Construction of brain functional hypernetwork and feature fusion analysis based on sparse group Lasso method
LI Yao1, ZHAO Yunpeng2, LI Xinyun1, LIU Zhifen3, CHEN Junjie1, GUO Hao1
1. College of Information and Computer Science, Taiyuan University of Technology, Jinzhong Shanxi 030600, China; 2. College of Art, Taiyuan University of Technology, Jinzhong Shanxi 030600, China; 3. Department of Mental Health, First Hospital of Shanxi Medical University, Taiyuan Shanxi 030000, China
摘要 功能超网络广泛地应用于脑疾病诊断和分类研究中，而现有的关于超网络创建的研究缺乏解释分组效应的能力或者仅考虑到脑区间组级的信息，这样构建的脑功能超网络会丢失一些有用的连接或包含一些虚假的信息，因此，考虑到脑区间的组结构问题，引入sparse group Lasso（sgLasso）方法进一步改善超网络的创建。首先，利用sgLasso方法进行超网络创建；然后，引入两组超网络特有的属性指标进行特征提取以及特征选择，这些指标分别是基于单一节点的聚类系数和基于一对节点的聚类系数；最后，将特征选择后得到的两组有显著差异的特征通过多核学习进行特征融合和分类。实验结果表明，所提方法经过多特征融合取得了87.88%的分类准确率。该结果表明为了改善脑功能超网络的创建，需要考虑到组信息，但不能逼迫使用整组信息，可以适当地对组结构进行扩展。
Abstract：Functional hyper-networks are widely used in brain disease diagnosis and classification studies. However, the existing research on hyper-network construction lacks the ability to interpret the grouping effect or only considers the information of group level information of brain regions, the hyper-network constructed in this way may lose some useful connections or contain some false information. Therefore, considering the group structure problem of brain regions, the sparse group Lasso (Least absolute shrinkage and selection operator) (sgLasso) method was introduced to further improve the construction of hyper-network. Firstly, the hyper-network was constructed by using the sgLasso method. Then, two groups of attribute indicators specific to the hyper-network were introduced for feature extraction and feature selection. The indictors are the clustering coefficient based on single node and the clustering coefficient based on a pair of nodes. Finally, the two groups of features with significant difference obtained after feature selection were subjected to multi-kernel learning for feature fusion and classification. The experimental results show that the proposed method achieves 87.88% classification accuracy by using the multi-feature fusion, which indicates that in order to improve the construction of hyper-network of brain function, the group information should be considered, but the whole group information cannot be forced to be used, and the group structure can be appropriately expanded.
李瑶, 赵云芃, 李欣芸, 刘志芬, 陈俊杰, 郭浩. 基于sparse group Lasso方法的脑功能超网络构建与特征融合分析[J]. 计算机应用, 2020, 40(1): 62-70.
LI Yao, ZHAO Yunpeng, LI Xinyun, LIU Zhifen, CHEN Junjie, GUO Hao. Construction of brain functional hypernetwork and feature fusion analysis based on sparse group Lasso method. Journal of Computer Applications, 2020, 40(1): 62-70.
