Functional MRI Core Facility

功能性核磁共振核心设施

• 批准号: 9152153
• 负责人: Peter Bandettini
• 金额: $ 475.4万
• 依托单位: NATIONAL INSTITUTE OF MENTAL HEALTH
• 依托单位国家: 美国
• 资助国家: 美国
• 起止时间: 至
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/9152153
• 关键词: Address; Administrator; Algorithms; Anatomy; Archives; Back; Basic Science; Boxing; Brain; Clinical Sciences; Clip; Collaborations; Communication; Communities; Computer software; Computers; Consult; Core Facility; Custom; Data; Data Set; Dependence; Development; Devices; Education; Electroencephalography; Electronics; Elements; Ensure; Eye; Floor; Focus Groups; Functional Magnetic Resonance Imaging; Galvanic Skin Response; Goals; Hour; Image; Individual; Information Technology; Institutes; Joystick; Laboratories; Lamivudine; Libraries; Link; MRI Scans; Magnetic Resonance Imaging; Maintenance; Maps; Methods; Mission; Modeling; Motion; Movement; Nasal bone structure; National Heart, Lung, and Blood Institute; National Institute of Mental Health; National Institute of Neurological Disorders and Stroke; National Institute on Alcohol Abuse and Alcoholism; National Institute on Deafness and Other Communication Disorders; Noise; Nuclear Magnetic Resonance; Optics; Outcome; Output; Paper; Participant; Patients; Pattern; Peer Review; Phase; Physiologic pulse; Play; Policies; Positioning Attribute; Principal Investigator; Procedures; Process; Production; Protocols documentation; Publications; Pythons; Relative (related person); Reporting; Research; Research Infrastructure; Research Personnel; Research Support; Resources; Role; Running; Scanning; Schedule; Science; Scientist; Series; Services; Signal Transduction; Site; Skin; Solutions; Specialist; Speed; System; Techniques; Technology; Testing; Time; Tooth structure; Training; Translating; United States National Institutes of Health; Update; Variant; Vendor; Work; Writing; base; blood oxygen level dependent; comparative efficacy; data format; data sharing; experience; feeding; image reconstruction; improved; independent component analysis; indexing; innovation; meetings; new technology; open source; operation; optimism; prisma; programs; prospective; radio frequency; reconstruction; research study; software development; square foot; symposium; tool; virtual; web site

Space Utilization: The Functional MRI Facility (FMRIF) currently occupies approximately 4800 sq ft of space, divided between the scanner bays, control rooms and electronics/machine rooms for 3TA, 3TB, 3TC, 3TD, and 7T MRI scanners located within the NMR center and office space on the second floor above the NMR center in the FMRIF/SFIM suite (approximately 1400 sq ft total, including shared conference space). The FMRIF staff (currently 14 full-time positions) consists of: the facility director, three staff scientists to keep the scanners running, six MRI technologists, an information technology specialist, a programmer, a technical laboratory manager, and an administrative laboratory manger. The functional MRI facility supports the research of over 30 Principal Investigators translating to over 300 researchers overall. Over 70 research protocols are active and making use of FMRIF scanners. Each scanner has scheduled operating hours of 105 hours per week. Since its inception in 2000 until July 2015, a total of 999 peer-reviewed publications from intramural investigators have used data acquired in the FMRIF core facility. The total is distributed among 672 papers from NIMH, 234 papers from NINDS, and 93 from the other institutes. These papers have been cited a total of 81,899 times for a combined h-index of 141. In other words, 141 papers using the FMRIF have been cited at least 141 times. A full listing of all of these publications is available at this link: https://fmrif.nimh.nih.gov/public/FMRIF_all_Aug2015.xlsx/at_download/file Projects: Multi-echo EPI: An ongoing collaboration with Dr. Bandettini's SFIM has been in the development and implementation of a multi-echo EPI acquisition and analysis pipeline in order to minimize non- blood oxygen level dependent (BOLD) and therefore, artifactual, signal changes. The basic concept is that fractional BOLD signal changes increase linearly with echo time. With multi-echo acquisition, the TE-dependence of the signal can be assessed at each TR. The basic procedure is to first perform Independent Component Analysis (ICA) on a time series of a composite concatenated image made up of the three echoes obtained with each TR, and to then determine the echo time dependence of each ICA component. Those ICA components that do not show linear TE dependence are removed. Centralized image reconstruction: Dr. Roopchansingh, Dr. Inati, and Joe Naegele as well as Dr. Michael Hansen (NHLBI) worked to develop and extend the Gadgetron framework. Briefly, the "Gadgetron" is a stand-alone recon engine that can receive raw data from a growing number of different datasets across scanner and vendor, and then pass it back to the scanner (or elsewhere) in the appropriate format. One