textures of types employed in psychophysical studies. Work undertaken during the previous period of support suggested that a fuller understanding of the roles of cortical cells will require the use of visual stimuli that tap the nonlinear aspects of their behavior. The study of texture discrimination should get directly at this because perceptual discrimination of textures appears to require nonlinear mechanisms with properties like those found in complex cells.

brain imaging /visualization /scanning; computer program /software; computer system design /evaluation

building /facility design /renovation; biomedical facility; animal colony;

ABSTRACT NOT PROVIDED