CRCNS: Sensory-Motor Integration in Mammalian Brian: experiment, analysis, modeling

CRCNS：哺乳动物布莱恩的感觉运动整合：实验、分析、建模

• 批准号: 9524750
• 负责人: RONALD R COIFMAN
• 金额: $ 20.53万
• 依托单位: YALE UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-07-15 至 2020-06-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9524750
• 关键词: Address; Animal Behavior; Animals; Area; Behavior; Behavioral; Biological; Biological Neural Networks; Brain; Chronic; Communication; Complex; Data; Data Analyses; Data Collection; Data Set; Development; Dimensions; Disease; Distributed Systems; Electrical Engineering; Engineering; Esthesia; Feedback; Formulation; Future; Gaussian model; Goals; Head; Huntington Disease; Instruction; Israel; Knowledge; Language; Lead; Learning; Literature; Machine Learning; Mathematics; Methodology; Methods; Microscopy; Modality; Modeling; Motor; Motor Activity; Motor Cortex; Movement; Mus; Nature; Neurons; Neurosciences; Noise; Organism; Parkinson Disease; Pharmacogenetics; Physiology; Population; Process; Prosthesis; Psychophysics; Research Proposals; Resolution; Robotics; Role; Sensory; Structural Models; Structure; System; Uncertainty; Universities; Viral; Work; analytical method; awake; base; brain computer interface; calcium indicator; cell type; control theory; cost; experience; experimental study; functional plasticity; high dimensionality; insight; interdisciplinary approach; mathematical analysis; medical schools; motor control; motor impairment; motor learning; neural circuit; novel; operation; optogenetics; sensorimotor system; signal processing; spatiotemporal; success; theories; tool; two-photon

A major goal facing organisms acting in the natural world is the selection of appropriate actions based on sensory information and prior knowledge accumulated through previous experience. Understanding how neural networks process information and control such actions requires a breakthrough both in large scale chronic data collection methods during behavioral tasks and in the development of new analysis and modeling tools that will be able to capture the dynamics and organization of such neural networks. To address key questions in the context of sensory-motor control and learning we propose a multidisciplinary approach that will synergize the expertise of the four groups involved in cellular and systems neuroscience, machine learning, signal processing, control theory and modeling. Our goal is to establish an empirically-grounded systems-level model explaining the interaction and integration within the sensory-motor system during behavioral tasks, which is consistent with the experimental data, and which provides concrete predictions for future experiments. More specifically, we intend to further our understanding on two main fronts. First, study cell type specific components that participate in the movement command and in sensory-motor error prediction. We hypothesize that layer 2-3 neurons subserve different roles from layer 5 neurons, and may be more strongly involved in error estimation rather than in control. Second, we intend to investigate whether and how the sensory and motor ends change in order to adapt to the new learned task. Here again we expect differences between the different cortical layers. Such a framework will not only provide new insights into the specific investigated system, but could be transferrable more generally to probe the structure and functionality of complex biological networks. In addition, the unique analysis methods developed and the deep understanding of biological sensory-motor systems may contribute invaluably to fields such as robotics and network control, and to the development of new prosthetics approaches within the field of Brain Computer Interfaces. RELEVANCE (See instructions): The goals of this research proposal are expected to provide novel insight on sensory-motor control as well as structural and functional plasticity processes of the cortical network during sensory-motor learning. Deep understanding of sensory-motor systems will aid in development of new treatment modalities for diseases that impair motor function, such as Parkinson's and Huntington's diseases.

在自然界中活动的生物体面临的一个主要目标是根据以下因素选择适当的行动： 感官信息和通过先前经验积累的先验知识。了解如何 神经网络处理信息和控制这种行为需要在大规模和大规模两个方面取得突破 在行为任务和开发新分析过程中的慢性数据收集方法， 建模工具，将能够捕捉这种神经网络的动态和组织。 为了解决感觉运动控制和学习背景下的关键问题，我们提出了一个 多学科的方法，将协同参与细胞和 系统神经科学、机器学习、信号处理、控制理论和建模。我们的目标是 建立一个系统级模型，解释系统内的相互作用和集成， 感觉运动系统在行为任务，这是与实验数据一致， 为未来的实验提供了具体的预测。更具体地说，我们打算进一步 在两个主要方面的理解。首先，研究细胞类型的具体组成部分，参与 运动命令和传感器-电机误差预测。我们假设第2 - 3层神经元 与第5层神经元的作用不同，并且可能更强烈地参与误差估计，而不是 比在控制。第二，我们打算调查是否以及如何感觉和运动结束改变， 以适应新的学习任务。这里我们再次期待不同的皮层之间的差异 层次。 这样的框架不仅将为具体的调查系统提供新的见解，而且可以 更广泛地说，它可以用来探测复杂生物网络的结构和功能。在 此外，独特的分析方法和对生物感觉运动的深刻理解 系统可能会对机器人和网络控制等领域做出宝贵的贡献， 脑机接口领域的新修复方法。 相关性（参见说明）： 这项研究计划的目标预计将提供新的见解，以及对感觉运动控制 作为感觉运动学习期间皮质网络的结构和功能可塑性过程。深 对感觉-运动系统的理解将有助于开发新的疾病治疗方式 如帕金森氏症和亨廷顿氏症。