Informatics Meta-Spaces for the Exploration of Human Neuroanatomy

用于探索人类神经解剖学的信息学元空间

• 批准号: 7937028
• 负责人: John Darrell Van Horn
• 金额: $ 50万
• 依托单位: UNIVERSITY OF CALIFORNIA LOS ANGELES
• 依托单位国家: 美国
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-09-30 至 2012-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7937028
• 关键词: Address; Advocate; American; Anatomy; Architecture; Archives; Area; Atlases; Automatic Data Processing; Behavior; Blast Cell; Brain; Brain Mapping; Brain imaging; Characteristics; Client; Clinical; Collection; Communities; Data; Data Set; Databases; Development; Diagnostic; Disease; Education; Elements; Environment; Genomics; Growth; Health; Human; Image; Imagery; Individual; Individual Differences; Informatics; Internet; Left; Link; Magnetic Resonance Imaging; Measurement; Measures; Mining; Neuroanatomy; Neurosciences; Neurosciences Research; Occupations; Outcome; Positioning Attribute; Process; Property; Protocols documentation; PubMed; Public Health; Relative (related person); Research; Research Personnel; Resources; Scanning; Scientist; Services; Shapes; Software Tools; Space Explorations; Structure; Surface; System; Technology; Text; TimeLine; base; bioimaging; brain shape; brain volume; cluster computing; computerized data processing; data mining; demographics; design; digital; empowered; insight; interest; neglect; neuroimaging; new technology; novel; regional difference; repository; statistics; tool; user-friendly; virtual; web page; web-enabled

DESCRIPTION (provided by applicant): This application addresses broad Challenge Area (06) Enabling Technologies and specific Challenge Topic, 06-MH-103: New Technologies for Neuroscience Research. This project will provide a powerful, content-driven approach to the identification of brains having similar geometry and shape, the clustering of neuroanatomically similar cases, and the interactive 3D visualization of the large collections contained in neuroimaging archives. Beginning with the example of the LONI Image Data Archive (IDA), we will design automated data processing meta-workflows that will decompose the thousands of whole brain MRI volumes into constituent 3D neuroanatomical regions. We will characterize the geometric properties of these regional parcellations, store these measurements, and systematically assess pair-wise regional "distances" between brains, and decompose the resulting similarity matrix using multidimensional scaling and related approaches. These processes will be automated to accommodate the continuous growth of the archive and be able to include content obtained from other neuroimaging archives as well. We will graphically represent the derived space of brain similarity via an interactive and freely available 3D browser. Finally, we will develop means for users to upload their own MR anatomical volumes for automated processing via this same process using a large grid computational architecture; decompositions of the uploaded data will be compared against the shape statistics derived from the previously processed archival data; content- based search results will be returned to users via the web in the form of a rank ordered list of brain volumes having similar neuroanatomical characteristics; hyperlinks to additional meta-data and online information, as well as a depiction of the position of their data with respect to other derived brain data using the interactive 3D browser. Meta-data concerning each object in the display will be easily available describing subject demographics, diagnostic group, scanning parameters, etc. This project does not seek to advocate or support the development of any new centralized neuroimaging database but will provide an unprecedented service to the neuroscience community for interacting with existing digital brain archives. The tools developed here will be capable of accommodating that of other neuroimaging repositories as well as user's local archives, thus having utility beyond a single data resource. We expect the outcomes of this project to draw considerable interest and excitement from the neuroimaging community in a similar manner to which BLAST has had for the genomics community. Following a two year development timeline, we anticipate that these informatics-based approaches and tool deliverables will be instrumental for researchers as part of their neuroscientific enterprise, helping to guide research directions, enhance education, and will provide significant new insights concerning large-scale neuroimaging repositories of health and disease. H The outcomes of this novel project for informatics and dynamic visualization to are expected to draw considerable interest and excitement from the neuroimaging community. This project will empower content-driven searches in a similar manner to BLAST has provided for genomics and will provide significant new insights concerning large-scale neuroimaging repositories of health and disease. Over a two-year period of development, during which a number of new American jobs will be created, this project will deliver a robust, content-driven informatics approach to the identification of brains having similar geometry and shape, the clustering of neuroanatomically similar cases, and the interactive 3D visualization of the large collections contained in neuroimaging archives.

描述(由申请人提供)：本申请涉及广泛的挑战领域(06)使能技术和具体的挑战主题，06-MH-103：神经科学研究的新技术。该项目将提供一种强大的、内容驱动的方法来识别具有相似几何和形状的大脑、神经解剖学相似病例的集群以及包含在神经成像档案中的大量集合的交互式3D可视化。从LONI图像数据档案(IDA)的例子开始，我们将设计自动化数据处理元工作流，将数千个全脑MRI体积分解为组成3D神经解剖区域。我们将表征这些区域分区的几何属性，存储这些测量结果，系统地评估大脑之间的成对区域“距离”，并使用多维缩放和相关方法分解得到的相似性矩阵。这些过程将是自动化的，以适应档案的持续增长，并能够包括从其他神经成像档案获得的内容。我们将通过交互式和免费可用的3D浏览器以图形方式表示大脑相似性的派生空间。最后，我们将开发一种方法，供用户通过使用大型网格计算架构的相同过程上传他们自己的MR解剖体积进行自动处理；上传数据的分解将与先前处理的档案数据得出的形状统计数据进行比较；基于内容的搜索结果将以具有类似神经解剖特征的大脑体积的排序列表的形式通过网络返回给用户；其他元数据和在线信息的超链接，以及使用交互式3D浏览器描绘他们的数据相对于其他派生的大脑数据的位置。关于显示器中每个物体的元数据将很容易获得，描述了受试者的人口统计、诊断组、扫描参数等。该项目并不寻求倡导或支持开发任何新的集中式神经成像数据库，而是将为神经科学界提供前所未有的服务，以便与现有的数字大脑档案进行互动。这里开发的工具将能够容纳其他神经影像存储库以及用户的本地档案，因此具有超越单一数据资源的实用价值。我们预计这个项目的结果将引起神经成像社区的相当大的兴趣和兴奋，就像BLAST对基因组学社区所做的那样。在两年的开发时间表之后，我们预计这些基于信息学的方法和工具交付成果将有助于研究人员作为其神经科学事业的一部分，帮助指导研究方向，加强教育，并将提供关于大规模神经成像健康和疾病存储库的重要新见解。H这一新的信息学和动态可视化项目的成果预计将引起神经成像界的极大兴趣和兴奋。该项目将以类似于BLAST为基因组学提供的方式支持内容驱动的搜索，并将提供关于大规模神经成像健康和疾病存储库的重要新见解。在为期两年的开发期间，该项目将为美国创造许多新的就业机会，该项目将提供一种强大的、内容驱动的信息学方法，用于识别具有相似几何和形状的大脑、神经解剖学相似病例的集群，以及包含在神经成像档案中的大量收藏的交互式3D可视化。