Computer Methods for Physiological Problems

描述(申请人提供)：借助计算机模拟，最好地理解解剖学和生物物理特性对神经元和神经回路功能的贡献。Neuron是我们为Mac OS X、MS Windows和Unix免费开发和提供的一个程序，它简化了非专业人士在数值方法和编程方面的神经模型的创建和分析。它可以在工作站、集群和大规模并行超级计算机上模拟单个神经元和神经元网络。模型的特性可以包括但不限于复杂的分支形态、多种通道类型、不均匀的通道分布、离子扩散、细胞外场、电子仪器和人工放电神经元。神经元被世界各地的神经科学家用来研究与先天性和获得性疾病(如癫痫、多发性硬化症和学习记忆障碍)有关的细胞和网络机制，以及药物和脑深部刺激等治疗干预措施如何影响这些机制。随着越来越多的研究人员、实验者和理论家将其纳入他们的研究计划，神经元使用者的社区正在稳步增长。由于神经科学是一个快速发展的领域，研究人员的需求总是在变化。因此，神经元必须在几个关键领域继续发展，才能继续满足神经科学研究不断发展的要求。需要更多的计算资源；我们将通过使用针对广泛使用的网络拓扑类别优化的可选方法来扩展神经元的新并行功能来解决这一问题。需要访问在其他生物科学、物理科学和工程学中可用的分析工具；为此，神经元正在采用一种现代编程语言(Python)。需要更灵活和更强大的方法来执行动态钳位实验；神经元可以在这种模式下获得极大的优势，但必须减少设置、硬件测试和实验配置的工作量。只有在易于安装和使用的情况下，神经元社区才能充分利用这些高级功能。我们还确定了需要创建新的图形用户界面工具以促进模型规范的使用模式：用于指定新的突触机制和人工尖峰细胞的有限状态机构建器，以及用于指定离子和第二信使积累机制的几何和动力学的工具。这些需求需要在提高性能、健壮性、打包、文档编制、培训以及用户咨询和协作等方面共同努力。在这项提案中，我们提出了一项旨在解决这些领域的研究计划。叙事性神经科学家使用神经元和神经回路的计算模型来更好地了解神经系统在健康和疾病中的功能。迫切需要开发模拟软件和用户支持，以便他们能够在并行计算机上有效地运行这些模型并访问最先进的分析工具。该项目通过构建神经元模拟器环境的优势来满足这些需求，该环境已经拥有大量的用户基础，并正在被越来越多的神经科学家采用。

DESCRIPTION (provided by applicant): The contributions of anatomy and biophysical properties to the function of neurons and neural circuits are best understood with the aid of computer simulations. NEURON, a program which we have developed and provide freely for Mac OS X, MS Windows, and UNIX, has simplified the creation and analysis of neural models for nonspecialists in numerical methods and programming. It can simulate individual neurons and networks of neurons on workstations, clusters, and massively parallel supercomputers. Model properties may include, but are not limited to, complex branching morphology, multiple channel types, inhomogeneous channel distribution, ionic diffusion, extracellular fields, electronic instrumentation, and artificial spiking neurons. NEURON is used by neuroscientists around the world to investigate cellular and network mechanisms that are involved in inborn and acquired disorders such as epilepsy, multiple sclerosis, and disorders of learning and memory, and how they are affected by therapeutic interventions such as medications and deep brain stimulation. The community of NEURON users is growing steadily as increasing numbers of investigators, experimentalists and theoreticians alike, incorporate it into their research plans. Since neuroscience is a rapidly advancing field, investigators' needs are always changing. Consequently NEURON must continue to develop in several critical areas in order to continue to satisfy the evolving requirements of neuroscience research. There is a need for ever greater computational resources; we will address this by extending NEURON's new parallel features with selectable methods optimized for widely used classes of network topology. There is a need to access analysis tools that have become available in other biological sciences, the physical sciences, and engineering; to this end, NEURON is adopting a modern programming language (Python). There is a need for more flexible and powerful ways to perform dynamic clamp experiments; NEURON can be employed in this mode to great advantage, but the setup, hardware testing, and experiment configuration effort must be reduced. The NEURON community can fully exploit these advanced capabilities only if they are easy to install and use. We have also identified patterns of usage that call for the creation of new GUI tools to facilitate model specification: a finite state machine builder for specifying new synaptic mechanisms and artificial spiking cells, and a tool for specifying the geometry and kinetics of ion and second messenger accumulation mechanisms. These needs require concerted efforts in the areas of increasing performance, robustness, packaging, documentation, training, and user consultation and collaboration. In this proposal we present a research plan that is designed to address these areas. Narrative Neuroscientists use computational models of neurons and neural circuits to better understand how the nervous system functions in health and disease. There is a critical need for development of simulation software and user support so that they can run these models efficiently on parallel computers and access state-of-the-art analysis tools. This project addresses these needs by building on the strengths of the NEURON Simulator Environment, which already has a large user base and is being adopted by increasing numbers of neuroscientists.