Connectivity of the Outer Plexiform Layer of the Mouse Retina

Abstract

The retina has two synaptic layers: In the outer plexiform layer (OPL), signals from the photoreceptors (PRs) are relayed to the bipolar cells (BCs) with one type of horizontal cell (HC) as interneuron. In the inner plexiform layer (IPL), the retinal ganglion cells (RGCs) receive input from the bipolar cells, modulated by multiple types of amacrine cells. The axons of the retinal ganglion cells form the optic nerve which transmit the visual signal to the higher regions of the brain (Masland 2012). Studies of signal processing in the retina usually focus on the inner plexiform layer. Here, the main computations take place such as direction selectivity, orientation selectivity and object motion detection (Gollisch and Meister 2010). However, to fully understand how these computations arise, it is also important to understand how the input to the ganglion cells is computed and thus to understand the functional differences between BC signals. While these are shaped to some extent in the IPL through amacrine cell feedback (Franke et al. 2017), they are also influenced by computations in the OPL (Drinnenberg et al. 2018). Accordingly, it is essential to understand how the bipolar cell signals are formed and what the exact connectivity in the OPL is. This thesis project aims at a quantitative picture of the mouse outer retina connectome. It takes the approach of systematically analyzing connectivity between the cell types in the OPL based on available high-resolution 3D electron microscopy imaging data (Helmstaedter et al. 2013). We reconstructed photoreceptor axon terminals, horizontal cells and bipolar cells, and quantified their contact statistics. We identified a new structure on HC dendrites which likely defines a second synaptic layer in the OPL below the PRs. Based on the reconstructed morphology, we created a biophysical model of a HC dendrite to gain insights into potential functional mechanisms. Our results reveal several new connectivity patterns in the mouse OPL and suggest that HCs perform two functional roles at two distinct output sites at the same time. The project emphasizes how large-scale EM data can boost research on anatomical connectivity and beyond and highlights the value of the resulting data for detailed biophysical modeling. Moreover, it shows how the known amount of complexity increases with the level of detail with which we can study a subject. Beyond that, this thesis project demonstrates the benefits of data sharing and open science which only enabled our studies.