Abstract
Abstract: :
Purpose: Retinal performance is limited by noise from photons and from biologically–generated mechanisms. At low photopic backgrounds, photon fluctuation is the major source of noise in cones (Rieke & Baylor, 2000), and a mammalian ganglion cell can discriminate contrast as low as 1% (Dhingra & Smith, 2004). This threshold is thought to be set by biologically–generated noise. Using simple "ideal" and "practical" models, we asked whether synaptic noise could explain the limits to ganglion cell performance. Methods: To determine "ideal" performance at the cone outer segment we evaluated the number of photon isomerizations in the ∼700 cones illuminated by a 300 µm dia. spot over 50 ms, calculating the contrast threshold as the increment generating a response equal to the associated photon noise. To explore the "practical" performance of the ganglion cell, we constructed a compartmental model of a brisk–sustained ganglion cell and a simple approximation of its presynaptic circuit that included arrays of cones and cone bipolar cells. Cones included a realistic light response with photon noise and synaptic output. Each cone bipolar collected signals from 4–7 cones through an inverting metabotropic synapse that included temporal filtering and noise properties. The ganglion cell collected synaptic inputs from all cone bipolars (n=240) within a threshold distance (10 µm) of its dendrites. Synapses had exponential gain (e–fold per 3–6 mV) modulating Poisson release of vesicles (50–500/s). The postsynaptic properties included nonlinear ligand binding, modulating a conductance without channel gating noise. Adaptation and feedback from horizontal and amacrine cells, and biophysical properties such as the ganglion cell spike generator were omitted. To estimate collective performance for cones and bipolar cells we multiplied the single–cell S/N ratio by the square root of the number of cells converging. Results: In the "ideal" model, contrast threshold at a low photopic background (20,000 photons/µm2/s) for 1 cone was ∼7%, and for the array of 700 cones was ∼0.2%, lower by 5–fold than the contrast threshold measured in real ganglion cells. In the "practical" compartmental model, sensitivity to the same flashed spot was similar, with threshold for the collective cone array 4–16 fold lower than for the ganglion cell, depending on the vesicle release rates and synaptic gains specified. Conclusions: Photon noise at low photopic backgrounds limits contrast discrimination performance in cones but not in the ganglion cell. In a simple model of the ganglion cell receptive field center, performance of cone bipolars and the ganglion cell is limited by fluctuation in vesicle release.
Keywords: retinal connections, networks, circuitry • photoreceptors: visual performance • ganglion cells