Decoding of visual stimuli from extracellular responses of a retinal ganglion cells population

The retina is a tissue placed on the posterior surface of the eye that encodes the visual information from the environment. This code is represented by action potentials or spikes transmitted along the optic nerve to the brain, whose characteristics depend on the visual stimulus that come to the retina. Currently, it is not still completely clear which coding strategies are employed by the retina to encode the visual information, although some of them have been assumed in the recent years. The present work evaluates some coding strategies to reconstruct visual stimuli corresponding to sequences of four light intensities randomly presented. This task was performed with combinations of three coding strategies: latency, first inter-spike interval and firing rate. The classification was based on linear discriminate analysis, with the aim to determine which combinations of coding strategies detect the information of the light intensities. The performance of the method depended on the coding strategies combinations and the specific light intensity to reconstruct.