The neural basis of insect pattern vision

视觉是我们进入物理世界的主要通道之一。由于其对人类和动物行为的核心重要性，了解神经处理如何产生视觉感知是神经科学研究的关键问题之一。为了从这个应用程序中的文本中提取含义，你目前正在使用视觉系统的一个核心功能：模式视觉。它在交流、物体识别和定向方面具有重要功能，构成了动物视觉的基本支柱。在其最一般的形式，它提供了一个大小，位置和方向不变的表示模式。鉴于我们的大脑有数亿个神经元用于这项任务，更令人惊讶的是，大脑比米粒还小的昆虫也能识别和记忆视觉模式。有些昆虫传粉者甚至具有概括模式特征的能力。因此，他们提供了一个模型来研究神经实现的不变模式识别有限的计算资源。虽然昆虫模式识别行为已被广泛研究，但对其神经元的实现知之甚少。我计划缩小这一差距，使用一个hawkmoth作为一个易于处理的模型来剖析昆虫的不变模式视觉电路从光子到behavior.The蜂鸟hawkmoth（Macroglossum stellatarum）结合了一个生理上可访问的神经系统和一个生物学相关的行为范式与研究不变模式识别。我计划用多种方法剖析这种行为背后的神经机制：（1）定量行为评估将确定天蛾模式视觉的参数空间，（2）细胞内记录将提供昆虫大脑中模式响应神经元的第一个特征。为了从行为到生理学的桥梁，（3）在虚拟现实中主动扫描模式的天蛾中的四极记录将揭示模式如何在行为昆虫中编码，以及主动视觉如何有助于模式编码。(4)一个计算模型将使机械的洞察模式视觉网络，超越那些可能在体内，其空中机器人的实施将揭示主动vision.The研究计划的成果将是视觉神经科学的最佳策略具有很强的指导意义。在天蛾中描述的神经机制将为跨昆虫物种的比较研究奠定基础。将天蛾的模式处理策略与脊椎动物的模式处理策略进行比较，将揭示昆虫用于科普其明显较低的处理能力的收敛神经策略以及发散解决方案。我们的计算模型的机器人实现将提供模式识别的独特应用。该项目将为动物视觉、神经科学、昆虫感觉生态学和机器人技术领域做出重要贡献。

Vision is one of our major access ports to the physical world. Because of its central importance to human and animal behaviour, understanding how neural processing gives rise to the visual percept is one of the key questions in neuroscience research. To extract meaning from the text in this application, you are currently using a central feature of your visual system: pattern vision. With essential functions in communication, object recognition, and orientation, it constitutes a fundamental pillar of animal vision. In its most generalised form it provides a size-, position- and orientation-invariant representation of a pattern. Given that our brain dedicates several hundred million neurons to this task, it is all the more astonishing that insects, with brains smaller than a grain of rice, recognise and memorise visual patterns as well. Some insect pollinators even possess the ability to generalise pattern features. They thus provide a model to study the neural implementation of invariant pattern recognition with limited computational resources. While insect pattern discrimination behaviour has been studied extensively, very little is known about its neuronal implementation. I plan to close this gap, using a hawkmoth as a tractable model to dissect the invariant pattern vision circuits of insects from photons to behaviour.The hummingbird hawkmoth (Macroglossum stellatarum) combines a physiologically accessible nervous system and a biologically relevant behavioural paradigm with which to study invariant pattern recognition. I plan to dissect the neural mechanisms underlying this behaviour with multiple methods: (1) a quantitative behavioural assessment will determine the parameter space for hawkmoth pattern vision, and (2) intracellular recordings will provide the first characterisation of pattern-responsive neurons in the insect brain. To bridge from behaviour to physiology, (3) tetrode recordings in hawkmoths that actively scan patterns in virtual reality will reveal how patterns are encoded in behaving insects and how active vision contributes to pattern encoding. (4) A computational model will enable mechanistic insights into the pattern vision network, beyond those possible in vivo, and its aerial robot implementation will reveal optimal strategies for active vision.The outcomes of this research programme will be highly instructive for visual neuroscience. The neural mechanisms described in hawkmoths will lay the basis for comparative investigations across insect species. Comparing the pattern processing strategies of hawkmoths to that in vertebrates will reveal converging neural strategies, as well as diverging solutions that insects use to cope with their distinctly lower processing power. The robotic implementation of our computational model will provide a unique application of pattern recognition. Together, this project will deliver key contributions to the fields of animal vision, neuroscience, insect sensory ecology, and robotics.