of the key outcomes of this project has been to provide converters for most major vendors' custom MRI raw data formats into a new standard ISMRMRD format, which allows reconstruction algorithms to be more easily shared and used. Joe Naegele developed components for the Gadgetron reconstruction framework to allow production of DICOM images from custom MRI scanner pulse sequences. This system is currently in use on the GE scanner platforms, specifically for Dr. Roopchansingh's custom B0 mapping sequence. Joe Naegele also developed a dynamic Python interface to the Gadgetron to allow writing reconstruction chains using Python tools. Joe Naegele played a critical role in developing solutions for continuous building, testing, and deployment of the Gadgetron, ISMRMRD and other related software projects. Motion Correction: In collaboration with GE, Dr. Roopchansingh has been working with scientists in the NMRF to evaluate the efficacy and applicability of a GE product "PROMO" that performs prospective motion correction. This has been implemented primarily to improve the quality of 3D anatomical data acquired from patients and subjects prone to motion. One of the initial projects compared the efficacy of PROMO in improving the consistency of metrics output by FreeSurfer. Dr. Roopchansingh also modified the sequence provided by GE to allow researchers to collect motion-compensated multi-echo MP-RAGE data. More recently, he worked with Dr. Basser's Section to provide motion-compensated MP2-RAGE capability using PROMO. The FMRIF Siemens Skyra (3TD) scanner has a KinetiCor Prospective Motion Correction (PMC) accessory. The KinetiCor system is a commercially available device that employs in-bore camera tracking with an external computer to identify subject motion modeled as a rigid body. The motion tracking data is then fed back to the MRI system and used to prospectively track the imaging scan volume with subject motion. The KinetiCor device is based around the Metria Innovation Moir Phase Tracking system that employs a single 12mm x 12mm marker with a Moir fringe pattern that changes with all 6 translational and rotational aspects of a rigid body transformation. The ability of the device to obtain the full rigid body motion from a single camera and marker together with favorable reports in early scientific publications generated optimism that the device would offer a straightforward solution to problems with patient motion during long imaging scans (3D anatomical and fMRI). Although the KinetiCor device works well during tests with phantoms, preliminary results employing the device with direct skin attachment to the nasion or nasal bone have been mixed with the Siemens 20-channel coil and poor with the 32-channel. As a result of our experiences and in collaboration with other groups (Maastricht, UCL, and others), we have concluded that directly mounting the marker on the skin is unreliable due to problems of skin movement relative to the brain, possible bumping of the marker against the coil elements and the likelihood of optical occlusion by the camera. Therefore, we have decided to implement a tooth-clip solution similar to that being used for motion logging at UCL. The in console-room manufacture of a patient-specific tooth-clip adds approximately 30 minutes additional time to an exam. While the FMRIF recognizes that an orally attached device will not be suitable for all patients, it will almost certainly help some research groups, and appears to be the most reliable way to use the KinetiCor device at present. Additionally, we have been in communication with KinetiCor to obtain the latest software for the camera system and MRI scanner, including the latest version of the XPace libraries for the scanner that offers the ability to simultaneously track up to 3 markers. It is hoped that this redundancy may provide some robustness to problems with optical occlusion of the markers due to the coil elements. Finally, we are programming and compiling KinetiCor aware pulse sequences locally which should allow PMC to be enabled with specialized pulse sequences in addition to the standard variants supplied with the KinetiCor device. Project sharing: Joe Naegele has deployed a project management system (Redmine) for project task tracking and collaborative science efforts. Joe maintains the FMRIF Github organization, used for collaborating on both open source and private software projects. He also created a project management system using Python for sharing data on POSIX file systems, which has greatly improved the ability of the FMRIF to collaborate scientifically and has made software development and testing efforts more efficient. Virtual Desktops: In 2014 our technologists needed a workstation solution that would allow them an uninterrupted desktop session while moving between scanner rooms within the MRI facility. Roark Maccado implemented a Virtual Desktop Infrastructure solution to replace our technologists' and administrators' MACs/Windows desktops.

