Please wait a minute...
浙江大学学报(工学版)
Journal of ZheJiang University (Engineering Science)  2020, Vol. 54 Issue (4): 694-703    DOI: 10.3785/j.issn.1008-973X.2020.04.008
Computer Technology, Inf ormation Engineering     
Learning effective connectivity network structure based on parallel searching of double firefly populations
Zi-long JI(),Jun-zhong JI*()
Beijing Municipal Key Laboratory of Multimedia and Intelligent Software Technology, Beijing University of Technology, Beijing 100124, China
Download: HTML     PDF(867KB) HTML
Export: BibTeX | EndNote (RIS)      

Abstract  

A new method to learn brain effective connectivity (EC) networks by parallel searching of double firefly populations in order to identify high-quality brain EC network from functional magnetic resonance imaging (fMRI) data. The double population optimization was used to learn brain EC networks. The initial population was divided into an elite population and a common population. Then the brain EC networks were gradually constructed through the directional movements of the elite population and the random movements of the common population. The population sizes of the elite population and the common population were dynamically adjusted in the process of network constructions, and the information exchange between the two populations was realized by using a population migration operation. An adaptive updating mechanism based on a diversity measure was used to dynamically update the two populations after a certain number of evolution iterations. The experimental results on many simulated datasets verify that the new algorithm has obvious advantages on the whole performance compared with other algorithms, and shows the potential practicability of the new algorithm on real datasets.

Key words： brain network      brain effective connectivity network      Bayesian network      firefly algorithm      parallel searching of double populations     
Received: 22 January 2019      Published: 05 April 2020
CLC:  TP 311  
  TP 18  
Corresponding Authors: Jun-zhong JI     E-mail: 403796237@qq.com;jjz01@bjut.edu.cn
Service
E-mail this article
Add to my bookshelf
Add to citation manager
E-mail Alert
RSS
Articles by authors
Zi-long JI
Jun-zhong JI
Cite this article:

Zi-long JI,Jun-zhong JI. Learning effective connectivity network structure based on parallel searching of double firefly populations. Journal of ZheJiang University (Engineering Science), 2020, 54(4): 694-703.

URL:

http://www.zjujournals.com/eng/10.3785/j.issn.1008-973X.2020.04.008     OR     http://www.zjujournals.com/eng/Y2020/V54/I4/694


基于双萤火虫种群并行搜索的脑效应连接网络学习方法

为了从功能磁共振成像（fMRI）数据中识别到高质量的脑效应连接网络，提出双萤火虫种群并行搜索的脑效应连接网络学习方法. 利用2个萤火虫种群寻优完成脑效应连接网络的学习，将萤火虫种群划分为精英种群和普通种群. 通过精英种群的定向移动和普通种群的随机移动，分别逐步构建脑效应连接网络；在网络构建过程中，利用种群迁移操作动态地调整精英种群和普通种群的种群规模，以实现两种群之间的信息交流. 每隔一定代数，使用基于多样性度量的种群自适应更新机制，动态地进行2个种群的更新. 在多组模拟数据上的实验结果表明，新算法与其他算法相比，在识别总体性能上具有明显优势. 在真实数据上的实验表明了算法的潜在实用性.


