The retina is 0.3 mm thick complex neurons layers and their connections. Three layers of cell bodies are separated by two layers of axons and dendrites. The retina’s first layer is the ganglion cell layer having a single axon each. In the retina, light is converted to a set of analogue photo-receptor voltages in either the low resolution rod system or the high resolution cone system dependent on ambient light level.
These voltages are modified to represent contrast values, color and spatial relationships between sets of receptors through the complex feed-forward and feed-back circuitry of the numerous bipolar, horizontal and amacrine cells. These conditioned voltages represent the processed visual image and are converted to a train of action potentials by the retinal ganglion cells and transmitted down the axons of the optic nerve with a mean frequency related to the voltage present at each location.
The higher the applied voltage at the ganglion cell, the higher the frequency at which the action potentials fire, The retina has one hundred million rods and six million cones where as the optic nerve contains only one million axons in total so some form of spatial encoding or time multiplexing is required. The rod system is known to sum a number of inputs (perhaps8–16) in a form of logical OR gate, which reduces the number of outputs (ganglion cell axons), needed considerably.
It is generally assumed that the retinal output arrays from the rods and from the different types of cones are transferred in a one-to-one spatial relationship to the lateral geniculate nucleus (LGN) located in the thalamus. The retinal ‘contrast image’ is then transferred from the LGN through the optic radiation to the visual cortex area V1 in a so called retinotopic mapping, which is then also replicated in higher visual processing centers.