Connectivity of a Locust’s Collision Detection Circuit

We can learn a lot about how a neuronal networks could function by elucidating the wiring pattern of the network. This can be done by reconstructing the neurons in 3D and by mapping their synaptic connections, so that it becomes clear which neurons within the network can communicate with each other and how numerically strong the connections are. We aim to elucidate the wiring within the neuronal network of a locusts collision detector. This detector is special because it enables the locust to avoid colliding with objects on a collision course, and we know that the avoidance reactions are mediated by two single linked neurons either side, termed LGMD1 and DCMD. The LGMD1 becomes excited whenever an object approaches the eye on a collision course. We also know that thousands of upstream neurons, from underneath the compound eye, contribute to the excitation of the LGMD1. But because these TmAs have very small processes, we yet have had no definite idea as to how they contribute to the LGMD1 excitation pattern. We assume that the TmAs signal changes in light levels in distinct facets of the compound eye and are able to inhibit each other to help the LGMD1 distinguish between approaching and passing objects. As we now have a novel, scanning electron microscope available to reconstruct entire neurons, we aim to reconstruct the TmAs and those neurons that pass signals onto them. We also aim to map their synaptic connections, to elucidate how the LGMD1 circuit works. 1

Understanding how neural networks function requires uncovering how individual neurons are connected. This can be achieved by creating 3D reconstructions of neurons and mapping their synaptic connections to reveal which cells communicate with each other and how strong these connections are. In our lab, we use a novel method based on scanning electron microscopy. We automatically generate ultrathin serial sections, photograph them, and then reconstruct the neurons and their synapses-section by section, cell by cell. In this project, we investigated the wiring of a unique neural circuit in the locust brain that detects approaching objects and triggers avoidance behavior. This collision detection system is remarkable because it helps the locust evade obstacles in its path. Two specific neurons-known as LGMD1 and DCMD-are central to this process. LGMD1 becomes active when an object approaches the eye on a direct collision course. Thousands of upstream neurons, located beneath the compound eye, provide input to the LGMD1. These so-called TmAs have extremely fine processes, making them difficult to study in detail until now. It is believed that each TmA is linked to a specific facet of the compound eye and can inhibit its neighboring TmAs. This inhibitory interaction likely enhances the LGMD1's ability to distinguish between objects that are approaching and those simply passing by. In our work, we fully reconstructed individual TmAs along their entire length and analyzed the terminal branches of 90 such neurons in the region surrounding the LGMD1. These detailed datasets form the foundation for computational models of this biologically advanced circuit. In the future, such bio-inspired models could help improve artificial collision detection systems-for example, in robotics, or in road and air traffic applications.