HIGH RES BRAIN ATLAS--DATA ACQUISITION AND USER ACCESS

高分辨率大脑图谱——数据采集和用户访问

• 批准号: 2703130
• 负责人: RICHARD L SIDMAN
• 金额: $ 21.06万
• 依托单位: HARVARD MEDICAL SCHOOL
• 依托单位国家: 美国
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-07-01 至 2000-04-30
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/2703130
• 关键词: brain imaging /visualization /scanning; brain mapping; computer program /software; confocal scanning microscopy; dissection; histology; laboratory mouse; magnetic resonance imaging; neuroanatomy; publications; sectioning

The goal of this project is to produce a high resoluation electronic atlas of the brain in three dimensions (3D). A number of technical problems are addressed, including how to minimize distortions during the processing of the brain through steps of hardening (fixation), embedding in a rigid support medium, sectioning and staining. Other problems relate to the massive computer database and memory requirements for the large number of images at the closely spaced intervals through the brain necessary for resolution of 1 um or less in all three planes of space (1 um isomorphic voxels). To solve these problems, we propose to make an atlas of the small brain of the mouse, as a prelude to tackling the human brain. We will begin at a voxel resolution of 10 um, and procede to 3 um and 1 um (greater than 550 GB of image data). Novel technical methods for whole brain embedding and sectionaing in epoxy resin and for confocal microscopy of stable 3D data space, and for partially automated volumetric segmentation of brains packed with hundreds of 3D structures, many with comples shapes. Images will be warped to a high resolution MRI standard of the intact fixed head (100 um digital "slices" obtained with a 9.4 Tesla magnet). The uniquely high level of histological resolution in 3D will allow us to develop a new "white matter anatomy" by tracing the exact trajectories of myelinated fiber groups that intermingle so intimately in the white matter that they have been impossible to sort out by previously available characterization of brain structures. Chemical, genetic, and pathological data will be superimposable digitally onto the basic cell and myelin atlas templates. For ease of data storage and transmission, the atlas images will be compressed with wavelet compression algorithms, with an added substraction method that will allow the origical focus on the elucidation of the genetic gases for individual differences in brain organization. Remote users with their own research, educational, and public welfare needs, will be able to obtain the entire database, including capabilities for rotations and digital "resectioning" in any desired orientation, on new high capacity DVD disks, or will gain access to specific 2D and 3D images via a dedicated WWW page outfitted with a systematic and friendly query interface and VRML-based browser. Users needing to download atlas images will access the WWW pageover new high speed Asynchronous Transfer mode transmission lines.

该项目的目标是生产高分辨率电子 三维 (3D) 大脑图谱。多项技术 问题得到了解决，包括如何最大限度地减少过程中的扭曲 通过硬化（固定）步骤对大脑进行处理， 包埋在刚性支持介质中、切片和染色。其他 问题与海量计算机数据库和内存有关 要求大量的图像在紧密的间隔 分辨率达到 1 微米或更小的大脑所需的间隔 所有三个空间平面（1 um 同构体素）。为了解决这些 问题，我们建议制作一个小鼠小大脑的图谱， 作为攻克人类大脑的前奏。我们将从体素开始 分辨率为10um，并进至3um和1um（大于550 GB 的图像数据）。全脑嵌入新技术方法 环氧树脂切片和稳定 3D 共焦显微镜 数据空间，以及部分自动化的体积分割 大脑中充满了数百个 3D 结构，其中许多具有复杂的结构 形状。图像将扭曲至高分辨率 MRI 标准 完整的固定头（用 9.4 特斯拉获得的 100 微米数字“切片” 磁铁）。 3D 中独特的高水平组织学分辨率将 让我们通过追踪精确的白质解剖学来开发新的“白质解剖学” 有髓纤维组的轨迹紧密地混合在一起 他们无法区分的白色物质 先前可用的大脑结构特征。化学、 遗传和病理数据将以数字方式叠加到 基本细胞和髓磷脂图谱模板。为了便于数据存储和 传输时，图集图像将被小波压缩 压缩算法，增加了减法方法 允许将最初的重点放在对遗传气体的阐明上 大脑组织的个体差异。远程用户及其 自己的研究、教育和公共福利需求，将能够 获取整个数据库，包括轮换功能和 在新的高容量上以任何所需方向进行数字“切除” DVD 光盘，或者将通过以下方式访问特定的 2D 和 3D 图像 专门的WWW页面配有系统且友好的查询 界面和基于 VRML 的浏览器。需要下载图集的用户 图像将通过新的高速异步访问 WWW 页面 传输模式传输线。