Multi-Channel Sparse Graph Transformer Network for Early Identification of Alzheimer′s Disease
Qiu Yali, Zhu Yun, Yu Shuangzhi, Song Xuegang, Wang Tianfu, Lei Baiying*
School of Biomedical Engineering, Shenzhen University, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, Shenzhen 518060, Guangdong, China
Abstract：Currently, there is no effective treatment for Alzheimer's disease (AD). Clinical intervention via early diagnosis can delay the AD progression and improve its prognosis. However, the existing methods only consider the neuroimaging features learned from group relationships, not the individual characteristics of the subjects. In this work, we designed a novel multi-modal multi-channel sparse graph transformer network (MSGTN). Our proposed network model included two parts, they are multi-modal data optimization and multi-modal feature learning. Firstly, we acquired the image information (e.g., diffusion tensor imaging (DTI) and functional magnetic resonance imaging (fMRI)) and non-image information (e.g., age and sex) of each subject. Secondly, we utilized locally weighted clustering coefficients (LWCC) to fuse functional and structural information. After that, the fused multi-modal image features were combined with the gender and age information of the subjects to construct a sparse graph. Finally, we input the sparse graph into the MSGTN network for early AD identification. We obtained a total of 170 subjects from the public database ADNI (Alzheimer's disease neuroimaging initiative), including 38 LMCI, 44 EMCI, 44 SMC, and 44 normal controls (NC). Our method achieved classification accuracy of 87.02%, 87.40%, 91.49%, 88.93%, 86.74% and 92.12%, respectively. The experimental results have proved that our proposed model not only can analyze NC versus three different early AD disease states, but also achieved superior classification performance in three different early AD disease states.
[1] Song X, Elazab A, Zhang Y. Classification of mild cognitive impairment based on a combined high-order network and graph convolutional network [J]. IEEE Access, 2020, 8: 42816-42827. [2] 韩璎, 陈观群. 中国AD临床前期联盟:开启我国阿尔茨海默病临床研究新时代 [J]. 医学研究杂志, 2019, 047(7): 1-4. [3] Patterson C. World Alzheimer report 2018- the state of the art of dementia research: new frontiers-an analysis of prevalence, incidence, cost and trends [R]. London, SE1 0BB, UK: Alzheimer′s Disease International, 2018. [4] Davatzikos C, Bhatt P, Shaw LM, et al. Prediction of MCI to AD conversion, via MRI, CSF biomarkers, and pattern classification [J]. Neurobiology of Aging, 2011, 32(12): 2319-2327. [5] Cherubini A, Péran P, Spoletini I, et al. Combined volumetry and DTI in subcortical structures of mild cognitive impairment and Alzheimer's disease patients [J]. Journal of Alzheimer's Disease, 2010, 19(4): 1273-1282. [6] Teipel SJ, Wohlert A, Metzger C, et al. Multicenter stability of resting state fMRI in the detection of Alzheimer's disease and amnestic MCI [J]. NeuroImage: Clinical, 2017, 14: 183-194. [7] Li F, Tran L, Thung K-H, et al. A robust deep model for improved classification of AD/MCI patients [J]. International Journal of Computer Applications, 2015, 19(5): 1610-1616. [8] Yang X, Jin Y, Chen X, et al. Functional connectivity network fusion with dynamic thresholding for MCI diagnosis [C]// Medical Image Computing and Computer-Assisted Intervention. Athens: Springer Verlag, 2016: 246-253. [9] Lei B, Yang P, Wang T, et al. Relational regularized discriminative sparse learning for Alzheimer's disease diagnosis [J]. IEEE Transactions on Cybernetics, 2017, 47(4): 1102-1113. [10] Wee C, Yap P, Zhang D, et al. Identification of MCI individuals using structural and functional connectivity networks [J]. NeuroImage, 2012, 59(3): 2045-2056. [11] Gao Y, Wee C, Kim M, et al. MCI identification by joint learning on multiple MRI data [C]// Medical Image Computing and Computer-Assisted Intervention. Munich: Springer Verlag, 2015: 78-85. [12] Zhu Y, Zhu X, Kim M, et al. Dynamic hyper-graph inference framework for computer-assisted diagnosis of neurodegenerative diseases [J]. IEEE Transactions on Medical Imaging, 2018, 38(2): 608-616. [13] Greicius MD, Supekar K, Menon V, et al. Resting-state functional connectivity reflects structural connectivity in the default mode network [J]. Cerebral Cortex, 2009, 19(1): 72-78. [14] Damoiseaux JS, Greicius MD. Greater than the sum of its parts: a review of studies combining structural connectivity and resting-state functional connectivity [J]. Brain Structure and Function, 2009, 213(6): 525-533. [15] Li W, Lin Y, Liu Y. The structure of weighted small-world networks [J]. Physica A: Statistical Mechanics and its Applications, 2007, 376: 708-718. [16] Yu S, Wang