Morphometric Analysis of Brains and Neurons

描述（由申请人提供）：现代生物医学成像技术允许生成许多微观生物对象的高分辨率数字3D图像。然而，有效的策略仍有待开发的自动定量和统计分析，这样的微观结构的三维形态。该建议旨在开发用于完整大脑中神经元的自动形态测量分析的算法。单个神经元的空间位置及其3D投影模式的定量和统计表征不仅对于理解具有单细胞分辨率的大脑的复杂性，多样性和可塑性至关重要，而且对于阐明各种神经/精神/心理障碍的微妙细胞病理机制也至关重要。将探索新的专业知识，以推进生物医学成像多个领域的技术，如图像计算和复杂组织的模拟。一个GAL 4-独立的二元转录系统已经被开发出来，专门标记整个形态的果蝇嗅觉学习和记忆中心，蘑菇体（MBs）。结合MARCM（具有可抑制细胞标记的镶嵌分析）技术，可以独立地标记同一大脑中的各种单个MB神经元和整个MB。与此同时，新的算法已经发展到进行自动变形（形态变形和匹配）的不规则形状的3D对象。虚拟平均MB将经由多个“标准”MB之间的成对变形的统计表征来构建。然后，不同MB的形态测定分析和单个MB神经元的空间映射将涉及在感兴趣的MB或MB（其中特定的单个MB神经元被差异标记）与统计模型MB之间建立点对点对应。因此，可以能够自动检测和定量描述任何给定MB的结构偏差，并基于其3D神经元位置/投影模式来识别单个MB神经元。

DESCRIPTION (provided by applicant): Modern biomedical imaging technologies allow for generation of high-resolution digital 3D images of many microscopic biological objects. However, effective strategies remain to be developed for automatic quantitative and statistical analysis of such microscopic structures' 3D morphologies. This proposal is aimed at developing algorithms for automatic morphometric analysis of neurons in intact brains. Quantitative and statistical characterization of individual neurons' spatial locations and their 3D projection patterns is not only essential for understanding brains' complexity, diversity, and plasticity with single-cell resolution, but also critical for elucidating subtle cellular pathological mechanisms underlying various neurological/mental/psychological disorders. New expertise will be explored to advance technologies in multiple areas of biomedical imaging, such as image computation and simulations of complex tissues. A GAL4-independent binary transcriptional system has been developed to label specifically the entire morphologies of the Drosophila olfactory learning and memory center, the mushroom bodies (MBs). In conjunction with MARCM (Mosaic Analysis with a Repressible Cell Marker) technologies, one can independently label various single MB neurons and the whole MBs in the same brains. Meanwhile, new algorithms have been developing to conduct automatic morphing (morphological deformation & matching) of irregular-shaped 3D objects. A virtual average MB will be constructed via statistical characterization of pair-wise morphing among multiple "standard" MBs. Morphometric analysis of distinct MBs and spatial mapping of individual MB neurons will then involve establishing point-to-point correspondence between the MBs of interest or the MBs, in which specific single MB neurons are differentially labeled, and the statistical model MB. Thus, one may be able to detect automatically and describe quantitatively any given MB's structural deviations and to identify individual MB neurons based on their 3D neuronal location/projection patterns.