Object discrimination using electrical and optogenetic artificial retinal stimulation

Abstract

The visual input constitutes of over 90% of the information one takes in about the surrounding world, therefore the sense of sight is arguably one of the most important. Despite technological and medical advancements, vision impairment still impacts millions of people world-wide and new artificial vision stimulation strategies for already existing prototypes or new gene therapies for sight restoration continue to be sought for. The current thesis aims at answering some of the open questions, contributing to the present artificial vision strategies. Throughout the studies presented here we used multi-electrode arrays (MEAs) to record retinal ganglion cells’ signals in response to either electrical or light stimulation. The more robust complementary metal oxide semiconductor (CMOS) MEA devices allowed us to both electrically stimulate retinal cells and record their neural activity. With the aid of a light stimulation add-on to the experimental setup, we were able to investigate both electrical stimulation strategies, as well as optogenetic ones in the same retinal samples.With electrical stimulation of the tissue and its relatively long history in the field of visual prosthetics on one side and optogenetic therapy as a newer field on the other side, one of the aims of this work was to compare the spatial resolution of the two techniques in mice expressing a gene mutation that leads to retinal degeneration (rd10). We have shown that, under the set experimental conditions, both techniques are capable of achieving high spatial resolution in adult rd10 mice.Next, we looked at optogenetics as a potential sight restoration therapy and performed experiments on both rd10 mice as well as normal sighted mice (C57BL/6) to see whether one cell type could be a better suited target. Comparing retinal ganglion cells (RGCs) with bipolar cells (BCs), both expressing channelrhodopsin- 2, showed that RGCs might be a better target when considering spatial resolution.In a subset of the experiments, we also considered ReaChR/gtACR1 expression in AII amacrine cells (ACs) and observed a bursting behavior similar to the one recorded in the ChR2 expressing BCs mice. We attributed this to the poorer spatial resolution we estimated for BCs and AII ACs optogenetic targets, but we acknowledge that further tests need to be performed. Sections 4.2.1 and 4.2.2 explored cell-class selective stimulation in either a mouse model of early retinal degeneration or a normal sighted, wild type mouse model.Lastly, we explored the question of the dependency of the cell type to the stimulus. To do so, we designed a stimulus inspired from the previous work; a two bar stimulus, where the bars are spatially separated and sinusoidals at 60 and 80 Hz as the temporal stimulus. We have classified cells as ON or OFF type based on their response to a green full filed flicker (1 Hz) and have subsequently compared the electrical responses of these two classes. For 60 Hz stimulation, we have not observed a strong difference between the response of ON or OFF type cells, but for the case of the 80 Hz stimulation, the OFF-type cells showed a lower activation threshold as opposed to the ON-type ones. However, one important conclusion of this study was that a cell’s response to these particular stimuli is rather influenced by the cell’s position with respect to the stimulus than the cell type.In conclusion, we have shown that under ideal laboratory conditions, high spatial resolution can be achieved with artificial stimulation, either electrical or optogenetic. However, cell-type specificity could not be conclusively shown with eitherof the tested approaches.