Our main strategy for developing theories of visual processing has been to work out, in quantitative detail, the constraints on visual performance imposed by anatomical and physiological mechanisms, and to carry out psychophysical experiments to determine what aspects of human performance might be accounted for by these mechanisms. The present proposal is to continue applying this research strategy to the study of spatial, contrast, shape and position discrimination. Three major projects are proposed. The first major project will be a quantitative analysis of the mean response and noise characteristics of a large population of neurons in primary visual cortex of cat and monkey. One of the main hypotheses that will be tested is that the variance of a cortical neuron's response is proportional to its mean response independent of the stimulus conditions that produce that mean response. We also use signal-detection methods to compute the discrimination and identification performance of cortical neurons along a wide range of stimulus dimensions including contrast, spatial frequency, temporal frequency, position/phase, and orientation. The second major project will involve the psychophysical investigation of three aspects of pattern detection and discrimination that have received relatively little attention in the past, but are crucial for developing general accounts of discrimination performance. One set of experiments will be directed at investigating the information pooling rules applied by the visual system in pattern detection and discrimination. Specifically, we propose several experiments to measure cycle summation rules for compound grating patterns. A second set of experiments will be directed at extending our efforts to understand how light/dark adaptation mechanisms contribute to pattern detection performance. Flashed-background increment-threshold functions will be measured for sinewave grating targets during early and long-term dark adaptation and during early light adaptation. A third set of experiments will be psychophysical tests of the hypothesis that the response variance of cortical neurons is proportional to the mean response (which seems to hold in primary visual cortex). The third major project will be directed at quantitative modeling of pattern detection and discrimination performance. The models we plan to investigate are unique in that they explicitly include (a) optical/receptor factors, (b) retinal light/dark adaptation factors, and (c) many of physiological properties observed in neurons in primary visual cortex.
Showing the most recent 10 out of 33 publications
