Characterization of the Distribution of Input and Output Synapses in Drosophila melanogaster Central Complex Neurons

Abstract

Unlike in the mammalian brain, most fly neurons are unipolar and thus have mixed synaptic inputs and outputs along the same arbor regions. Previous literature has yet to characterize the spatial distributions of input and output synapses, specifically for these neurons with mixed synaptic profiles. Recently, the Murthy and Seung labs developed FlyWire, an electron microscopy (EM) and AI based full Drosophila melanogaster connectome, which includes synaptic connectivity data at the single neuron level. This novel dataset was used to determine the extent to which inputs and outputs are mixed in the Drosophila melanogaster central complex (CX), a brain region implicated in head direction representation, spatial navigation, and sleep. An agglomerative clustering algorithm employing an along-skeleton distance function proved to be an effective technique for separating synapses into distinct groups based on arbor location. The results showed 1) CX neurons tend to exhibit a greater number of outputs compared to inputs across neurons as a whole, 2) neurons with three clusters tend to have an input dominated arbor, an output dominated arbor, and an arbor with a more equal number of inputs and outputs, and 3) neurons of the same cell type exhibit consistency, both in terms of raw numbers of inputs and outputs and ratios of inputs and outputs. While further research is needed to answer questions raised by the present findings, the novel clustering technique and discovered distributions of mixed inputs and outputs are important because they can be applied to analysis of synaptic connectivity.