S, Xiao X, et al. Multi-scale enhanced graph convolutional network for early mild cognitive impairment detection [C]// Medical Image Computing and Computer-Assisted Intervention. Lima: Springer Science and Business Media Deutschland GmbH, 2020: 228-237. [17] Zhu X, Suk H, Huang H, et al. Low-rank graph-regularized structured sparse regression for identifying genetic biomarkers [J]. IEEE Transactions on Big Data, 2017, 3(4): 405-414. [18] Chen M, Wei Z, Huang Z, et al. Simple and deep graph convolutional networks [C]// International Conference on Machine Learning. Seattle: International Machine Learning Society (IMLS), 2020: 1703-1713. [19] Kazi A, Shekarforoush S, Krishna SA, et al. InceptionGCN: receptive field aware graph convolutional network for disease prediction [C]// Information Processing in Medical Imaging. Hong Kong: Springer Verlag, 2019: 73-85. [20] Wan S, Gong C, Zhong P, et al. Hyperspectral image classification with context-aware dynamic graph convolutional network [J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 59(1): 597-612. [21] Perozzi B, Alrfou R, Skiena S. DeepWalk: online learning of social representations [C]// 20th ACM SIGKDD International Conference on Knowledge discovery and Data Mining (KDD). New York: Association for Computing Machinery, 2014: 701-710. [22] Abuelhaija S, Kapoor A, Perozzi B, et al. N-GCN: Multi-scale graph convolution for semi-supervised node classification [EB/OL]// arXiv.1802.08888, 2018-02-24/2021-07-11. [23] Chandra R, Optican LM. Detection, classification, and superposition resolution of action potentials in multiunit single-channel recordings by an on-line real-time neural network [J]. International Journal of Computer Applications, 1997, 44(5): 403-412. [24] Zhu G, Li Y, Wen P. Analysis and classification of sleep stages based on difference visibility graphs from a single-channel EEG signal [J]. International Journal of Computer Applications, 2014, 18(6): 1813-1821. [25] Wang S, Wang X, Hu Y, et al. Diabetic retinopathy diagnosis using multichannel generative adversarial network with semisupervision [J]. IEEE Transactions on Automation Science and Engineering, 2020, 18(2): 574-585. [26] Kipf TN, Welling M. Semi-supervised classification with graph convolutional networks [C]// International Conference on Learning Representations. Toulon: ICLR, 2016. [27] Yun S, Jeong M, Kim R, et al. Graph transformer networks [C]// Neural Information Processing Systems (NeurIPS). Vancouver: Curran Associates Inc., 2019. [28] Cui Z, Zhong S, Xu P, et al. PANDA: a pipeline toolbox for analyzing brain diffusion images [J]. Frontiers in Human Neuroscience, 2013, 7(42): 1-16. [29] Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL [J]. NeuroImage, 2004, 23: S208-S219. [30] 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. [31] Vorlí?ková J. Least absolute shrinkage and selection operator method [D]. Institut Ekonomickch Studií; Univerzita Karlova, 2017. [32] Jacob L, Vert J-p, Bach FR. Clustered multi-task learning: A convex formulation [C]// Neural Information Processing Systems. Vancouve: Curran Associates Inc., 2009: 745-752. [33] Li Y, Liu J, Tang Z, et al. Deep spatial-temporal feature fusion from adaptive dynamic functional connectivity for MCI identification [J]. IEEE Transactions on Medical Imaging, 2020, 39(9): 2818-2830. [34] Lei B, Nina C, Frangi AF, et al. Self-calibrated brain network estimation and joint non-convex multi-task learning for Identification of early Alzheimer's disease [J]. Medical Image Analysis, 2020, 61: 101652. [35] Song X, Zhou F, Frangi AF, et al. Graph convolution network with similarity awareness and adaptive calibration for disease-induced deterioration prediction [J]. Medical Image Analysis, 2021, 69: 101947. [36] Wang Z, Liu Q, Dou Q. Contrastive cross-site learning with redesigned net for COVID-19 CT classification [J]. IEEE Journal of Biomedical and Health Informatics, 2020, 24(10): 2806-2813. [37] Parisot S, Ktena SI, Ferrante E, et al. Disease prediction using graph convolutional networks: application to autism spectrum disorder and Alzheimer's disease [J]. Medical Image Analysis, 2018, 48: 117-130. [38] Yu R, Zhang H, An L, et al. Connectivity strength-weighted sparse group representation-based brain network construction for MCI classification [J]. Human Brain Mapping, 2017, 38(5): 2370-2383. [39] Yang P, Zhou F, Ni D, et al. Fused sparse network learning for longitudinal analysis of mild cognitive impairment [J]. IEEE Transactions on Cybernetics, 2019, 51(1): 233-246. [40] Zhang Y, Zhang H, Adeli E, et al. Multiview feature learning with multiatlas-based functional connectivity networks for MCI diagnosis [J]. IEEE Transactions on Cybernetics, 2020: 1-12.
