Harnessing big-data for plasticity and rehabilitation in translational SCI

脊髓损伤和紊乱(SCI/D)是影响退伍军人的重大健康问题，其发生率更高 而不是平民百姓。年，SCI/D的总经济负担估计为90亿美元/年至4000亿美元 终生医疗和生产力损失成本。最常见的临床表现是颈椎高SCI/D 这会产生广泛的问题，包括手功能丧失，自主性，感觉变化， 痉挛、疼痛和自主神经功能障碍，深刻影响生活质量。恢复这些功能是 再生和康复治疗脊髓损伤/脊髓损伤的目标 脊髓损伤联合会(VA-GMSCIC)是由VA资助的一项努力，旨在开发脊髓损伤的晚期翻译疗法 非人灵长类动物(NHP)模型，为临床测试紧急治疗方法做准备。之前 目前的资金主要集中在通过5个不同的中心收集每个主题的多方面数据 协作收集数据，每个数据都在其特定的专业领域(生理学、行为学、组织学、 神经康复和分子生物学)。这是NHP模型的理想用法，因为最大信息量是 收集了有关治疗方法在少数NHP中的表现的资料。来自于此的数据 重要模型的经典特征是大数据的3V：大容量(大图像)、高多样性 (多模式数据)和高速(机器人康复；生理学；神经调节)，这两者都是一个挑战 以及新发现的机会。现代数据科学工具的应用可以帮助兑现承诺 治疗SCI/D的翻译精准医学。正如我们先前的工作所证明的，VA的有效管理 NHP大数据使我们能够有效地利用VA-GMSCIC数据来推动新发现。然而， 集成这些NHP大数据需要持续以数据为驱动的机器人康复、运动学、组织学、 和医疗信息。提取有意义的发现需要大量的计算工作。这个 拟议续订的目的是在我们不断取得成功的基础上，为 集成新型高分辨率数据支持新药安全性/有效性研究的VA-GMSCIC 治疗学。我们的数据科学团队完全有能力实现这一目标。我们的团队已经提供了分析 支持VA-GMSCIC，帮助集成来自UCSD、UCLA、UCI、UC Davis和UCSF的数据以进行测试 13年以上的脊髓损伤康复和再生治疗。我们支持开发不同的 损伤模型、行为评估、电生理学、运动学测量和治疗方法 100多名受试者。在我们目前的功勋奖(截至2020年11月)下，我们的团队建立在这样的历史背景下 建立可实现快速结构化数据的功能性灵长类数据共享(PDC-SCI)基础架构 跨VA-GMSCIC站点的共享、数据集成和分析支持。该项目帮助退伍军人管理局- GMSCIC从专注的发现项目发展到后期的翻译研究， 大规模收集大数据。我们的目标是扩大我们的知识发现渠道，以满足以下关键需求 翻译的SCI/D大数据，以支持计划的安全性/有效性研究。具体地说，更新将建立在 我们的成功并通过支持整合以下翻译电子产品来扩大我们的工作范围：目标1 健康记录(TEHR)，AIM 2)先进的机器人康复数据，AIM 3)来自大脑的神经调制数据- 机器接口，以及AIM 4)先进的机器学习分析管道，用于快速集成 多维数据。其目标是帮助VA-GMSCIC有效地测试重要的候选治疗药物 翻译到人类，同时促进现代数据管理，遵守联邦批准的博览会 (可查找、可访问、可互操作和可重用)数据共享原则。这将确保现有的 通过数字技术最大限度地利用VA在数据收集方面的投资，以实现 从SCI/D这一有价值的NHP模型中发现知识。

Spinal cord injury and disorders (SCI/D) are substantial health concerns impacting veterans at a higher rate than the civilian population. The total economic burden of SCI/D is estimated at $9 billion/year to $400 billion in lifetime medical and loss-of-productivity costs. The most common clinical presentation is high cervical SCI/D which produces a broad spectrum of issues, including loss of hand function, autonomy, sensory changes, spasticity, pain and autonomic dysfunction, profoundly impacting quality of life. Restoring these functions is the goal of regenerative and rehabilitative therapeutic approaches for SCI/D. The VA Gordon Mansfield Spinal Cord Injury Consortium (VA-GMSCIC) is a VA-funded effort to develop late-stage translational therapeutics in a nonhuman primate (NHP) model in preparation for testing emergent therapeutic approaches clinically. Prior and current funding has focused on multifaceted data collection on each subject with 5 different centers collaboratively collecting data, each within their specific domain of expertise (physiology, behavior, histology, neurorehabilitation, and molecular biology). This is an ideal use of the NHP model, as maximal information is collected about the performance of therapeutic approaches in a small number of NHPs. Data from this important model is characterized by the classic ‘3Vs of Big Data’: high volume (large images), high variety (multi-modal data), and high velocity (robotic rehab; physiology; neuromodulation), providing both a challenge and opportunity for novel discoveries. Application of modern data science tools can help deliver on the promise of translational precision medicine for SCI/D. As our prior work demonstrates, effective management of VA NHP big data enables us to effectively harness VA-GMSCIC data to drive new discoveries. However, integrating these NHP big data requires ongoing data-driven integration of robotic rehab, kinematics, histology, and medical information. Extraction of meaningful discoveries requires extensive computational work. The purpose of the proposed renewal is to build on our ongoing success in assembling a data commons for the VA-GMSCIC by integrating new types of high-resolution data in support of safety/efficacy studies of novel therapeutics. Our data science team is well positioned to achieve this goal. Our team has provided analytical support for the VA-GMSCIC, helping to integrate data from UCSD, UCLA, UCI, UC Davis and UCSF for testing SCI rehab and regenerative therapies in NHPs for over 13 years. We have supported development of different injury models, behavioral assessments, electrophysiology, kinematic measures, and therapeutic approaches in 100+ subjects. Under our current merit award (ending Nov 2020), our team built on this historical background to establish a functional primate data commons (PDC-SCI) infrastructure that enables rapid, structured data sharing, data integration, and analytics support across the VA-GMSCIC sites. The project has helped the VA- GMSCIC evolve from focused discovery projects to late-stage translational studies with highly-sophisticated, large-scale “big-data” collection. We aim to expand our knowledge-discovery pipeline for these critical translational SCI/D big data to support planned safety/efficacy studies. Specifically, the renewal will build on our successes and expand the scope of our work by supporting integration of: Aim 1) translational electronic health records (tEHR), Aim 2) advanced robotic rehabilitation data, Aim 3) neuromodulation data from brain- machine interfaces, and Aim 4) advanced machine learning analytical pipelines for rapidly integrating multidimensional data. The goal is to help VA-GMSCIC efficiently test important therapeutic candidates for translation to humans while promoting modern data stewardship adhering to the federally-endorsed FAIR (Findable, Accessible, Interoperable, and Reusable) data sharing principles. This will ensure that the existing VA investment in data collection is leveraged to the maximal extent through digital technologies for enduring knowledge-discovery from this valuable NHP model of SCI/D.