Are dendritic integration rules in retinal ganglion cells adapted to the statistics of the natural environment?

动物生活的环境中，不同的视觉信息存在于视野的不同部分，因此，福尔斯落在不同的视网膜区域。例如，许多陆生动物的世界被地平线一分为二：对于老鼠来说，捕食者更有可能出现在上部视野中，而食物则出现在下部视野中。为了更有效和鲁棒地编码与生存相关的信息，视觉系统被认为是适应自然环境的统计数据。这种适应已经在视网膜中开始了，视网膜通常可以分为不同的区域：例如，在这些区域中，感光器群体的光谱组成似乎专门用于不同的任务，例如小鼠中的猛禽检测或斑马鱼幼虫中的狩猎。目前，我们只是在开始了解这些区域特定的适应。然而，在没有深入了解区域回路和潜在的适应机制的情况下，很难全面了解视网膜如何处理动物的自然视觉环境。在本项目中，我们将重点关注视网膜网络的输出神经元，即视网膜神经节细胞。在这些细胞中，树突形态的变化通常是区域特化的标志。因此，我们将研究神经节细胞树突状动脉整合突触前回路输入的规则，并评估这些规则是否以及如何在视网膜上发生变化。然后，我们问这样的适应可以subserve有效处理的自然视觉environment.Specifically，使用功能性树突状成像，形态重建和计算建模，我们将研究如何不同类型的小鼠RGC处理上游电路的突触输入，以及如何在一个RGC类型的树突整合配置文件的区域变化可能有助于编码这个视网膜区域遇到的自然场景。我们将在“最小”生物物理RGC模型中捕获这些适应的本质，该模型使我们能够匹配来自不同视网膜区域的测量数据，从而提取确定RGC功能的重要参数。除了深入了解RGC中树突整合和加工的机制外，我们还希望进一步了解RGC如何促进视网膜32个以上输出通道的产生（目标1）。我们期望了解所选类型的RGC的树突整合功能如何以及在多大程度上随着视网膜位置和视野位置依赖性自然输入统计而变化（目标2），并使用最小生物物理模型提取这些适应的基本一般原则（目标3）。

Animals live in environments in which different visual information is present in distinct parts of the visual field and, hence, falls onto different retinal regions. For example, the world of many terrestrial animals is bisected by the horizon: For mice, predators are more likely to appear in the upper visual field, whereas food is rather found in the lower visual field. To encode survival-relevant information more efficiently and robustly, the visual system is thought to be adapted to the statistics of the natural environment. This adaptation starts already in the retina, which often can be divided into different zones: in these, for example, the spectral composition of the photoreceptor population seems to be specialized for distinct tasks, such as raptor detection in mice or hunting paramecia in zebrafish larvae. Currently, we are only at the beginning of understanding these region-specific adaptations. However, without an in-depth knowledge of the regional circuits and the underlying adaptational mechanisms, a comprehensive understanding of how the retina processes an animals’ natural visual environment is difficult.In this project, we will focus on the output neurons of the retinal network, the retinal ganglion cells. In these cells, changes in dendritic morphology are often a tell-tale sign for regional specializations. Therefore, we will study the rules according to which ganglion cell dendritic arbours integrate input from the presynaptic circuitry and evaluate if and how these rules change across the retina. We then ask how such adaptations could subserve efficient processing of the natural visual environment.Specifically, using functional dendritic imaging, morphological reconstruction, and computational modelling, we will study how different types of mouse RGC process the synaptic input of the upstream circuitry and how regional variations in an RGC type’s dendritic integration profile may contribute to encoding the natural scene encountered by this retinal region. We will capture the essence of these adaptions in "minimal" biophysical RGC models which allow us to match measured data from different retinal regions and, thereby, extracting important parameters that determine RGC function. In addition to insights into the mechanisms of dendritic integration and processing in RGCs, we expect to further our understanding of how RGCs contribute to the generation of the more than 32 output channels of the retina (Objective 1). We expect to learn how and to what extent the dendritic integration function of selected types of RGC changes with retinal location and visual field position-dependent natural input statistics (Objective 2) and extract the underlying general principles of these adaptations using minimal biophysical models (Objective 3).