Search for a command to run...
Abstract Photoreceptor degeneration is among the leading causes of blindness and optogenetics as a potential therapeutic measure has garnered much attention over recent years. In this approach, light-sensitive molecules like Channelrhodopsin-2 (ChR2) are inserted into neurons in the retina to play the role of light-sensing elements after the loss of photoreceptors. Previous studies have shown that retinal ganglion cells (RGCs) in blind animal models with optogenetically modified retinas can respond reliably to steps in light intensity or similar diagnostic stimuli. Yet, little is known about how responses to natural stimuli in optogenetically treated retinas compare to normal, photoreceptor-mediated responses and how any differences might be counteracted by adjusting the stimulation. In this work, using mice of both sexes with intact photoreceptors as well as ChR2 expression in RGCs, we directly compared the encoding of natural images by individual RGCs under photoreceptor and optogenetic stimulation. We observed that evoked firing rates under optogenetic stimulation, relative to photoreceptor stimulation, often display reduced thresholding effects and a more linear dependence on receptive-field activation as well as reduced sensitivity to local spatial contrast and reduced dynamic range. Based on these differences, we devised modifications of the natural images, including thresholding and scaling of pixel intensities together with spatial low-pass filtering, and found that using such modified images under optogenetic stimulation can lead to stronger responses that are also more similar to the original photoreceptor-evoked responses. These findings may help optimize stimulation of optogenetically modified retinas to achieve more natural vision in future therapeutic applications. Significance Statement Degenerative diseases that lead to the loss of photoreceptors, the eye’s light sensors in the retina, are a major cause of blindness. One promising therapy approach uses optogenetics to place light-sensitive proteins into retinal neurons, allowing them to detect light in the photoreceptors’ stead. We compared how the retina responds to natural images when stimulated via the inserted light-sensitive proteins versus normal activation and observed systematic differences in how images are represented, owing to reductions in response range, signal thresholding, and contrast sensitivity. Yet, by modifying the presented images, including spatial blurring as well as intensity thresholding and scaling, we managed to restore more natural image responses. These results suggest ways to improve visual quality from optogenetic treatments of blindness.