Joshua L. Morgan, Ph.D. - US grants

PROJECT SUMMARY Many computations in the nervous system occur at the level of individual neurites within large extensively branched arbors (i.e., subcellular processing). Most neurites operate in dense neuropils, in which processes of diverse cell types are tightly packed and abundantly interconnected. To understand subcellular processing, we need to measure neurite responses to physiological stimuli and relate them to local patterns of synaptic inputs. To delineate the functional architecture of neuropils and reveal the logic of their connectivity, we need to characterize neurite responses and synapse patterns at high density. Neurite responses can be observed by two-photon imaging, and synaptic inputs can be reconstructed in serial-section electron microscopy (ssEM). A number of technical obstacles have precluded the combination of these techniques (i.e., functional connectomics) to study subcellular processing in dense neuropils. Here, we develop new tools and approaches to overcome these obstacles. In Aim 1, we develop genetic, viral, and computational tools for multispectral two-photon calcium imaging and signal demixing to enable dense functional characterization of neuropils. In Aim 2, we devise a novel strategy for combining two-photon imaging and ssEM (i.e., multimodal imaging), and establish a high-throughput ssEM method for analyzing local synaptic connectivity patterns in the context of larger-scale circuit wiring (i.e., multiresolution imaging). We use our advances to study amacrine cells (ACs), a diverse class of retinal interneurons. The neurites of more than 50 AC types extract salient visual information in a dense neuropil the inner retina. We will acquire a complete functional connectomic dataset of ACs. This dataset, which will be made publicly available, will form the basis of a future R01 application to study the mechanisms of subcellular processing in ACs, the functional architecture of the AC neuropil, and the logic of its connectivity.

The functional architecture of the nervous system is usually thought of in terms of groups of neurons with similar functions being clustered together. However, groups of synaptic connections with similar functions also cluster together resulting in a fine-scale (1-10 ?m) functional architecture of nervous tissue. The significance of this fine-scale microcircuitry has been difficult to asses because it can involve dozens of neurons computing signals that are not readily detectable at the level of neuronal cell bodies. Here we propose an approach to studying developmental and functional principles of local microcircuits in the mouse visual thalamus. In the dorsal lateral geniculate nucleus (dLGN), axons from the retina (RGCs) form glia encapsulated clusters of synapses (glomeruli) with thalamocortical relay cells and inhibitory neurons. The functional significance of these local microcircuits is unknown. Based on the diverse patterns of glomerular connectivity we observed in our previous studies of mouse dLGN, we believe these microcircuits are preforming visual-channel-specific computations ranging from simple relay of signals to complex feature detection. We also hypothesize that, central to the organization of these microcircuits, is the use of developmental retinal activity to group functionally related RGCs and inhibitory neurites together in the same glomeruli. Our first step in testing the above hypotheses is to create a detailed mapping of the development of mouse dLGN glomerular microcircuits using serial section electron microscopy that will reveal which cell types initiate glomerulus formation and which aspects of the microcircuit?s connectivity appear during activity dependent synaptic remodeling. We will then test whether visual deprivation or transgenic silencing of subsets of RGC inputs alters the grouping together of neurites or the specificity with which they form synapses. The results of these developmental studies will reveal the extent to which dLGN microcircuit structure is the result of visual experience. We will then probe the function of dLGN glomeruli by first recording the response properties of thalamocortical cells and then reconstructing the glomerular microcircuitry of those same neurons with electron microscopy. By matching receptive field properties to their microcircuit configurations, we will learn whether different glomerular types are specific to different channels of visual processing and will gain insight into the computations likely to be executed by different local microcircuits. We will next use optogenetic stimulation of RGCs to determine how different glomerular configurations integrate signals from converging inputs. The proposed experiments will reveal the origin, organization and function of the fine-scale microcircuitry of an important model system for vision and circuit development.