空间利用率： 功能性 MRI 设施 (FMRIF) 目前占用约 4800 平方英尺的空间，分为位于 NMR 中心内的 3TA、3TB、3TC、3TD 和 7T MRI 扫描仪的扫描仪隔间、控制室和电子/机器房以及 FMRIF/SFIM 套件中 NMR 中心上方二楼的办公空间（总共约 1400 平方英尺，包括 共享会议空间）。 FMRIF 工作人员（目前有 14 个全职职位）包括：设施主任、三名负责保持扫描仪运行的科学家、六名 MRI 技术专家、一名信息技术专家、一名程序员、一名技术实验室经理和一名行政实验室经理。 功能性 MRI 设施支持 30 多名首席研究员的研究，即总共 300 多名研究人员的研究。超过 70 个研究方案正在使用 FMRIF 扫描仪。每台扫描仪的预定运行时间为每周 105 小时。 自 2000 年成立以来至 2015 年 7 月，校内研究人员总共发表了 999 篇经过同行评审的出版物，使用了 FMRIF 核心设施获取的数据。总共分布在 NIMH 的 672 篇论文、NINDS 的 234 篇论文和其他机构的 93 篇论文中。这些论文总共被引用了 81,899 次，h 指数合计为 141。换句话说，使用 FMRIF 的 141 篇论文至少被引用了 141 次。 所有这些出版物的完整列表可通过以下链接获取：https://fmrif.nimh.nih.gov/public/FMRIF_all_Aug2015.xlsx/at_download/file 项目： 多回波 EPI： 与 Bandettini 博士的 SFIM 持续合作，开发和实施多回波 EPI 采集和分析流程，以最大限度地减少非血氧水平依赖性 (BOLD) 以及因此产生的人为信号变化。基本概念是分数 BOLD 信号变化随回波时间线性增加。通过多回波采集，可以在每个 TR 处评估信号的 TE 依赖性。基本过程是首先对由每个 TR 获得的三个回波组成的复合串联图像的时间序列执行独立分量分析 (ICA)，然后确定每个 ICA 分量的回波时间依赖性。那些不表现出线性 TE 依赖性的 ICA 分量被删除。 集中图像重建： Roopchansingh 博士、Inati 博士、Joe Naegele 以及 Michael Hansen 博士 (NHLBI) 致力于开发和扩展 Gadgetron 框架。 简而言之，“Gadgetron”是一个独立的侦察引擎，可以从扫描仪和供应商中越来越多的不同数据集接收原始数据，然后以适当的格式将其传回扫描仪（或其他地方）。该项目的主要成果之一是为大多数主要供应商的定制 MRI 原始数据格式提供转换器，将其转换为新的标准 ISMRMRD 格式，从而使重建算法更容易共享和使用。 Joe Naegele 为 Gadgetron 重建框架开发了组件，以允许从定制 MRI 扫描仪脉冲序列生成 DICOM 图像。该系统目前在 GE 扫描仪平台上使用，专门用于 Roopchansingh 博士的定制 B0 映射序列。 Joe Naegele 还开发了 Gadgetron 的动态 Python 接口，以允许使用 Python 工具编写重建链。 Joe Naegele 在开发持续构建、测试和部署 Gadgetron、ISMRRMRD 及其他相关软件项目的解决方案方面发挥了关键作用。 