Department of Artificial Intelligence, Hanyang University, Seoul 04763, South Korea.
Department of Electronic Engineering, Hanyang University, Seoul 04763, South Korea.
Curr Biol. 2023 Jan 23;33(2):321-335.e6. doi: 10.1016/j.cub.2022.12.014. Epub 2023 Jan 4.
Visual systems extract multiple features from a scene using parallel neural circuits. Ultimately, the separate neural signals must come together to coherently influence action. Here, we characterize a circuit in Drosophila that integrates multiple visual features related to imminent threats to drive evasive locomotor turns. We identified, using genetic perturbation methods, a pair of visual projection neurons (LPLC2) and descending neurons (DNp06) that underlie evasive flight turns in response to laterally moving or approaching visual objects. Using two-photon calcium imaging or whole-cell patch clamping, we show that these cells indeed respond to both translating and approaching visual patterns. Furthermore, by measuring visual responses of LPLC2 neurons after genetically silencing presynaptic motion-sensing neurons, we show that their visual properties emerge by integrating multiple visual features across two early visual structures: the lobula and the lobula plate. This study highlights a clear example of how distinct visual signals converge on a single class of visual neurons and then activate premotor neurons to drive action, revealing a concise visuomotor pathway for evasive flight maneuvers in Drosophila.
视觉系统使用并行神经回路从场景中提取多种特征。最终，单独的神经信号必须汇聚在一起，以协调地影响行动。在这里，我们描述了果蝇中一个整合与即将到来的威胁相关的多种视觉特征的回路，以驱动逃避性的运动转弯。我们使用遗传扰动方法鉴定了一对视觉投射神经元 (LPLC2) 和下行神经元 (DNp06)，它们是对侧向运动或接近的视觉物体做出逃避性飞行转弯的基础。使用双光子钙成像或全细胞膜片钳技术，我们表明这些细胞确实对平移和接近的视觉模式都有反应。此外，通过测量 LPLC2 神经元在遗传沉默前向运动感觉神经元后的视觉反应，我们表明它们的视觉特性是通过整合两个早期视觉结构（小脑中叶和小脑中叶板）中的多个视觉特征而出现的。这项研究突出了一个明显的例子，即不同的视觉信号如何汇聚到单个视觉神经元类上，然后激活运动前神经元来驱动行动，揭示了果蝇中逃避性飞行机动的简洁视觉运动通路。
