Description. The signal-to-noise ratio (SNR) of ganglion cells determines the quality of retinal output and limits visual performance. Currently, there are no quantitative accounts of how retinal circuits minimize noise. The long term goals of the applicant have been to relate circuit structure to function in the cat retina using anatomical, electrophysiological, and computational approaches. The applicant proposes to measure the SNR of the generator potential of alpha and beta ganglion cells. He will determine how much noise arises from synapses upon these ganglion cells, compared to noise from ionic channels in ganglion cell membrane, and noise introduced by the circuitry between the cones and the ganglion cells. The applicant then proposes to use quantitative structural information in conjunction with physiological data to model this circuitry. This should provide information on how these circuits improve the SNR of ganglion cells.