Computational Infrastructure for Brain Research: EAGER: Next-Generation Neural Data Analysis (NGNDA) Platform: Massive Parallel Analysis of Multi-Modal Brain Networks

在过去的十年中，前所未有的技术进步促进了对大脑时空细节的精细研究。大规模努力的雄心勃勃的目标，包括那些由BRAIN Initiative支持的目标，包括长时间同时测量数千个神经元，并生成非常高分辨率的大脑及其活动图像。然而，这些技术将产生大量的数据，而对这些大型数据集的大规模分析几乎不可能用目前可用的计算工具来完成。该项目的重点是通过开发一种新颖且可广泛访问的下一代神经数据分析（NGNDA）平台来分析和集成大量异构大脑数据，从而解决许多计算限制。这是来自哈佛医学院研究计算、波士顿儿童医院/哈佛医学院认知神经科学实验室、贝斯以色列女执事医疗中心/哈佛医学院睡眠和炎症系统实验室以及波士顿儿童医院癫痫部门和南卡罗来纳州医科大学的神经科学研究人员、算法开发人员和计算技术专家的共同努力。 最终目标是不仅要了解健康的大脑，还要了解影响越来越多人口的复杂疾病和障碍，从而导致巨大的社会经济成本。因此，该项目符合NSF促进科学进步和促进国家健康、繁荣和福利的使命，是开发新的计算基础设施NGNDA的先驱，该基础设施将专门为合作研究而设计，以促进跨物种大脑连接体的大规模分析、模拟和建模。其总体目标是开发大神经数据的智能并行处理方法，并通过实施创新算法和动态利用和整合共享的机构和国家高性能计算（HPC）资源，有效估计神经组织规模上的连接体。NGNDA平台将在哈佛医学院研究计算提供的Orchestra HPC资源上实现，并在国家科学基金会支持的极端科学和工程发现环境（XSEDE）国家计算联盟的资源上实现。NGNDA将使用四个非常高维的神经数据集进行验证，每个数据集都构成了独特的计算挑战。NGNDA旨在促进神经科学研究的“融合”方法，将不同学科的专业知识和见解与尖端资源和工具相结合，以全面研究大脑。NGNDA计算基础设施将可供数千名用户访问，其所有算法和验证数据将免费提供给社区; NGNDA最终也可能成为一个新的电子学习平台，用于下一代神经科学家的多方面协作学习和教育。NGNDA还将促进测试研究结果的可重复性和概括性。CISE高级网络基础设施部门的探索性研究（EAGER）早期概念赠款奖由SBE行为和认知科学部门共同支持，与NSF理解大脑活动相关的资金，包括用于开发神经科学的国家研究基础设施，并与国家战略计算计划下的NSF目标保持一致。

Unprecedented technological advances over the last decade have facilitated investigation of the brain at exquisite levels of spatial-temporal detail. Ambitious goals of large-scale efforts, including those supported by the BRAIN Initiative, include simultaneously measuring from thousands of neurons for long periods of time, and generating very high resolution images of the brain and its activity. However, enormous volumes of data will be produced by these technologies, and grand-scale analyses of these large datasets are virtually impossible to accomplish with currently available computational tools. This project is focused on addressing a number of these computational limitations by developing a novel and broadly accessible Next-Generation Neural Data Analysis (NGNDA) platform to analyze and integrate large volumes of heterogeneous brain data. This is a collaborative effort of neuroscience researchers, algorithm developers, and computing technology experts from Harvard Medical School Research Computing, the Laboratory of Cognitive Neuroscience at Boston Children's Hospital/Harvard Medical School, the Sleep and Inflammatory Systems Laboratory at Beth Israel Deaconess Medical Center/Harvard Medical School and the Epilepsy Divisions at Boston Children's Hospital and the Medical University of South Carolina. The ultimate goal is to understand not only the healthy brain but also complex diseases and disorders that are affecting progressively larger populations resulting in enormous socioeconomic costs. This project therefore aligns with the NSF mission to promote the progress of science and to advance the national health, prosperity and welfare.This project is a pioneer effort to develop a new computational infrastructure, NGNDA, which will be specifically designed for collaborative research, to facilitate grand-scale analysis, simulation and modeling of brain connectomes across species. The overarching goal is to develop the means for intelligent parallel processing of big neural data, and efficient estimation of connectomes across scales of neural organization, via implementation of innovative algorithms and dynamic leveraging and integrating of shared institutional and national high performance computing (HPC) resources. The NGNDA platform will be implemented on the Orchestra HPC resource provided by Harvard Medical School Research Computing, and on resources of the Extreme Science and Engineering Discovery Environment (XSEDE) national computing consortium supported by the National Science Foundation. NGNDA will be validated with four very high-dimensional neural datasets, each posing a unique computational challenge. NGNDA aims to facilitate a "convergence" approach to Neuroscience research whereby expertise and insights from distinct disciplines are merged with cutting-edge resources and tools for a comprehensive investigation of the brain. The NGNDA computational infrastructure will be accessible to thousands of users and all its algorithms and validation data will be freely available to the community; and NGNDA may also eventually serve as a novel e-learning platform for multifaceted collaborative learning and education of next-generation neuroscientists. NGNDA will also facilitate testing the reproducibility and generalization of research findings.This Early-concept Grants for Exploratory Research (EAGER) award by the CISE Division of Advanced Cyberinfrastructure is jointly supported by the SBE Division of Behavioral and Cognitive Sciences, with funds associated with the NSF Understanding the Brain activity including for developing national research infrastructure for neuroscience, and alignment with NSF objectives under the National Strategic Computing Initiative.