Dendritic integration of spatial and object features in a Drosophila sensorimotor circuit.

Research Abstract Single neurons perform complex computations that depend critically on the interactions between often precisely targeted synaptic inputs and the intrinsic active and passive properties of the neuron. While some functional implications of synaptic topography and intrinsic properties for neural computation have been explored, their interactions remain poorly understood. Difficulties in simultaneously localizing synaptic inputs, effectively activating synapses with realistic time courses and functional tuning, and manipulating neuron intrinsic properties, pose challenges in exploring these interactions which collectively shape neuronal computations. This study aims to address this gap by using a combination of computational and experimental approaches, focusing on the looming detection circuits of the fruit fly Drosophila melanogaster, which drive motor responses to objects approaching on a direct collision course. These circuits contain descending neurons (DNs) that integrate overlapping inputs from visual projection neuron (VPN) populations, that encode different features of the looming stimulus (such as angular velocity or size). By leveraging experimental access to VPNs and DNs, together with the comprehensive electron microscopy dataset of the full adult fly brain, we develop detailed biophysical models to investigate how the spatiotemporally realistic activation of VPN synapses influences integration processes and interacts with intrinsic properties of DNs. Through experimental manipulations and computational simulations, we aim to uncover the mechanisms that underlie the integration of synaptic inputs in DNs to drive realistic responses. Overall, this study will provide insights into visual feature integration within VPN-DN circuits and general mechanisms underlying dendritic integration.