Enabling Shared Analysis and Processing of Large Neurophysiology Data

项目摘要/摘要 了解大脑功能是改善医疗保健和推进一系列科学举措的关键。 阿尔茨海默氏症、帕金森氏症和肌萎缩侧索硬化症等退行性脑部疾病的治疗正在成为 随着当前美国人口老龄化和预期寿命的增加，这一问题变得越来越重要。所需成本 阿尔茨海默氏症和其他痴呆症仅在2016年就估计超过2000亿美元，更不用说人类 这些疾病造成的破坏。自闭症、成瘾、抑郁、癫痫、创伤性脑损伤和疼痛 治疗只是与大脑功能相关的几个至关重要的健康问题。认知 科学也处于计算和自主系统研究的前沿，有可能 彻底改变人机交互，应对新出现的全球挑战。由于这些原因， 大脑的其他重要机会(通过推进创新的神经技术进行大脑研究) 总统倡议是为了改善国家未来的健康和竞争力而创建的， 加速大脑研究的根本目标。 为了实现这一目标，大脑研究界开发了无国界神经数据： 支持大规模协作和研究的神经生理学(NWB)文件格式和规范。这 开放格式创建于2014年，已经对基于细胞的神经生理学产生了影响， 像艾伦脑科学研究所这样的组织，生成和共享数据集 脑天文台和艾伦细胞类型数据库。尽管这项初步工作是有希望的，但在 由于缺乏易于浏览、处理、分析和可视化的软件工具，研究社区发展缓慢 NWB数据，同时促进可复制性。因此，这项工作的目的是生产这样的工具来支持大脑 通过支持和发展NWB社区，发起和实施其他大规模的脑研究项目。 这里提出的工作涉及基于细胞的神经生理学中的三个重要工作流程：o​ptic 在刺激下对神经元成像的生理学，从表面检测棘波事件的硅探针记录 大脑皮层下经深部​结构和​体外切片电生理记录时相​时变刺激 以及来自神经元​的电反应。将创建新的多规模软件工具，以实现高效 浏览、处理、分析和可视化基于NWB的大脑数据；将实验刺激与 观察到的反应。还将开发转换实用程序，将现有数据转换为NWB格式。作为 数据庞大、复杂，可能分布在许多站点上，软件工具将是基于Web的， 使研究人员能够以可复制的方式远程访问和处理数据，并使用可扩展的云 计算资源。该软件将在开源许可下发布，并将放置在 正式的软件流程，以促进整个研究社区的共享。这些工具将被构思并 在艾伦科学家的帮助下创建，他们还将使用这三个工作流程执行最终验证。

Project Summary / Abstract Understanding brain function is key to improving health care and advancing a number of scientific initiatives. The treatment of degenerative brain diseases such as Alzheimer's, Parkinson’s, and ALS is becoming increasingly important as the current US population ages and life expectancies increase. The costs of Alzheimer's and other dementias is estimated at over $200 billion in 2016 alone, not to mention the human devastation that these diseases incur. Autism, addiction, depression, epilepsy, traumatic brain injury, and pain treatment are just a few more of the critically important health concerns related to brain function. Cognitive science is also at the forefront of research into computational and autonomous systems, with the potential to revolutionize human computer interaction and tackle emerging global challenges. As a result of these and other significant opportunities the BRAIN (Brain Research through Advancing Innovative Neurotechnologies) Presidential initiative was created to improve the future health and competitiveness of the nation, with the fundamental goal of accelerating brain research. Consistent with this goal, the brain research community has developed the Neurodata Without Borders: Neurophysiology (NWB) file format and specification to support large-scale collaboration and research. This open format was created in 2014 and is already making an impact on cellular-based neurophysiology, with organizations such as the Allen Institute for Brain Science generating and sharing datasets such as the Allen Brain Observatory and the Allen Cell Types Database. Although this preliminary work is promising, progress in the research community is slowed by a lack of software tools to readily browse, process, analyse, and visualize NWB data, while promoting replicability. Thus this work aims to to produce such tools in support of the BRAIN initiative and other large-scale brain research programs by supporting and growing the NWB community. The work proposed here addresses three important workflows in cellular-based neurophysiology: o​ ptical physiology to image neurons under stimuli, silicon probe recordings to detect spike events from the surface of the cortex down through deeps​ structures, and​ in vitro slice electrophysiology to record t​ ime-varying stimulus and electrical response from a neuron​. Novel multiscale software tools will be created to enable efficient browsing, processing, analysis, and visualization of NWB-based brain data; linking experimental stimuli to observed responses. Conversion utilities will also be developed to convert existing data into NWB form. As the data is large, complex, and may be distributed across many sites, the software tools will be web-based, enabling researchers to remotely access and process data in a reproducible manner, and to use scalable cloud computing resources. The software will be released under open source licences and will be placed under formal software process to facilitate sharing across the research community. The tools will be conceived and created with the help of Allen scientists, who will also perform final validation using these three workflows.