Purchase this article with an account.
Lauren E Wool, Joanna D Crook, Orin Packer, Qasim Zaidi, Dennis M Dacey; Selective wiring is not needed for color opponency in midget ganglion cells. Invest. Ophthalmol. Vis. Sci. 2016;57(12):No Pagination Specified.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Biologists have been reticent to embrace a theory of retinal red-green color opponency predicated on random connections. The claim for selective circuitry is based on the appearance of antagonistic long (L) and medium (M)–wavelength cone signals in midget ganglion cells across the retina, despite mixed L and M inputs to both center and surround (e.g., Buzás et al., J. Neurosci. 2006). But evidence for such circuitry is either lacking (Crook et al., J. Neurosci, 2011) or limited (Field, et al., Nature 2010). We tested whether a random-connection difference-of-Gaussians receptive field model could produce L–M cone-opponent midget cells across the retina.
Hexagonal retinal cone mosaics were randomly populated from a distribution of realistic L:M cone ratios; cone density and receptive-field size changed as a function of eccentricity. Cone assignments to receptive-field centers were made by applying a Gaussian kernel to the mosaic and weighting all cones within 1σ. Antagonistic surround radii were 6× larger with 75% of the center gain. Response amplitude and phase were evaluated from the Fourier transform of the receptive field for spatial frequencies 1/128–8 cpd. Chromatic (L–M) and achromatic (L+M) tuning were evaluated for 5000 model cells and compared to data from 128 midget cells recorded in vitro at 3.25–10mm eccentricity.
Our model produces canonical “double-duty” (Gouras & Zrenner, Science 1979) midgets with chromatic responses at low spatial frequencies and achromatic responses at high. Those cells with L–M opponency decrease with distance from fovea, while cells with purely achromatic responses appear more peripherally, consistent with greater L and M cone convergence. Median eccentricity for model cells with chromatic tuning is 5.72±1.47mm, compared to 7.04±1.56mm for purely achromatic cells. Median eccentricity for 76 empirical double-duty cells and 52 purely achromatic cells is 5.48±1.46mm and 6.89±1.73mm, respectively.
A biologically realistic model of center-surround receptive-field structure that draws indiscriminately from the random cone mosaic generates peripheral midget cells that maintain red-green color opponency, despite increased L and M inputs to centers and surrounds. Hence, both the prevalence of L–M opponency across the retina and the appearance of purely achromatic midget cells in the retinal periphery can be explained without recourse to selective wiring.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.
This PDF is available to Subscribers Only