[1] ZENG L, SHEN H, LIU L, et al. Identifying major depression using whole-brain functional connectivity:a multivariate pattern analysis[J]. Brain, 2012, 135(5):1498-1507. [2] NIXON N L, LIDDLE P F, NIXON E, et al. Biological vulnerability to depression:linked structural and functional brain network findings[J]. The British Journal of Psychiatry, 2014, 204(4):283-289. [3] HUANG S, LI J, SUN L, et al. Learning brain connectivity of Alzheimer's disease by sparse inverse covariance estimation[J]. NeuroImage, 2010, 50(3):935-949. [4] JIE B, WEE C Y, SHEN D, et al. Hyper-connectivity of functional networks for brain disease diagnosis[J]. Medical Image Analysis, 2016, 32:84-100. [5] LIU M, ZHANG J, YAP P T, et al. Diagnosis of Alzheimer's disease using view-aligned hypergraph learning with incomplete multi-modality data[C]//Proceedings of the 2016 International Conference on Medical Image Computing and Computer-Assisted Intervention, LNCS 9900. Cham:Springer, 2016:308-316. [6] 彭瑶,祖辰,张道强.基于超图的多模态特征选择算法及其应用[J].计算机科学与探索,2018,12(1):112-119.(PENG Y, ZU C, ZHANG D Q. Hypergraph based multi-modal feature selection and its application[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(1):112-119.) [7] 靳研艺,郭浩,陈俊杰.基于elasticnet方法的静息态脑功能超网络构建优化[J].计算机应用研究,2018,35(11):3276-3280,3297.(JIN Y Y, GUO H, CHEN J J. Optimization of resting-state brain functional hyper-network construction based on elastic net[J]. Application Research of Computers, 2018, 35(11):3276-3280, 3297.) [8] 张帆,陈俊杰,郭浩.基于脑功能超网络的多特征融合分类方法[J].计算机工程与应用,2018,54(21):120-127.(ZHANG F, CHEN J J, GUO H. Machine learning classification method combining multiple features of brain function hyper-network[J]. Computer Engineering and Applications, 2018, 54(21):120-127.) [9] GU S, YANG M, MEDAGLIA J D, et al. Functional hypergraph uncovers novel covariant structures over neurodevelopment[J]. Human Brain Mapping, 2017, 38(8):3823-3835. [10] ZU C, GAO Y, MUNSELL B, et al. Identifying disease-related subnetwork connectome biomarkers by sparse hypergraph learning[J]. Brain Imaging and Behavior, 2019, 13(4):879-892. [11] GALLAGHER S R, GOLDBERG D S. Clustering coefficients in protein interaction hypernetworks[C]//Proceedings of the 2013 International Conference on Bioinformatics, Computational Biology and Biomedical Informatics. New York:ACM, 2013:552-560. [12] ZOU H, HASTIE T. Regularization and variable selection via the elastic net[J]. Journal of the Royal Statistical Society:Series B (Statistical Methodology), 2005, 67(2):301-320. [13] LIU X, GONCALVES A R, CAO P, et al. Modeling Alzheimer's disease cognitive scores using multi-task sparse group lasso[J]. Computerized Medical Imaging and Graphics, 2018, 66:100-114. [14] GUO H, LI Y, XU Y, et al. Resting-state brain functional hyper-network construction based on elastic net and group lasso methods[J]. Frontiers in Neuroinformatics, 2018, 12:No.25. [15] FRIEDMAN J, HASTIE T, TIBSHIRANI R. A note on the group lasso and a sparse group lasso[R]. Stanford:Stanford University, 2010. [16] OGUTU J O, PIEPHO H P. Regularized group regression methods for genomic prediction:Bridge, MCP, SCAD, group bridge, group lasso, sparse group lasso, group MCP and group SCAD[J]. BMC Proceedings, 2014, 8(S5):No.S7. [17] MATSUI H. Sparse group lasso for multiclass functional logistic regression models[J]. Communications in Statistics-Simulation and Computation, 2019, 48(6):1784-1797. [18] GOLDBERG D S, ROTH F P. Assessing experimentally derived interactions in a small world[J]. Proceedings of the National Academy of Sciences of the United States of America, 2003, 100(8):4372-4376. [19] LATAPY M, MAGNIEN C, VECCHIO N D. Basic notions for the analysis of large two-mode networks[J]. Social Networks, 2008, 30(1):31-48. [20] FORNITO A, ZALESKY A, BREAKSPEAR M. Graph analysis of the human connectome:promise, progress, and pitfalls[J]. NeuroImage, 2013, 80:426-444. [21] KAUFMANN M, VAN KREVELD M, SPECKMANN B. Subdivision drawings of hypergraphs[C]//Proceedings of the 2008 International Symposium on Graph Drawing, LNCS 5417. Berlin:Springer, 2008:396-407. [22] 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[J]. NeuroImage, 2002, 15(1):273-289. [23] MEIER L, VAN DE GEER S, BVHLMANN P. The group lasso for logistic