运动校正： Roopchansingh 博士与 GE 合作，一直与 NMRF 的科学家合作，评估执行前瞻性运动校正的 GE 产品“PROMO”的功效和适用性。实施这一点主要是为了提高从易于运动的患者和受试者处获取的 3D 解剖数据的质量。 最初的项目之一比较了 PROMO 在提高 FreeSurfer 指标输出的一致性方面的功效。 Roopchansingh 博士还修改了 GE 提供的序列，使研究人员能够收集运动补偿的多回波 MP-RAGE 数据。最近，他与 Basser 博士的部门合作，使用 PROMO 提供运动补偿 MP2-RAGE 功能。 FMRIF 西门子 Skyra (3TD) 扫描仪具有 KinetiCor 前瞻性运动校正 (PMC) 附件。 KinetiCor 系统是一款商用设备，它采用内孔摄像头跟踪和外部计算机来识别建模为刚体的对象运动。 然后，运动跟踪数据被反馈到 MRI 系统，并用于前瞻性地跟踪对象运动的成像扫描体积。 KinetiCor 设备基于 Metria Innovation 莫尔相位跟踪系统，该系统采用单个 12mm x 12mm 标记，其莫尔条纹图案随刚体变换的所有 6 个平移和旋转方面而变化。 该设备能够通过单个摄像头和标记获得完整的刚体运动，再加上早期科学出版物中的好评报告，人们乐观地认为该设备将为长时间成像扫描（3D 解剖和功能磁共振成像）期间患者运动问题提供直接的解决方案。 尽管 KinetiCor 设备在模型测试中工作良好，但使用直接将皮肤附着在鼻根或鼻骨上的设备的初步结果与西门子 20 通道线圈混合在一起，而与 32 通道线圈的效果较差。 根据我们的经验以及与其他小组（马斯特里赫特、伦敦大学学院等）的合作，我们得出的结论是，由于皮肤相对于大脑的运动问题、标记可能与线圈元件碰撞以及相机光学遮挡的可能性，直接将标记安装在皮肤上是不可靠的。 因此，我们决定实施一种类似于伦敦大学学院运动记录所用的齿夹解决方案。在控制台室中制造患者专用牙夹，使检查时间增加了约 30 分钟。 虽然 FMRIF 认识到口腔附着装置并不适合所有患者，但它几乎肯定会对一些研究小组有所帮助，并且似乎是目前使用 KinetiCor 装置的最可靠方法。 此外，我们一直在与 KinetiCor 沟通，以获得用于摄像头系统和 MRI 扫描仪的最新软件，包括用于扫描仪的最新版本 XPace 库，该库能够同时跟踪最多 3 个标记。 希望这种冗余可以为由于线圈元件而导致的标记的光学遮挡问题提供一定的鲁棒性。最后，我们正在本地编程和编译 KinetiCor 感知脉冲序列，除了 KinetiCor 器件提供的标准变体之外，这还允许使用专门的脉冲序列启用 PMC。 项目分享： Joe Naegele 部署了一个项目管理系统 (Redmine)，用于项目任务跟踪和协作科学工作。 Joe 维护着 FMRIF Github 组织，用于在开源和私有软件项目上进行协作。他还使用Python创建了一个项目管理系统，用于在POSIX文件系统上共享数据，这极大地提高了FMRIF的科学协作能力，使软件开发和测试工作更加高效。 虚拟桌面： 2014 年，我们的技术人员需要一种工作站解决方案，使他们能够在 MRI 设施内的扫描仪室之间移动时进行不间断的桌面会话。 Roark Maccado 实施了虚拟桌面基础设施解决方案来取代我们技术人员和管理员的 MAC/Windows 桌面。