Please wait a minute...
浙江大学学报(工学版)
浙江大学学报(工学版)  2021, Vol. 55 Issue (7): 1270-1278    DOI: 10.3785/j.issn.1008-973X.2021.07.006
计算机与控制工程     
基于宽度学习系统的fMRI数据分类方法
刘嘉诚(),冀俊忠*()
北京工业大学 信息学部，北京 100124
Classification method of fMRI data based on broad learning system
Jia-cheng LIU(),Jun-zhong JI*()
Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China
 全文: PDF(1078 KB)   HTML
摘要：

提出基于宽度学习系统的功能性磁共振成像（fMRI）数据分类方法，通过简单结构提取fMRI数据的深层特征，加快分类速度. 使用fMRI中感兴趣区域体素均值的时间序列构造输入数据，分别提取fMRI数据的浅层和深层特征，映射为宽度学习的特征节点和增强节点并构建模型框架，利用岭回归逆计算分类模型的连接权值，实现对fMRI数据的分类. 使用ABIDE Ⅰ、ABIDE Ⅱ和ADHD-200数据集，将所提方法与6种分类方法进行对比实验，结果表明，所提方法可以在保持良好的分类准确率的同时，大幅度降低训练时间.
关键词： 功能性磁共振成像（fMRI）数据分类;  深度学习;  宽度学习系统;  随机特征映射;  特征增强;  岭回归逆    
Abstract:

A functional magnetic resonance imaging (fMRI) data classification method based on broad learning system was proposed. The deep features of fMRI data were extracted through a simple structure to speed up the classification. Using the time series of the mean values of the voxel in the region of interest in fMRI the input data was constructed. The shallow and deep features of fMRI data were extracted respectively, mapped to feature nodes and enhancement nodes for broad learning, and a model framework was built. Ridge regression was used to inversely calculate the connection weights of the classification model to achieve fMRI data classification. ABIDE I, ABIDE II and ADHD-200 were used to compare the proposed method with six classification methods. Results show that the proposed method can maintain good classification accuracy while reduce training time greatly.
Key words: functional magnetic resonance imaging (fMRI) data classification    deep learning    broad learning system    random feature mapping    feature enhancement    ridge regression inverse
收稿日期: 2020-03-04 出版日期: 2021-07-05
CLC:  TP 391  
基金资助: 国家自然科学基金资助项目（61672065，61906010）；北京市教委科技计划一般项目（KM202010005032）
通讯作者: 冀俊忠     E-mail: from_soldier@sina.com;jjz01@bjut.edu.cn
作者简介: 刘嘉诚（1995—），男，硕士生，从事人工智能算法研究. orcid.org/0000-0003-3489-2011. E-mail： from_soldier@sina.com
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章　　
刘嘉诚
冀俊忠

引用本文:

刘嘉诚,冀俊忠. 基于宽度学习系统的fMRI数据分类方法[J]. 浙江大学学报(工学版), 2021, 55(7): 1270-1278.

Jia-cheng LIU,Jun-zhong JI. Classification method of fMRI data based on broad learning system. Journal of ZheJiang University (Engineering Science), 2021, 55(7): 1270-1278.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2021.07.006    或    https://www.zjujournals.com/eng/CN/Y2021/V55/I7/1270

