Development of a Brain Template for Diffusion-Tensor MRI

描述（由申请人提供）：开发用于扩散张量成像（DTI）的大脑模板对于比较不同人群的神经元结构完整性和大脑连通性至关重要。DTI是一种非侵入性技术，可提供有关脑组织微结构特征的独特信息。然而，患者群体和健康志愿者之间DTI结果的比较主要集中在扩散张量的标量上，忽略了张量中可用的部分信息。造成这种情况的主要原因是无法对DTI数据进行精确的空间归一化，这是比较某些方向相关张量信息的必要步骤。传统的DTI数据采集是基于回波平面成像（EPI），存在失真和图像伪影的问题，这在一定程度上降低了DTI数据空间归一化的精度。此外，为了提高归一化的精度，必须在配准过程中使用张量的所有信息。然而，大脑结构与其张量的精确匹配需要非线性配准方法，而非线性配准方法对张量的噪声很敏感。最后，不存在既包含解剖特征又包含低噪声DTI信息的脑模板。以上因素均降低了DTI数据登记的准确性，妨碍了脑组织某些弥散和结构特征的组间比较，从而限制了DTI的临床应用潜力。相比之下，Turboprop-DTI是一种成像技术，与基于epi的DTI相比，它提供的DTI数据具有更少的伪影。然而，与基于epi的DTI相比，Turboprop-DTI的特点是数据采集速度较慢，噪声水平较高。我们最近提出了一种基于非均匀快速傅里叶变换（NUFFT）的涡桨成像迭代图像重建方法，与传统的涡桨成像技术相比，该方法可以提高精度并降低噪声水平。因此，该项目的主要目标是：a)开发Turboprop-DTI数据采集策略，结合我们最近引入的图像重建技术，将在临床可接受的时间内提供低噪声内容和最小伪影的数据；b)开发对张量噪声不太敏感的鲁棒配准技术，以便c)为DTI生成准确的大脑模板。这项研究的成功完成将允许DTI数据的准确登记，并在人群中进行结构完整性和大脑连接的全面比较。因此，本研究的结果将增强DTI作为广泛临床问题的诊断工具的作用。开发用于扩散张量成像（DTI）的大脑模板对于比较不同人群的神经元结构完整性和大脑连通性至关重要。该项目的总体目标是开发稳健的配准技术，并为DTI生成准确的大脑模板。这项研究的成功完成将允许对不同人群的结构完整性和大脑连通性进行全面比较，并将增强DTI作为广泛临床问题诊断工具的作用。

DESCRIPTION (provided by applicant): The development of a brain template for diffusion tensor imaging (DTI) is crucial for comparisons of neuronal structural integrity and brain connectivity across populations. DTI is a non-invasive technique that provides unique information regarding the microstructural characteristics of brain tissue. However, comparisons of DTI results between populations of patients and healthy volunteers have primarily focused only on scalar quantities derived from the diffusion tensor, overlooking portion of the information available in the tensor. The primary reason for this is the inability to perform accurate spatial normalization of DTI data, which is a necessary step for the comparison of certain orientation-dependent tensor information. The accuracy of spatial normalization of DTI data is compromised partly due to the fact that conventional DTI data acquisitions are based on echo- planar imaging (EPI), which suffers from distortions and image artifacts. Furthermore, to increase the accuracy of normalization, all the information of the tensor must be used in the registration process. However, accurate matching of brain structures and their tensors requires non-linear registration methods, which are sensitive to the tensors' noise. Finally, a brain template that contains not only anatomical features but also DTI information with low noise content does not exist. All of the above factors reduce the accuracy in registration of DTI data and prevent intergroup comparisons of certain diffusion and structural characteristics of brain tissue, thus limiting the clinical potential of DTI. In contrast, Turboprop-DTI is an imaging technique that provides DTI data with significantly fewer artifacts than EPI-based DTI. However, Turboprop-DTI is characterized by slower data acquisition and higher noise-levels than EPI-based DTI. We recently introduced an iterative image reconstruction method for Turboprop imaging based on the non-uniform fast Fourier transform (NUFFT), which can increase accuracy and reduce noise levels compared to conventional Turboprop reconstruction techniques. Therefore, the broad objectives of this project are: a) to develop Turboprop-DTI data acquisition strategies that, in combination with our recently introduced image reconstruction technique, will provide data with low noise content and minimal artifacts in a clinically acceptable time, b) to develop robust registration techniques that are less sensitive to the tensors' noise, in order to c) produce an accurate brain template for DTI. The successful completion of this research will allow for accurate registration of DTI data and for comprehensive comparisons of structural integrity and brain connectivity across populations. Therefore, the results of this research will enhance the role of DTI as a diagnostic tool for a wide range of clinical problems. PUBLIC HEALTH RELEVANCE The development of a brain template for diffusion tensor imaging (DTI) is crucial for comparisons of neuronal structural integrity and brain connectivity across populations. The broad objective of this project is to develop robust registration techniques and produce an accurate brain template for DTI. The successful completion of this research will allow for comprehensive comparisons of structural integrity and brain connectivity across populations, and will enhance the role of DTI as a diagnostic tool for a wide range of clinical problems.