Characterization of visual feature encoding in Drosophila Visual Projection Neurons LPLC1 and LPLC2 - Abstract A critical function performed by visual systems is the fast and reliable detection of features within the environment, such as moving objects or approaching threats. Neurons predicted to encode these higher order, behaviorally relevant visual features have been identified in a variety of species. However, a lack of cell-type specific genetic tools, accessible neural anatomy, and clearly mapped neural circuitry has limited the identification of these cell-types, and led to an incomplete understanding of the mechanisms which govern feature encoding in visual systems. Here, we investigate visual feature encoding by leveraging Drosophila melanogaster. We specifically investigate mechanisms governing visual feature encoding in visual projection neurons (VPNs), a class of neurons predicted to selectively encode visual features. We focus on two VPN populations, Lobula Plate/Lobula Columnar type 1 (LPLC1) and Lobula Plate/Lobula Columnar type 2 (LPLC2). Previous studies have suggested these two neuron types encode the visual features of an object approaching on a direct collision course, however their hypothesized selectivity to visual features and the origins of this selectivity have yet to be thoroughly investigated. Differences in recording location and stimulus parameters during Ca2+ imaging have made it difficult to resolve feature tuning across studies. Additionally, our preliminary data that demonstrate these neurons are spiking, contrary to prior assumptions, suggest feature selectivity may emerge from a spike timing code, which has to date been ignored. Thus, there is a critical need to study these neuron populations using whole-cell electrophysiology and precisely designed stimuli to understand the mechanisms governing their feature encoding and reveal their selectivity to specific visual features. The goal of this project is to uncover feature tuning for these neurons at the level of individual spikes and investigate the mechanisms governing their feature selectivity. To accomplish this goal, we will first map the receptive fields of LPLC1 and LPLC2 and characterize responses to an array of visual stimuli to establish their feature selectivity (Aim 1). Following this, we will investigate the mechanisms of inhibition that shape both the feature and receptive field selectivity (Aim 2). Finally, we will investigate how neuron active properties further govern feature selectivity within these neurons through multicompartment modelling (Aim 3). Findings from our work will reveal the specificity and underlying mechanisms for feature encoding within two VPN types, which may serve as a foundation for investigating feature detecting neurons in other, more complex species. Drosophila possess a visual system with striking similarities to vertebrate visual systems, so our findings may provide general principles for collision detection. In the longer term, a more fundamental understanding of visual feature encoding in fruit flies may help the development of human therapeutics for those suffering from blindness.
