A cell population model of retinal ganglion cell layer thickness

Abstract

Purpose: The thickness of the retinal ganglion cell (RGC) layer of the eye is readily measured using optical coherence tomography, and is observed to vary across individuals in overall size and distribution of tissue. To relate this variation to computational models of psychophysical performance, we require a means of estimating the classes of RGCs and their quantities that contribute to the observed layer thickness. Using results from prior histology studies, we have developed a model of RGC layer thickness that is based upon the volume occupied by retinal ganglion cells of different classes across the retina. Methods: We took an estimate of total RGC density as a function of eccentricity (Curcio & Allen 1990) and determined the proportion of midget (Dacey 1993) and bistratified cells (Dacey 1993); the parasol fraction was assumed to be the remainder. The density of displaced amacrine cells was also modeled (Curcio & Allen 1990). Estimates of cell body size as a function of eccentricity were obtained from human (Liu et al 2017; Dacey 1993; Perry, Oehler, Cowey 1984) and macaque studies (Dacey 1990). We assumed spherical cell bodies and Keppler’s limit (0.74) for packing density. We compared the predicted thickness of the RGC layer to empirical measures (Curcio et al 2011). Results: In histology, the maximum thickness of the RGC layer in the temporal meridian is 61 microns at 4.4° eccentricity. Our model returns a maximum thickness of 44 microns at 3.3°. Improved agreement between the predicted and observed RGC layer thickness is obtained if the assumed fraction of midget cells is reduced by ~5% (commensurately increasing the parasol fraction). Conclusion: A model of RGC cell population density accurately models layer thickness. The model is sensitive to the assumed midget fraction, and thus provides another means of specifying this elusive value.

Acknowledgement: U01EY025864