Tools for Model Discovery, Validation and Selection in Neuroscience with NeuroML

 描述（由申请人提供）：建立计算模型是促进我们对神经系统的正常功能和与衰老、神经创伤或疾病相关的病理学的理解的最佳方式之一。生物解剖学详细的模型提供了一个框架，跨空间尺度整合数据，并探索有关神经元和网络动力学的生物机制的假设。然而，随着模型复杂性的增加，在模型的创建、验证、交换和重用方面出现了更多的障碍。NeuroML项目旨在通过提供描述神经科学中多尺度模型的标准格式来解决这些问题。NeuroML由30多个工具和数据库支持，是开源大脑模型交换的基础，374名用户正在57个公共建模项目上进行合作。尽管在神经科学界这种有希望的运动模型共享，这是非常罕见的，看到一个具体的，严格的声明用于评估模型在模型开发过程中的标准，和多个模型相同的离子通道和神经元没有比较一致性与同一套实验数据。该项目的总体目标是创建一个灵活的基础设施，用于评估神经科学中计算模型的范围和质量，并使这些信息广泛提供给社区的一大类模型。目标1专注于增强现有工具，以便无缝协作，根据实验数据验证NeuroML模型。目标2侧重于开发一个专门的门户网站，将自动模型验证纳入现有的模型共享平台，并创建文档、教程、论坛和其他外联活动，以促进采用和获得用户反馈。目标3包括在多个大规模神经网络建模环境中测试验证工具链。拟议的活动将建立连接多个现有倡议的桥梁，以支持模型开发，验证，交换，选择和重用，并将实验数据与建模工作相结合，以提高效率，更好的透明度和计算模型在神经科学研究中的更大影响。

 DESCRIPTION (provided by applicant): Building computational models is one of the best ways of furthering our understanding of both the normal function of the nervous system and the pathology associated with aging, neural trauma, or disease. Biophysically detailed models provide a framework for integrating data across spatial scales and for exploring hypotheses about the biological mechanisms underlying neuronal and network dynamics. However, as models increase in complexity, additional barriers emerge to the creation, validation, exchange and re-use of models. The NeuroML project aims to address these issues by providing a standard format for describing multiscale models in neuroscience. NeuroML is supported by over 30 tools and databases and is the basis for model exchange at Open Source Brain, where 374 users are collaborating on 57 public modeling projects. In spite of this promising movement toward model sharing in the neuroscience community, it is extremely rare to see a specific, rigorous statement of the criteria used for evaluating models during model development, and multiple models for the same ion channels and neurons are not compared for concordance with the same suite of experimental data. The overall goal of this project is to create a flexible infrastructure for assessing the scope and quality of computational models in neuroscience and to make this information broadly available to the community for a large class of models. Aim 1 focuses on enhancing existing tools to work together seamlessly for validation of NeuroML models against experimental data. Aim 2 concentrates on the development of a dedicated web portal, incorporation of automated model validation into existing model sharing platforms, and the creation of documentation, tutorials, forums and other outreach for promoting uptake and obtaining user feedback. Aim 3 includes testing of the validation tool chain in multiple large-scale neuronal network modeling environments. The proposed activities will build bridges that connect multiple, existing initiatives in support of model development, validation, exchange, selection, and re-use, and will integrate experimental data with modeling efforts for more efficiency, better transparency, and greater impact of computational models in neuroscience research.