Developing Cloud-based tools for Big Neural Data

描述（由申请人提供）：大数据有可能以类似于改变遗传学的方式极大地推进电生理学生物数据科学。这两个领域之间的差异决定了应用大数据工具的不同方法，以及为研究社区提供成功资产的方法。首先，神经数据集本质上是非常异构的。在不了解数据采集协议、实验范例和记录的受试者的生理状态的具体情况下，难以解释数据。许多神经数据集都补充了复杂的元数据集，这应该是与其他研究人员整合和共享这些数据的任何努力的一个组成部分。该项目的目标是开发新颖的，可推广的大数据工具，以促进复杂的多尺度神经数据的云分析。癫痫研究将被用作指导工具开发的特定用例。一组进行癫痫研究的资深研究人员将在他们的实验室中使用和验证这些工具。癫痫研究目前局限于其狭隘的集中在单一的模型（动物或人类）在个别中心和实验室。正如遗传学通过大数据技术发生了革命性变化一样，癫痫研究也可以通过新的方法来改变，以标准化，共享和挖掘研究人员之间的数据。在过去的几年里，我共同开发了一个由NINDS资助的基于云的数据平台：ieeg.org，这让我在开发神经数据的大数据解决方案方面发挥了核心作用，例如定制数据共享，大规模基于云的数据分析，以及复杂数据和元数据的搜索和询问技术。我对这个项目的科学目标是：（1）开发可推广的工具来管理，分析和询问多尺度神经数据，（2）创建一个平台，激发一个专注于共享数据的研究社区，以及推进基础和转化神经科学大数据研究的方法。同样重要的是，我提出了一个培训计划，为我的学术生涯做好准备，专注于神经科学中的大数据。该计划补充了我在生物工程和神经数据统计建模方面的背景，在数据集成和机器学习方面具有更广泛的数据科学专业知识，以及临床神经科学的更深入领域知识。我已经召集了一组合作者，基础研究人员和临床科学家，他们将使用本项目开发的工具来分析 验证他们的数据和方法。我将使用这个项目的结果作为R 01资助申请的基础，在该项目中，我将扩展开发的平台和工具，以针对其他研究领域（TBI，急救护理，心脏），以及整合其他数据模式，如成像和基因组学。OMB编号0925-0001/0002（2012年8月批准至2015年8月31日修订版）页码续页格式页码

DESCRIPTION (provided by applicant): Big data has the potential to dramatically advance the electrophysiology biodata sciences in similar ways that it has transformed Genetics. Differences between these two areas dictate separate approaches to apply Big Data tools, and methods in order to provide successful assets to the research community. For one, neural datasets are very heterogeneous by nature. The data is difficult to interpret without knowing specifics about the data acquisition protocol, the experimental paradigm and the physiological state of the recorded subject. Many neural datasets are complemented with complex meta-data sets, which should be an integral component in any effort to integrate and share these data with other researchers. The goal of this project is to develop novel, generalizable Big Data tools to facilitate cloud-base analysis of complex multi-scale neural data. Epilepsy research will be used as a specific use case to guide the development of the tools. A cohort of established senior investigators performing epilepsy research will use and validate these tools in their laboratories. Epilepsy research is currently limited by its narrow focus on single models (animal or human) in individual centers and laboratories. Just as Genetics was revolutionized through Big Data techniques, so too can Epilepsy research be transformed through novel approaches to standardize, share, and mine data across groups of investigators. Over the past several years I have co-developed a NINDS funded cloud-based data platform, ://ieeg.org, giving me a central role in developing Big Data solutions for neural data, such as customized data sharing, large-scale cloud-based data analysis, and search and interrogation techniques for complex data and metadata. My scientific objectives for this project are: (1) to develop generalizable tools to curate, analyze, and interrogate multi-scale neural data, and (2) to create a platform that will galvanize a research community focused on sharing data, and methods to advance Big Data research in the basic and translational neurosciences. Equally important to this proposal, I present a training plan to prepare me for an academic career focused on Big Data in the neurosciences. This plan supplements my background in bioengineering and statistical modeling of neural data with broader data-science expertise in data integration and machine learning, and deeper domain knowledge of the clinical neurosciences. I have assembled a group of collaborators, basic investigators and clinician scientists, who will use the tools developed in this project to analyze and validate their data and methods. I will use the results of this project as the foundation for a R01 Grant application, in which I will expand the developed platform and tools to target other research domains (TBI, Emergency Care, Cardiac), as well as integrate other data-modalities such as Imaging, and Genomics. OMB No. 0925-0001/0002 (Rev. 08/12 Approved Through 8/31/2015) Page Continuation Format Page