判别性子图挖掘方法及其在MCI分类中的应用

南京大学学报(自然科学版) ›› 2015, Vol. 51 ›› Issue (2): 328–334.

判别性子图挖掘方法及其在MCI分类中的应用

费飞,王立鹏,接标,张道强

出版日期:2015-03-30 发布日期:2015-03-30
作者简介:(南京航空航天大学计算机科学与技术学院南京 210016)
基金资助:
江苏省自然科学基金杰出青年基金( B K 2 0 1 3 0 0 3 4 ) , 高等学校博士学科点专项基金( 2 0 1 2 3 2 1 8 1 1 0 0 0 9 ) , 南京航空航天大
学基本科研业务费( N E 2 0 1 3 1 0 5 ) , 中央高校基本科研业务专项资金( N Z 2 0 1 3 3 0 6

Discriminative subgraph mining with application in MCI classification

Fei Fei, Wang Lipeng, Jie Biao, Zhang Daoqiang*

Online:2015-03-30 Published:2015-03-30
About author:(Colleague of Computer Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing, 210016, China)

摘要/Abstract

摘要： 最近，脑连接网络已经被用于神经退行性疾病（如阿尔茨海默病AD以及轻度认知障碍MCI）的诊断和分类。以往典型方法是从脑连接网络中提取一些特征（如局部聚类系数等）构成一个长特征向量，并用其训练一个分类器用于最终的分类。然而，上述方法的一个缺点是未能充分考虑网络的拓扑结构信息，从而限制了分类性能的进一步提升。有鉴于此，本文提出了一种基于判别子图挖掘的脑连接网络分类方法。首先分别从正类训练样本集和负类训练样本集中挖掘频繁子网络（即频繁子图）。然后，利用基于图核的方法来衡量频繁子网络的判别性能，并选择那些最具判别性的频繁子网络作为判别子网络用于后续的分类。最后，在真实MCI数据集上的实验验证了本文方法的有效性。

Abstract: Brain connectivity networks have been recently used for classification of neurodegenerative diseases, e.g., Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Accurate diagnosis of AD, as well as its prodromal stage (MCI), is very important for possible delay and early treatment of the disease. In conventional connectivity-networks-based classification method, features (e.g., local clustering coefficient, etc.) are often extracted from connectivity networks and concatenated into a long vector to train a classifier for final classification. However, one disadvantage of those methods is that some useful network topological information was not fully considered, which limits the further improvement of classification performance. Accordingly, in this paper, we propose a novel brain connectivity-network classification method based on discriminative subgraph mining, which can reflect the intrinsic disease pathology. Specifically, in preprocessing stage, we first use the specific application tool to original fluoro-deoxy-glucose-positron emission tomography (FDG_PET) images data from 27 subjects. For each preprocessed original FDG-PET image, 90 regions of interest (ROIs) are labeled by an atlas warping algorithm. And then, we construct the connectivity networks according to the fiber matter between these brain regions. After, we extract a set of frequent subnetworks (i.e., subgraphs) with using frequent subgraph mining algorithm from each of the two groups (i.e., MCI and NC), respectively. Finally, we measure the discriminative ability of those frequent subnetworks using graph-kernel-based classification method and select the most discriminative subnetworks for subsequent classification. The classification experiment scheme is tested on the MCI dataset by leave-one-out (LOO) cross-validation method. The experimental results show the efficacy of our proposed method with comparison to the state-of-the-art method for connectivity-networks based MCI classification, and our method can gain a better insight in the disease pathology

费飞,王立鹏,接标,张道强. 判别性子图挖掘方法及其在MCI分类中的应用[J]. 南京大学学报(自然科学版), 2015, 51(2): 328–334.

Fei Fei, Wang Lipeng, Jie Biao, Zhang Daoqiang*. Discriminative subgraph mining with application in MCI classification[J]. Journal of Nanjing University(Natural Sciences), 2015, 51(2): 328–334.

