Figure 1

A. Simple two-pathway retinal circuit model. A stimulus (s) is presented and transmitted to separate ON and OFF pathways, which receive correlated corrupting noises η₊ and η_-, respectively. The signals are passed through encoding nonlinearities to produce inputs r₊ = f₊(s+ η₊) + ζ₊ and r_- = f_-(-s- η _- ) + ζ_- to retinal ganglion cells; these responses have been further corrupted by correlated noises ζ₊ and ζ_-. We calculate the optimal shape of the nonlinearities f₊(z) and f _- (z) as functions of the noise and stimulus distribution parameters. B. The optimal encoding nonlinearities for low pre- and post-nonlinearity noise variance. C. Large noise variances. D. Very large post-nonlinearity noise.