Institution of Information Processing and Automation, College of Information Engineering, Zhejiang University of Technology, Hangzhou, China.
The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, China.
Neurosci Lett. 2022 Aug 10;785:136724. doi: 10.1016/j.neulet.2022.136724. Epub 2022 Jun 10.
Diffusion magnetic resonance imaging tractography allows investigating brain structural connections in a noninvasive way and has been widely used for understanding neurological disease. Quantification of brain connectivity along with its length by dividing a fiber bundle into multiple segments (node) is a powerful approach to assess biological properties, which is termed as tractometry. However, current tractometry methods face challenges in node identification along with the length of complex bundles whose morphology is difficult to summarize. In addition, the anatomic measure reflecting the macroscopic fiber cross-section has not been followed in previous tractometry. In this paper, we propose an automated fiber bundle quantification, which we refer to as ClusterMetric. The ClusterMetric uses a data-driven approach to identify fiber clusters corresponding to subdivisions of the white matter anatomy and identify consistent space nodes along the length of clusters across individuals. The proposed method is demonstrated by applicating to our collected dataset including 23 Alzheimer's disease (AD) patients and 22 healthy controls (HCs) and a public dataset of ADNI including 53 AD patients and 85 HCs. The altered white matter tracts in AD group are observed using both datasets, which involve several major fiber tracts including the corpus callosum, corona-radiata-frontal, arcuate fasciculus, inferior occipito-frontal fasciculus, uncinate fasciculus, thalamo-frontal, superior longitudinal fasciculus, inferior cerebellar peduncle, cingulum bundle, and extreme capsule. These fiber clusters represent the white matter connections that could be most affected in AD, suggesting the ability of our method in identifying potential abnormalities specific to local regions within a fiber cluster.
弥散磁共振成像纤维束示踪允许以非侵入性的方式研究大脑的结构连接，并已广泛用于了解神经疾病。通过将纤维束划分为多个片段（节点）来量化脑连接及其长度，是评估生物学特性的一种强大方法，称为束测法。然而，当前的束测法方法在节点识别以及形态难以概括的复杂束的长度方面面临挑战。此外，以前的束测法没有反映宏观纤维横截面积的解剖学测量。在本文中，我们提出了一种自动纤维束量化方法，称为 ClusterMetric。ClusterMetric 使用数据驱动的方法来识别对应于白质解剖细分的纤维簇，并在个体之间的簇长度上识别一致的空间节点。该方法通过应用于包括 23 名阿尔茨海默病（AD）患者和 22 名健康对照（HC）的我们收集的数据集以及包括 53 名 AD 患者和 85 名 HC 的 ADNI 公共数据集来进行演示。使用这两个数据集观察到 AD 组中改变的白质束，这些束涉及几个主要的纤维束，包括胼胝体、辐射冠额部、弓状束、下额枕额束、钩束、丘脑额部、上纵束、小脑下脚、扣带回束和极囊。这些纤维簇代表了在 AD 中最可能受到影响的白质连接，表明我们的方法在识别纤维簇内特定局部区域的潜在异常方面的能力。