参考文献

[1] Brookmeyer R, Johnson E, Ziegler-Graham K, et al. Forecasting the global burden of Alzheimer’s disease. Alzheimers Dement, 2007, 3(3): 186~191.
[2] Petersen, R C, Doody R, Kurz A, et al. Current concepts in mild cognitive impairment. Archives of neurology, 2001, 58(12): 1985~1992.
[3] Zhang D, Wang Y, Zhou L, et al. Multimodal classification of Alzheimer’s disease and mild cognitive impairment. Neuroimage, 2011, 55(3): 856~867.
[4] 程波,张道强.基于多模态向量回归的老年痴呆症临床变量预测.南京大学学报(自然科学), 2012, 48(2): 140~146.
[5] Wee C Y, Yap P T, Zhang D, et al. Identification of MCI individuals using structural and functional connectivity networks. Neuroimage, 2012, 59(3): 2045~2056.
[6] Jie B, Zhang D, Gao W, et al. Integration of network topological and connectivity properties for neuroimaging classification. IEEE transactions on biomedical engineering, 2014, 1.
[7] Wang J, Zuo X, Dai Z, et al. Disrupted functional brain connectome in individuals at risk for Alzheimer’s disease. Biological Psychiatry, 2013, 73(5): 472~481.
[8] He Y, Chen Z, Evans A C. Small-world anatomical networks in the human brain revealed by cortical thickness from MRI. Cerebral Cortex, 2007, 17(10): 2407~2419.
[9] Iturria-Medina Y, Canales-Rodriguez E J, Melie-Garcia L, et al. Characterizing brain anatomical connections using diffusion weighted MRI and graph theory. Neuroimage, 2007, 36(3): 645~660.
[10] Wee C Y, Yap P T, Li W, et al. Enriched white matter connectivity networks for accurate identification of MCI patients. Neuroimage, 2011, 54: 1812~1822.
[11] Chen G, Ward B D, Xie C, et al. Classification of Alzheimer disease, mild cognitive impairment, and normal cognitive status with large-scale network analysis based on resting-state functional MR imaging. Radiology, 2011, 259: 213~221.
[12] Yan X, Han J. gSpan: Graph-based substructure pattern mining. In: 2002 IEEE International Conference on Data Mining. Maebashi City: IEEE, 2002, 721~724.
[13] Inokuchi A, Washio T, Motoda H. An apriori-based algorithm for mining frequent substructures from graph data. In: Principles of data mining and knowledge discovery. Berlin Heidelberg: Springer, 2000, 1910: 13~23.
[14] Jiang C, Coenen F, Zito M. A survey of frequent subgraph mining algorithms. Knowledge Engineering Review, 2013, 28: 75~105.
[15] Gartner T, Flach P, Wrobel S. On graph kernels: Hardness results and efficient alternatives. Learning Theory and Kernel Machines, 2003, 2777: 129~143.
[16] Alvarez M A, Qi X, Yan C. A shortest-path graph kernel for estimating gene product semantic similarity. Journal of Biomedical Semantics, 2011, 2: 3.
[17] Shervashidze N, Schweitzer P, van Leeuwen E J, et al. Weisfeiler-Lehman graph kernels. Journal of Machine Learning Research, 2011, 12: 2539~2561.
[18] Harchaoui Z, Bach F. Image classification with segmentation graph kernels. In: 2007 IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis: IEEE Press, 2007, 1-8: 612~619.
[19] Borgwardt K M, Ong C S, Schonauer S, et al. Protein function prediction via graph kernels. Bioinformatics, 2005, 21: I47~I56.
[20] Tzourio-Mazoyer N, Landeau B, Papathanassiou D, et al. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage, 2002, 15: 273~289.
[21] Jie B, Zhang D, Wee C Y, et al. Structural feature selection for connectivity network-based MCI diagnosis. In: The 2nd International Workshop on Multimodal Brain Image Analysis. Nice: Springer, 175~184.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed