Qin Jiaolong, Wei Maobin, Liu Haiyan, Chen Jianhuai, Yan Rui, Hua Lingling, Zhao Ke, Yao Zhijian, Lu Qing
Key Laboratory of Child Development and Learning Science (Ministry of education), Research Centre for Learning Science, Southeast University, Nanjing 210096, China.
Department of Psychiatry, Affiliated Nanjing Brain Hospital of Nanjing Medical University, Nanjing 210029, China.
Magn Reson Imaging. 2014 Dec;32(10):1314-20. doi: 10.1016/j.mri.2014.08.037. Epub 2014 Aug 29.
Previous studies had explored the diagnostic and prognostic value of the structural neuroimaging data of MDD and treated the whole brain voxels, the fractional anisotropy and the structural connectivity as classification features. To our best knowledge, no study examined the potential diagnostic value of the hubs of anatomical brain networks in MDD. The purpose of the current study was to provide an exploratory examination of the potential diagnostic and prognostic values of hubs of white matter brain networks in MDD discrimination and the corresponding impaired hub pattern via a multi-pattern analysis. We constructed white matter brain networks from 29 depressions and 30 healthy controls based on diffusion tensor imaging data, calculated nodal measures and identified hubs. Using these measures as features, two types of feature architectures were established, one only included hubs (HUB) and the other contained both hubs and non hubs. The support vector machine classifiers with Gaussian radial basis kernel were used after the feature selection. Moreover, the relative contribution of the features was estimated by means of the consensus features. Our results presented that the hubs (including the bilateral dorsolateral part of superior frontal gyrus, the left middle frontal gyrus, the bilateral middle temporal gyrus, and the bilateral inferior temporal gyrus) played an important role in distinguishing the depressions from healthy controls with the best accuracy of 83.05%. Moreover, most of the HUB consensus features located in the frontal-parieto circuit. These findings provided evidence that the hubs could be served as valuable potential diagnostic measure for MDD, and the hub-concentrated lesion distribution of MDD was primarily anchored within the frontal-parieto circuit.
以往的研究探讨了重度抑郁症（MDD）结构神经影像数据的诊断和预后价值，并将全脑体素、分数各向异性和结构连接性作为分类特征。据我们所知，尚无研究考察MDD中大脑解剖网络枢纽的潜在诊断价值。本研究的目的是通过多模式分析，对MDD鉴别中白质脑网络枢纽的潜在诊断和预后价值以及相应的受损枢纽模式进行探索性研究。我们基于扩散张量成像数据，从29名抑郁症患者和30名健康对照者构建了白质脑网络，计算节点指标并识别枢纽。以这些指标为特征，建立了两种类型的特征架构，一种仅包含枢纽（HUB），另一种同时包含枢纽和非枢纽。特征选择后，使用具有高斯径向基核的支持向量机分类器。此外，通过一致性特征估计特征的相对贡献。我们的结果表明，枢纽（包括双侧额上回背外侧部分、左侧额中回、双侧颞中回和双侧颞下回）在区分抑郁症患者和健康对照者方面发挥了重要作用，最佳准确率为83.05%。此外，大多数HUB一致性特征位于额顶叶回路。这些发现提供了证据，表明枢纽可作为MDD有价值的潜在诊断指标，且MDD的枢纽集中性病变分布主要定位于额顶叶回路。
