Cortical Reconstruction and Analysis Software

这项提议的目标是开发基本上自动化和用户友好的软件工具，用于查找和 在体积磁共振(MR)图像中数学地表示大脑皮层。这些 软件工具将公开和免费提供，并将在涉及1)的研究中有用 开发正常和病变皮质几何结构的描述，2)标志性生成 可变形图谱配准，3)结构/功能关系的统计分析，4) 衰老和疾病中形态变化的特征。将形成以下基础的算法 所开发的软件工具来源于我们自己在大脑皮层重建和分析方面的研究工作。我们 提出五个具体目标：1)将核心的体积处理算法纳入MIPAV，a) NIH开发的商业质量但免费提供的软件包；2)实现专用的图形用户界面- 基于软件包，用于从磁共振图像执行大脑皮层的完整重建；3) 在我们的专用软件包中实施基于表面的处理和分析工具，从而 皮质上感兴趣区域的定义、皮质展开、表面编辑和脑沟标记；4) 对软件包执行两个阶段的测试，然后进行全面公开发布；以及5) 实施增强的脚本和基于客户端/服务器的软件版本，允许进行定制 的整体处理方法，并将计算密集型算法卸载到联网的 服务器。这些目标将通过1)将现有的内部软件移植到Java编程中来实现 使用MIPAV应用程序编程接口的语言，2)开发图形用户界面 可视化和验证，以及3)根据测试人员的反馈进行改进。所有人都可以获得源代码 开发的软件将公之于众。

The goal of this proposal is to develop largely automated and user-friendly software tools for finding and mathematically representing the cerebral cortex in volumetric magnetic resonance (MR) images. These software tools will be made publicly and freely available and will be useful in studies involving 1) the development of a description of normal versus diseased cortical geometry, 2) landmark generation for deformable atlas registration, 3) statistical analyses of structure/function relationships, and 4) characterization of morphological changes in aging and disease. The algorithms that will form the basis for the developed software tools stem from our own research work on cortical reconstruction and analysis. We propose five specific aims: 1) incorporate the core volumetric processing algorithms into MIPAV, a commercial-quality but freely available software package developed at NIH; 2) implement a dedicatedGUI- based software package for performing a full reconstruction of the cerebral cortex from MR images; 3) implement surface-based processing and analysis tools within our dedicated software package that will allow definition of regions of interest on the cortex, cortical unfolding, surface editing, and sulcal labeling; 4) perform two phases of testing of the software package followed by a general public release; and 5) implement enhanced scripting and client/server-based versions of the software that will allow customization of the overall processing methodology and offload computationally intensive algorithms to a networked server. These aims will be achieved by 1) porting existing in-house software into the Java programming language using the MIPAV application programming interface, 2) developing graphical user interfaces for visualization and validation, and 3) making refinements based on tester feedback. Source code to all developed software will be made publicly available.