FMRI Dynamic Phantom for ImprovedDetection of Resting-­‐State Brain Networks

摘要 功能磁共振成像(FMRI)已迅速成为人类神经科学中的主导工具 研究，并准备成为精神病学和神经学领域的一项变革性技术 诊断。近年来，大规模(1.5亿-12亿美元)和长期(10-12年)的国际 投资(例如，NIDA青少年大脑认知发展研究，NIH人类连接组项目， 白宫大脑倡议、英国生物库、欧盟人脑项目)扩大了人类fMRI的研究范围 包括更快的脉冲序列和更复杂的分析工具、更高的(≥7T)场强、与 多尺度实验和建模，强调跨多台扫描仪/研究集成数据 网站。这一代功能磁共振成像研究超越了最初专注于激活的简单化模型 研究大脑中的连接、网络和动态的非线性电路。新的思维方式 正被应用于一些对我们的社会利益影响最大的领域，从上瘾，抑郁， 自闭症和脑损伤，到基于年龄的认知退化。然而，尽管功能磁共振研究戏剧性地 加速，产生这些发现所需的核磁共振机器的质量保证协议已经滞后 远远落在后面。因此，今天的神经成像中心通常使用过时的静态幻影方案，现在 不能针对与当前和新兴应用相关的质量控制问题。石溪 Dynamic Phantom就是为满足这种迫切需求而设计的。以第一代工作原型为基础构建 幻影(正在申请专利)，以及第二代原型的工程改进 旨在提高其耐用性和可靠性，这里我们重点介绍商业化的下一个合乎逻辑的步骤。 第一阶段专注于质量控制和建立市场附加值，通过展示我们的幻影的 动态保真度测量以直接和具体的方式提供数据质量的唯一信息性测量 对人类数据的解释的影响。纳入由我们的高级顾问提供的反馈 然后，第二阶段分三步走向商业化。首先，ALA科学仪器公司将 使工程适应批量生产。第二，石溪大学的学术团队和 马萨诸塞州总医院/哈佛医学院将开发算法，使用表征 扫描仪噪声以清除相关伪像的数据，用于单一对象(临床)应用程序以及 用于多站点研究的扫描仪之间的标准化。最后，硬件和软件将被集成 一款产品，将由15家神经成像领域的国际领先者进行现场测试。反馈 从这个组中，根据初始学习曲线和/或日常使用情况确定潜在的摩擦点 将在最终设计中实施，以确保我们最终制造的硬件， 软件和文档使体模操作起来尽可能舒适，并且在实际中尽可能有用。在… 至此，我们的设备将准备好进行商业分销。

Abstract Functional magnetic resonance imaging (fMRI) has rapidly become the dominant tool in human neuroscience research, and is poised to become a transformative technology in the areas of psychiatric and neurological diagnostics. In recent years, large-scale ($150M-$1.2B USD) and long-term (10-12 year) international investments (e.g., NIDA Adolescent Brain Cognitive Development Study, NIH Human Connectome Project, White House BRAIN Initiative, UK Biobank, EU Human Brain Project) have expanded the reach of human fMRI to include faster pulse sequences and more complex analytic tools, higher (≥7T) field strength, integration with multi-scale experiments and modeling, and an emphasis on integration of data across multiple scanner/study sites. This generation of fMRI studies goes beyond the original simplistic models that focused upon “activation maps,” to investigate connections, networks, and dynamic nonlinear circuits in the brain. New ways of thinking are being applied to some of our highest-impact areas of societal interest, ranging from addiction, depression, autism, and brain injury, to age-based cognitive degeneration. However, while fMRI research dramatically accelerates, quality assurance protocols for the MRI machines needed to generate these findings have lagged far behind. Thus, today's neuroimaging centers typically use outdated static phantom protocols that are now incapable of targeting quality control issues relevant to current and emerging applications. The Stony Brook Dynamic Phantom is designed to address this urgent need. Building upon a 1st generation working prototype of the phantom (patent pending), as well as engineering improvements in the 2nd generation prototype designed to increase its durability and reliability, here we focus on the next logical steps to commercialization. Phase I focuses on quality control and establishing added value to the market, by showing that our phantom's measure of dynamic fidelity provides a uniquely informative measure of data quality with direct and concrete implications for the interpretation of human data. Incorporating feedback provided by our Senior Advisory Board, Phase II then proceeds towards commercialization in three steps. First, ALA Scientific Instruments will adapt the engineering for mass production. Second, the academic teams at Stony Brook University and Massachusetts General Hospital/Harvard Medical School will develop algorithms that use characterization of scanner noise to clean data of associated artifact, for both single-subject (clinical) applications as well as normalization across scanners for multi-site research. Finally, hardware and software will be then integrated into one product, which will be field-tested by 15 international leaders in the neuroimaging field. Feedback from this group, identifying potential friction points in terms of the initial learning curve and/or day-to-day usage of the phantom, will be implemented in the final design, to ensure that our final manufacturing for the hardware, software, and documentation make the phantom as pleasant to operate and practically useful as possible. At this point, our device will be ready for commercial distribution.