regression[J]. Journal of the Royal Statistical Society:Series B (Statistical Methodology), 2008, 70(1):53-71. [24] PARK H S, JUN C H. A simple and fast algorithm for K-medoids clustering[J]. Expert Systems with Applications, 2009, 36(2):3336-3341. [25] YUAN M, LIN Y. Model selection and estimation in regression with grouped variables[J]. Journal of the Royal Statistical Society:Series B (Statistical Methodology), 2006, 68(1):49-67. [26] FRIEDMAN J, HASTIE T, TIBSHIRANI R. Regularization paths for generalized linear models via coordinate descent[R]. Stanford:Stanford University, 2009. [27] SIMON N, FRIEDMAN J, HASTIE T, et al. A sparse-group lasso[J]. Journal of Computational and Graphical Statistics, 2013, 22(2):231-245. [28] LIU J, JI S, YE J. SLEP:sparse learning with efficient projections[EB/OL].[2019-01-20].http://read.pudn.com/downloads701/sourcecode/math/2820891/SLEP/manual.pdf. [29] FASANO G, FRANCESCHINI A. A multidimensional version of the Kolmogorov-Smirnov test[J]. Monthly Notices of the Royal Astronomical Society, 1987, 225(1):9-20. [30] BENJAMINI Y, HOCHBERG Y. Controlling the false discovery rate:a practical and powerful approach to multiple testing[J]. Journal of the Royal Statistical Society:Series B (Statistical Methodology), 1995, 57(1):289-300. [31] ZHANG D, WANG Y, ZHOU L, et al. Multimodal classification of Alzheimer's disease and mild cognitive impairment[J]. NeuroImage, 2011, 55(3):856-867. [32] ZHU J, SHEN X, QIN J, et al. Altered anatomical modular organization of brain networks in patients with major depressive disorder[C]//Proceedings of the 2016 International Conference on Biological Sciences and Technology. Paris:Atlantis Press, 2016, 2:284-289. [33] LIU F, HU M, WANG S, et al. Abnormal regional spontaneous neural activity in first-episode, treatment-naive patients with late-life depression:a resting-state fMRI study[J]. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 2012, 39(2):326-331. [34] JIN C, GAO C, CHEN C, et al. A preliminary study of the dysregulation of the resting networks in first-episode medication-naive adolescent depression[J]. Neuroscience Letters, 2011, 503(2):105-109. [35] GUO H, CAO X, LIU Z, et al. Machine learning classifier using abnormal brain network topological metrics in major depressive disorder[J]. Neuroreport, 2012, 23(17):1006-1011. [36] QIU L, HUANG X, ZHANG J, et al. Characterization of major depressive disorder using a multiparametric classification approach based on high resolution structural images[J]. Journal of Psychiatry and Neuroscience, 2014, 39(2):78-86. [37] LORD A, HORN D, BREAKSPEAR M, et al. Changes in community structure of resting state functional connectivity in unipolar depression[J]. PLoS One, 2012, 7(8):No.e41282. [38] LIU F, GUO W, LIU L, et al. Abnormal amplitude low-frequency oscillations in medication-naive, first-episode patients with major depressive disorder:a resting-state fMRI study[J]. Journal of Affective Disorders, 2013, 146(3):401-406. [39] ROLLS E T, CHENG W, GILSON M, et al. Effective connectivity in depression[J]. Biological Psychiatry:Cognitive Neuroscience and Neuroimaging, 2018, 3(2):187-197. [40] LUI S, WU Q, QIU L, et al. Resting-state functional connectivity in treatment-resistant depression[J]. American Journal of Psychiatry, 2011, 168(6):642-648. [41] GUO H, YAN P, CHENG C, et al. fMRI classification method with multiple feature fusion based on minimum spanning tree analysis[J]. Psychiatry Research:Neuroimaging, 2018, 277:14-27. [42] ZHANG J, WANG J, WU Q, et al. Disrupted brain connectivity networks in drug-naive, first-episode major depressive disorder[J]. Biological Psychiatry, 2011, 70(4):334-342. [43] FITZGERALD P B, LAIRD A R, MALLER J, et al. A meta-analytic study of changes in brain activation in depression[J]. Human Brain Mapping, 2008, 29(6):683-695. [44] 皇甫浩然,杨剑,杨阳.基于fMRI动态功能连接的抑郁症患者分类研究[J].计算机应用研究,2017,34(3):678-682.(HUANGFU H R, YANG J, YANG Y. Classifying patients with depression based on fMRI dynamic functional connectivity[J]. Application Research of Computers, 2017, 34(3):678-682.) [45] ROSA M J, PORTUGAL L, HAHN T, et al. Sparse network-based models for patient classification using fMRI[J]. NeuroImage, 2015, 105:493-506.
2020, Vol. 40 
