CONNECTOME PROJECT DATA MANAGEMENT AND IMAGE ASSEMBLY WITH CLAY REID

这个子项目是许多利用资源的研究子项目之一 由NIH/NCRR资助的中心拨款提供。子项目的主要支持 而子项目的主要调查员可能是由其他来源提供的， 包括其它NIH来源。 列出的子项目总成本可能 代表子项目使用的中心基础设施的估计数量， 而不是由NCRR赠款提供给子项目或子项目工作人员的直接资金。 该项目的目标是获取，组装和分析的图像， 实验动物的脑回路，目的是了解 神经回路的结构和功能 我们最初的工作是 最大可能的连续切片透射电子显微镜（TEM） 哈佛半自动TEM捕获系统产生的图像集。 这些数据正被用于研究鼠标中电路的操作 视觉皮层和共配准与体内钙图像的 相同的标本，同时暴露于各种视觉刺激。 目前的哈佛仪器以每秒20兆像素的速度工作， 50 nm厚的矩形切片的成像集，最大为1200 × 750微米 以4 nm的面内分辨率在不到一个小时内完成。 这些数据包括 数以千计的相机帧，每个约10百万像素，必须对齐， 2D可生成高达56千兆比特的全平面图像。 飞机必须 在3D中对齐，以生成可分析的配准体积数据 比单独的平面或相机框架更容易。 前两 数据集，目前正在获得超过6个月的时间，是更大的 每个人都超过10个。 本阶段开发的方法将应用于 更快的下一代捕获系统生成的更大数据集 将在未来两年内建成。 由此产生的PetaByte数据将涵盖 体积高达1 mm 3，足以容纳整个皮质 从大脑表面延伸到白色物质中的柱状物。

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. Primary support for the subproject and the subproject's principal investigator may have been provided by other sources, including other NIH sources. The Total Cost listed for the subproject likely represents the estimated amount of Center infrastructure utilized by the subproject, not direct funding provided by the NCRR grant to the subproject or subproject staff. The goal of this project is to acquire, assemble, and analyze images of brain circuitry in experimental animals with the aim of understanding the structure and function of neural circuits. Our initial work is to align the largest possible serial section transmission electron microscopy (TEM) image sets produced by Harvard's semi-automated TEM capture system. These data are being used to study the operation of circuits in the mouse visual cortex and are co-registered with in-vivo calcium images of the same specimen while exposed to various visual stimuli. The current Harvard instrument, operating at ~20 Mpixels/sec, captures mosaic image sets of 50nm thick rectangular sections up to 1200 by 750 microns at 4 nm in-plane resolution in less than an hour each. These data consist of thousands of camera frames, ~10 Mpixels each, that must be aligned in 2D to produce full planar images up to 56 GBytes. The planes must then be aligned in 3D to produce registered volumetric data that can be analyzed much more easily than individual planes or camera frames. The first two datasets, currently being aquired over a period of ~6 months, are larger than 10 TBytes each. Methods developed at this stage will be applied to even larger datasets produced by a faster next generation capture system to be built during the next 2 years. The resulting PetaByte data will cover volumes up to 1mm3 which are sufficient to contain an entire cortical column extending from the brain surface into the white matter.