图 1  宽度学习系统的基本结构
图 2  基于宽度学习的fMRI数据分类方法示意图
数据集 样本量/个 正常被试量/个 患者量/个 机构量/个
ABIDE Ⅰ 1 096 569 527 17
ABIDE Ⅱ 1 043 556 487 16
ADHD-200 445 277 168 4
表 1  3个数据集的基本情况
图 3  在准确率不同参数下的取值
数据集 N1 N2 n3
ABIDE I 10 10 10 000
ABIDE II 10 10 10 000
ADHD-200 9 11 5 000
表 2  3个数据集的参数设置
方法 结构
SVM 使用Puthon中的默认模块函数，设置神经节点数为10
RF 使用Puthon中的默认模块函数，设置神经节点数为10
KNN 使用Puthon中的默认模块函数，设置神经节点数为10
DNN [6 670，1 000，600，96，2]
GCN [116*116，32@116*116，64@1*116，128@116*1，96，2]
CCNN [116*116，32@116*116，64@1*116，128@116*1，96，2]
表 3  对比方法的参数设置
方法类别 方法 Acc /% Pr /% Sn /% Sp /% F-measure
传统机器学习 SVM 57.81 55.92 88.37 24.89 68.48
RF 62.40 62.52 69.24 55.09 65.59
KNN 58.80 56.73 87.60 27.77 68.81
深度学习 GCN 64.59 62.62 64.33 58.36 63.33
CCNN 65.60 64.61 76.77 53.49 69.66
DNN 63.98 65.24 67.13 60.71 65.86
宽度学习 本文 64.48 78.21 63.12 50.85 69.91
表 4  7种算法在ABIDE Ⅰ数据集上的实验结果
方法类别 方法 Acc /% Pr /% Sn /% Sp /% F-measure
传统机器学习 SVM 54.26 53.90 98.66 3.57 69.70
RF 61.18 61.43 73.38 47.26 66.83
KNN 56.63 58.85 63.74 48.61 60.76
深度学习 GCN 62.03 64.52 72.73 54.17 68.38
CCNN 65.47 66.73 70.69 59.49 68.52
DNN 66.07 66.58 74.11 74.11 69.90
宽度学习 本文 65.29 86.67 62.93 40.28 72.55
表 5  7种算法在ABIDE Ⅱ数据集上的实验结果
方法类别 方法 Acc /% Pr /% Sn /% Sp /% F-measure
传统机器学习 SVM 59.04 64.98 36.97 80.65 46.84
RF 59.21 59.83 53.53 64.77 56.40
KNN 58.93 61.93 45.79 71.76 52.22
深度学习 GCN 60.61 57.97 48.57 66.21 50.77
CCNN 60.28 59.27 59.11 60.00 58.88
DNN 62.73 59.40 61.85 63.17 61.82
宽度学习 本文 61.19 51.33 64.11 72.92 56.20
表 6  7种算法在ADHD-200数据集上的实验结果
图 4  7种方法在3个数据集上的训练时间与准确率比较
1 BELLIVEAU J W, KENNEDY D N, MCKINSTRY R C, et al Functional mapping of the human visual cortex by magnetic resonance imaging[J]. Science, 1991, 254 (5032): 716- 719
doi: 10.1126/science.1948051
2 COX D D, SAVOY R L Functional magnetic resonance imaging (fMRI)“brain reading”: detecting and classifying distributed patterns of fMRI activity in human visual cortex[J]. Neuro Image, 2003, 19 (2): 261- 270
3 CHENG B, LIU M, SHEN D, et al Multi-domain transfer learning for early diagnosis of Alzheimer ’ s disease[J]. Brain, 2012, 135 (5): 1498- 1507
doi: 10.1093/brain/aws059
4 ROSA M J, PORTUGAL L, HAHN T, et al Sparse network-based models for patient classification using fMRI[J]. Neuroimage, 2015, 105: 493- 506
doi: 10.1016/j.neuroimage.2014.11.021
5 SACCHET M D, PRASAD G, FOLAND-ROSS L C, et al Support vector machine classification of major depressive disorder using diffusion-weighted neuroimaging and graph theory[J]. Frontiers in Psychiatry, 2015, 6: 21
6 KHAZAEE A, EBRAHIMZADEH A, BABAJANI-FEREMI A Application of advanced machine learning methods on resting-state fMRI network for identification of mild cognitive impairment and Alzheimer ’ s disease[J]. Brain Imaging and Behavior, 2016, 10 (3): 799- 817
doi: 10.1007/s11682-015-9448-7
7 KHAZAEE A, EBRAHIMZADEH A, BABAJANI-FEREMI A. Automatic classification of Alzheimer's disease with resting-state fMRI and graph theory [C]// 2014 21th Iranian Conference on Biomedical Engineering (ICBME). Tehran: IEEE, 2014: 252-257.
8 CHENG B, LIU M, SHEN D, et al Multi-domain transfer learning for early diagnosis of Alzheimer ’ s disease[J]. Neuroinformatics, 2017, 15 (2): 115- 132
doi: 10.1007/s12021-016-9318-5
9 LI H, XUE Z, ELLMORE T M, et al. Identification of faulty DTI-based sub-networks in autism using network regularized SVM [C]// 2012 9th IEEE International Symposium on Biomedical Imaging(ISBI). Barcelona: IEEE, 2012: 550-553.
10 DODERO L, MINH H Q, SAN BIAGIO M, et al. Kernel-based classification for brain connectivity graphs on the Riemannian manifold of positive definite matrices [C]// 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI), Brooklyn: IEEE, 2015: 42-45.
11 WEE C Y, YAP P T, SHEN D Diagnosis of autism spectrum disorders using temporally distinct resting-state functional connectivity networks[J]. CNS Neuroscience and Therapeutics, 2016, 22 (3): 212- 219
