My broad research interests focus on the neural mechanisms underlying visual performance. As an undergraduate, I helped develop a model for computing motion based on the known properties of cortical cells in MT. I also gained an interest in modeling biological mechanisms using analog VLSI. In specific, I was interested in improving performance of silicon models of the retina. However, I decided that I first needed to learn more about how the anatomy and physiology of the retina.
My thesis work focuses on the processing of visual signals in the primate retina. Specifically, I am developing a computational model to investigate how anatomical structures in the retina affect the spatial processing of luminance and spectral information. To evaluate this processing, I extend the ideal observer paradigm to incorporate the known anatomy of the retina. This includes coupling between foveal cones, horizontal cell feedback at the cone terminals, and the various types of ganglion cell arrays proposed to carry chromatic information. By using the ideal observer to perform psychophysical discriminations at different stages, various theories about retinal function can be tested. This includes whether spatial arrangement of cone types is important for spatial and color vision, the relative advantage of establishing a cone receptive field, and which ganglion cell type might code chromatic information. Furthermore, an ideal observer enables the comparison of retinal performance with the known human psychophysics, providing insight into the contribution of each retinal stage towards overall visual performance.