Purchase this article with an account.
M. Y. Lipin, W. R. Taylor; Maximizing Contrast Sensitivity in the Outer Retina. Invest. Ophthalmol. Vis. Sci. 2008;49(13):2430. doi: https://doi.org/.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Barlow (1953) maintained that the retina acts as a filter rejecting redundant information and passing salient information. But what information is salient? As van Hateren (1993) has pointed out, the power spectrum of natural images is the same as that of a step function, or edge. Thus one aim was to determine those properties of the outer retina that provide its highest sensitivity to edges. A second goal was to examine the factors that determine temporal contrast sensitivity.
The outer retina was modeled as a continuous 2D- extension of the discrete 1D- model of Yagi et al. (1989). The photoreceptors excite bipolar and horizontal cells, and the horizontal cells feedback to inhibit the photoreceptors and feed-forward to inhibit the bipolar cells. The space and time constants of the photoreceptor layer and horizontal cell layer were set by their longitudinal and leakage conductances along with the membrane capacitance. The spatio-temporal impulse response of the model was determined. A light boundary and an onset of full-field illumination were used as spatial and temporal contrast, respectively.
The maximum response to spatial contrast occurred when the ratio of the space-constant of the horizontal cell layer to that of photoreceptor layer was equal to twice the square root of the feedback loop gain. The maximum response to temporal contrast occurred when the ratio of the time-constant of horizontal cell layer to that of the photoreceptor layer was equal to quadruple the feedback loop gain. Responses to contrast of bipolar cells at high feedback loop-gain were at most twice as large as responses in the absence of negative feedback. A negative feedback gain from horizontal cells to photoreceptors of one half, equalized the dynamic range of the photoreceptors and horizontal cells.
At fixed noise levels, the parameters for the model that maximize the response amplitude will also maximize the signal-to-noise ratio. The analytical model predicts a low negative feedback gain of ~0.5 from horizontal cells to photoreceptors. Examination of previous data indicates that real systems may conform to the expectations of the model.
This PDF is available to Subscribers Only