doi: 10.1111/cns.12499
12 ANDERSON A, DOUGLAS P K, KERR W T, et al Non-negative matrix factorization of multimodal MRI, fMRI and phenotypic data reveals differential changes in default mode subnetworks in ADHD[J]. Neuroimage, 2014, 102: 207- 219
doi: 10.1016/j.neuroimage.2013.12.015
13 MITRA J, SHEN K, GHOSE S, et al Statistical machine learning to identify traumatic brain injury (TBI) from structural disconnections of white matter networks[J]. Neuroimage, 2016, 129: 247- 259
doi: 10.1016/j.neuroimage.2016.01.056
14 RICHIARDI J, GSCHWIND M, SIMIONI S, et al Classifying minimally disabled multiple sclerosis patients from resting state functional connectivity[J]. Neuroimage, 2012, 62 (3): 2021- 2033
doi: 10.1016/j.neuroimage.2012.05.078
15 ARBABSHIRANI M R, KIEHL K, PEARLSON G, et al Classification of schizophrenia patients based on resting-state functional network connectivity[J]. Frontiers in Neuroscience, 2013, 7: 133
16 KENDALL A, GAL Y, CIPOLLA R. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics [C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 7482-7491.
17 MESZLENYI R J, BUZA K, VIDNYANSZKY Z Resting state fMRI functional connectivity-based classification using a convolutional neural network architecture[J]. Frontiers in Neuroinformatics, 2017, 11: 61
doi: 10.3389/fninf.2017.00061
18 PARISOT S, KTENA S I, FERRANTE E, et al. Spectral graph convolutions for population-based disease prediction [C]// International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017.[S.l.]: Springer, 2017: 177-185.
19 PARISOT S, KTENA S I, 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
doi: 10.1016/j.media.2018.06.001
20 HEINSFELD A S, FRANCO A R, CRADDOCK R C, et al Identification of autism spectrum disorder using deep learning and the ABIDE dataset[J]. Neuroimage: Clinical, 2018, 17: 16- 23
doi: 10.1016/j.nicl.2017.08.017
21 CHEN C L P, LIU Z Broad learning system: An effective and efficient incremental learning system without the need for deep architecture[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 29 (1): 10- 24
22 YU W, ZHAO C Broad convolutional neural network based industrial process fault diagnosis with incremental learning capability[J]. IEEE Transactions on Industrial Electronics, 2019, 67 (6): 5081- 5091
23 CHRN C L P, YU D, LIU L Automatic leader-follower persistent formation control for autonomous surface vehicles[J]. IEEE Access, 2018, 7: 12146- 12155
24 WANG J, ZHAO C. Broad learning system based visual fault diagnosis for electrical equipment thermography images [C]// 2018 Chinese Automation Congress (CAC). Xi'an: IEEE, 2018: 1632-1637.
[1] 何立,庞善民. 结合年龄监督和人脸先验的语音-人脸图像重建[J]. 浙江大学学报(工学版), 2022, 56(5): 1006-1016.
[2] 张雪芹,李天任. 基于Cycle-GAN和改进DPN网络的乳腺癌病理图像分类[J]. 浙江大学学报(工学版), 2022, 56(4): 727-735.
[3] 褚晶辉,史李栋,井佩光,吕卫. 适用于目标检测的上下文感知知识蒸馏网络[J]. 浙江大学学报(工学版), 2022, 56(3): 503-509.
[4] 程若然,赵晓丽,周浩军,叶翰辰. 基于深度学习的中文字体风格转换研究综述[J]. 浙江大学学报(工学版), 2022, 56(3): 510-519, 530.
[5] 陈彤,郭剑锋,韩心中,谢学立,席建祥. 基于生成对抗模型的可见光-红外图像匹配方法[J]. 浙江大学学报(工学版), 2022, 56(1): 63-74.
[6] 任松,朱倩雯,涂歆玥,邓超,王小书. 基于深度学习的公路隧道衬砌病害识别方法[J]. 浙江大学学报(工学版), 2022, 56(1): 92-99.
[7] 刘兴,余建波. 注意力卷积GRU自编码器及其在工业过程监控的应用[J]. 浙江大学学报(工学版), 2021, 55(9): 1643-1651.
[8] 陈雪云,黄小巧,谢丽. 基于多尺度条件生成对抗网络血细胞图像分类检测方法[J]. 浙江大学学报(工学版), 2021, 55(9): 1772-1781.
[9] 金立生,华强,郭柏苍,谢宪毅,闫福刚,武波涛. 基于优化DeepSort的前方车辆多目标跟踪[J]. 浙江大学学报(工学版), 2021, 55(6): 1056-1064.
[10] 许佳辉,王敬昌,陈岭,吴勇. 基于图神经网络的地表水水质预测模型[J]. 浙江大学学报(工学版), 2021, 55(4): 601-607.
[11] 王虹力,郭斌,刘思聪,刘佳琪,仵允港,於志文. 边端融合的终端情境自适应深度感知模型[J]. 浙江大学学报(工学版), 2021, 55(4): 626-638.
[12] 张腾,蒋鑫龙,陈益强,陈前,米涛免,陈彪. 基于腕部姿态的帕金森病用药后开-关期检测[J]. 浙江大学学报(工学版), 2021, 55(4): 639-647.
[13] 徐利锋,黄海帆,丁维龙,范玉雷. 基于改进DenseNet的水果小目标检测[J]. 浙江大学学报(工学版), 2021, 55(2): 377-385.
[14] 陈涵娟,达飞鹏,盖绍彦. 基于竞争注意力融合的深度三维点云分类网络[J]. 浙江大学学报(工学版), 2021, 55(12): 2342-2351.
[15] 陈雪云,夏瑾,杜珂. 基于多线型特征增强网络的架空输电线检测[J]. 浙江大学学报(工学版), 2021, 55(12): 2382-2389.