Online Tools for Customizable ADNI Atlases: Technology and Application

描述(由申请人提供)：这份建议书描述了我们计划建立一套软件工具和在线能力，以便从ADNI图像和患者数据数据库中生成高分辨率、特定于研究的地图集。建议的工具将依赖于大变形差分同胚度量映射(LDDMM)技术，以生成高质量的多模式脑地图集。这个项目的目标不仅仅是一个单一的地图集，也不仅仅是一套新的算法。相反，我们将提供在线服务，用户可以通过该服务请求他们通过Web界面指定的自定义地图集。计算基础设施将把这些地图集提供给用户。 我们还将在地图集建设框架中寻求根本和技术上的改进，并对这些技术进行验证。为了提供这些进展的临床背景，我们将使用这些工具与犹他大学神经病学系的合作研究人员一起进行一项基于图谱的痴呆症代谢不对称研究。 与公共卫生相关：该项目的目标是多方面的。我们将开发用于生成LDDMM的开源软件和一个向世界各地的研究人员开放的网络界面，以根据ADNI数据库查询自动生成和提供高质量的按需地图集。拟议的脑地图集门户将按照用户的具体情况制作：(1)多模式图像，包括受试者与平均值之间的转换；(2)地图集的可靠性指数；(3)量化队列的结构和代谢变异性；(4)一组纵向的多模式地图集。

DESCRIPTION (provided by applicant): This proposal describes our plans to build a set of software tools and an online capability for generating high-resolution, study-specific atlases from the ADNI database of images and patient data. The proposed tools will rely on the technology for large deformation diffeomorphic metric mappings (LDDMMs), to generate high-quality, multimodal brain atlases. The goal of this project is not merely a single atlas nor is it simply a set of new algorithms. Rather, we will provide an online service through which users can request custom atlases that they specify through a web interface. A computational infrastructure will deliver these atlases to the user. We will also pursue fundamental and technological improvements in the atlas construction framework as well as the validation of these technologies. In order to provide a clinical context for these developments, we will use these tools to conduct an atlas-based study on metabolic asymmetry in dementia with coinvestigators in the Department of Neurology at the University of Utah. PUBLIC HEALTH RELEVANCE: The goals of this project are multifold. We will be developing open-source software for generating LDDMMs and a web interface open to researchers around the world for automatically generating and delivering high-quality, on-demand atlases from ADNI database queries. The proposed Brain Atlas Portal will produce, as specified by the user: (i) multimodal imagery, with transformations between subjects and the mean, (ii) reliability index for the atlas, (iii) quantification of structural and metabolic variability of the cohort, and (iv) a set of longitudinal multimodal atlases for a cohort.