Intelligent connectomic analysis tool for dense neuronal circuits

用于密集神经元回路的智能连接组分析工具

• 批准号: 10019731
• 负责人: Shih-Jong J Lee
• 金额: $ 33.04万
• 依托单位: DRVISION TECHNOLOGIES, LLC
• 依托单位国家: 美国
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2020-12-18
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10019731
• 关键词: 3-Dimensional; Active Learning; Biological Models; Brain; Brain Diseases; Classification; Complex; Consumption; Data; Data Set; Databases; Dendrites; Detection; Disease; Electron Microscopy; Evaluation; Failure; Feedback; Fluorescence; Fluorescence Microscopy; Generations; Goals; Health; Human; Image; Infrastructure; Intelligence; Intelligence Tests; Joints; Machine Learning; Maps; Modeling; Mus; Nervous system structure; Neurons; Performance; Pharmaceutical Preparations; Phase; Phenotype; Process; Resolution; Retina; Structure; Synapses; Synaptic Vesicles; Testing; Time; Tissues; Training; Update; Validation; Vertebral column; Zebrafish; annotation system; automated analysis; brain disorder therapy; cell type; commercialization; data exchange; experimental study; fluorescence imaging; innovation; insight; interest; light microscopy; microscopic imaging; molecular marker; nervous system disorder; neuronal circuitry; next generation; novel diagnostics; novel therapeutics; prototype; reconstruction; terabyte; tool; usability

Intelligent Connectomic Analysis Tool for Dense Neuronal Circuits Project Summary: The lack of basic understanding of neuronal functions and disease processes is a big factor of failures in creating drugs for neurological diseases. High-resolution maps of the complex connectivity of neuronal circuits correlating with functional and/or molecular markers offer invaluable insights into the functional organization of the neuronal structures, which is a key to understanding the brain in health and disease. There is a strong interest in elucidating and quantifying the connectomics of brain networks with subcellular resolution using electron microscopy (EM) and correlate with functional fluorescence microscopy data. The ultimate goal is to elucidate human brain functions and the mechanisms of human brain disorders. This is critically important to enable new diagnostics and therapies for brain disorders. The reconstruction and analyses of neuronal networks is challenging in part due to the joint requirement of large volume and high resolution and a large gap in connectomic analysis solutions. There is a strong need for next generation, well supported, integrated, easy to use and highly automated analysis tools to detect and classify neurons, trace arbor branches, identify synapses, spines and synaptic vesicles that increase the throughput of otherwise prohibitively time-consuming analyses in connectomic experiments. There is also a strong need for tools to perform downstream data-driven analysis such as functional inference from structure and phenotypic discovery. Powered by machine learning and DRVision innovations and collaborating with Dr. Rachel Wong and 9 additional labs, this project proposes to create an intelligent connectomic analysis (ICA) tool optimized for dense neuronal circuits. The tool will be commercially supported and integrated with DRVision’s flagship product Aivia to (1) provide accurate and automated neuron tracing in 3D EM and 3D fluorescence data up to multi-terabytes, (2) identify pre- and post-synaptic dendrite segments, (3) correlate light and electron microscopy data, quantify and classify neurons and sub-cellular components, (4) extract and analyze neuron circuits, (5) provide tools for phenotype discoveries, (6) seamlessly integrate the pipeline of ground truth (GT) annotation, editing, and machine learning workflow, and (7) access the required computing infrastructure, database connection, and exchange of data with other tools.

密集神经元回路的智能连接组分析工具 项目概要： 缺乏对神经元功能和疾病过程的基本了解是治疗失败的一个重要因素。 制造治疗神经系统疾病的药物神经元复杂连接的高分辨率图 与功能和/或分子标记相关的电路提供了对功能的宝贵见解， 神经元结构的组织，这是了解大脑健康和疾病的关键。 有一个强烈的兴趣，阐明和量化的脑网络的连接与亚细胞 使用电子显微镜（EM）进行分辨率分析，并与功能荧光显微镜数据相关联。的 最终目的是阐明人脑功能和人脑疾病的机制。这是 这对大脑疾病的新诊断和治疗至关重要。 神经元网络的重建和分析是具有挑战性的，部分原因是由于联合 大体积和高分辨率的要求以及在连接组分析解决方案中的大差距。有 对下一代、支持良好、集成、易于使用和高度自动化的分析工具的强烈需求 检测和分类神经元，追踪乔木分支，识别突触，棘和突触囊泡， 增加连接组实验中原本极其耗时的分析的吞吐量。 还强烈需要工具来执行下游数据驱动的分析，如函数推理 从结构和表型发现。 由机器学习和DRVision创新提供支持，并与Rachel Wong博士合作， 9个额外的实验室，该项目建议创建一个智能连接组分析（伊卡）工具 优化了密集的神经元回路。该工具将得到商业支持，并与 DRVision的旗舰产品Aivia（1）在3D EM和3D中提供准确和自动化的神经元追踪 高达数TB的荧光数据，（2）识别突触前和突触后树突片段，（3） 光学和电子显微镜数据，定量和分类神经元和亚细胞成分，（4）提取 并分析神经元电路，（5）为表型发现提供工具，（6）无缝集成管道 地面实况（GT）注释，编辑和机器学习工作流程，以及（7）访问所需的 计算基础设施、数据库连接以及与其他工具的数据交换。