关键词： 脑网络,  脑效应连接网络,  贝叶斯网,  萤火虫算法,  双种群并行搜索 
Fig.1 Framework of DFA-EC algorithm
Fig.2 Example of directional movement of firefly individual i
Fig.3 Example of random movement of firefly individual i
数据组号 Nnodes Ss/min Ttr /s Nn /% HHRF.std.dev /s 其他因素
1 5 2.5 3.00 1.0 0.5 ?
2 5 5 3.00 1.0 0.5 ?
3 5 60 3.00 1.0 0.5 ?
4 5 10 3.00 1.0 0.5 ?
5 10 10 3.00 1.0 0.5 ?
6 15 10 3.00 1.0 0.5 ?
7 50 10 3.00 1.0 0.5 ?
8 5 10 3.00 1.0 0.5 2组测试
9 5 10 3.00 1.0 0.5 外部输入
Tab.1 Specific parameters of simulated dataset
被试组别 被试数量 男的数量/女的数量 年龄 平均年龄
正常组 50 27/23 70~85 77.6
AD患者组 50 20/30 69~83 74.8
Tab.2 Information of subjects
脑叶 感兴趣区域 编号 脑叶 感兴趣区域 编号
额叶 Frontal_Sup_L（左背外侧额上回） 1 枕叶 Occipital_Mid_L（左枕中回） 19
额叶 Frontal_Sup_R（右背外侧额上回） 2 枕叶 Occipital_Mid_R（右枕中回） 20
额叶 Frontal_Mid_L（左额中回） 3 枕叶 Occipital_Inf_L（左枕下回） 21
额叶 Frontal_Mid_R（右额中回） 4 枕叶 Occipital_Inf_R（右枕下回） 22
额叶 Rectus_L（左回直肌） 5 颞叶 Temporal_Sup_L（左颞上回） 23
额叶 Rectus_R（右回直肌） 6 颞叶 Temporal_Sup_R（右颞上回） 24
额叶 Cingulum_Ant_L（左前扣带和旁扣带脑回） 7 颞叶 Temporal_Mid_L（左颞中回） 25
额叶 Cingulum_Ant_R（右前扣带和旁扣带脑回） 8 颞叶 Temporal_Mid_R（右颞中回） 26
顶叶 Cingulum_Post_L（左后扣带回） 9 颞叶 Temporal_Inf_L（左颞下回） 27
顶叶 Cingulum_Post_R（右后扣带回） 10 颞叶 Temporal_Inf_R（右颞下回） 28
顶叶 Precuneus_L（左楔前叶） 11 颞叶 Hippocampus_L（左海马） 29
顶叶 Precuneus_R（右楔前叶） 12 颞叶 Hippocampus_R（右海马） 30
顶叶 Parietal_Sup_L（左顶上回） 13 颞叶 ParaHippocampal_L（左海马旁回） 31
顶叶 Parietal_Sup_R（右顶上回） 14 颞叶 ParaHippocampal_R（右海马旁回） 32
顶叶 Parietal_Inf_L（左顶下缘角回） 15 颞叶 Angular_L（左角回） 33
顶叶 Parietal_Inf_R（左顶下缘角回） 16 颞叶 Angular_R（右角回） 34
枕叶 Occipital_Sup_L（左枕上回） 17 颞叶 Fusiform_L（左梭状回） 35
枕叶 Occipital_Sup_R（右枕上回） 18 颞叶 Fusiform_R（右梭状回） 36
Tab.3 Selected regions of interest
Fig.4 Comparison of Fd of DFA-EC on Sim1-9 when parameters change
数据组号 算法 Fd t/ms 数据组号 算法 Fd t/ms
1 FAR-EC 1 168 6 FAR-EC 0.81 14 387
1 DFA-EC 1 93 6 DFA-EC 0.82 12 073
2 FAR-EC 0.98 250 7 FAR-EC 0.77 245 154
2 DFA-EC 1 167 7 DFA-EC 0.77 231 212
3 FAR-EC 1 2 177 8（a） FAR-EC 0.96 236
3 DFA-EC 1 1 791 8（a） DFA-EC 1 183
4 FAR-EC 1 384 8（b） FAR-EC 1 230
4 DFA-EC 1 317 8（b） DFA-EC 1 186
5 FAR-EC 0.89 6 135 9 FAR-EC 0.83 396
5 DFA-EC 0.93 5 871 9 DFA-EC 0.92 346
Tab.4 Results of DFA-EC and FAR-EC on the Sim1-9
算法 Fc Fd
Sim1 Sim2 Sim3 Sim1 Sim2 Sim3
GES 1 1 1 0.8 1 0.8
LiNGAM 1 0.91 1 0.6 0.73 1
GC 0.83 1 1 0.33 0.6 0.6
GS 1 1 1 0.6 0.8 1
Patel 1 1 1 0.8 0.98 1
AIAEC 1 1 1 0.96 0.98 1
ACOEC 1 1 1 0.96 0.98 1
DFA-EC 1 1 1 1 1 1
Tab.5 Identification results of eight algorithms on Sim1-3
算法 Fc Fd
Sim4 Sim5 Sim6 Sim7 Sim4 Sim5 Sim6 Sim7
GES 1 1 1 1 0.4 0.64 0.5 0.59
LiNGAM 0.91 0.96 0.95 0.98 0.91 0.87 0.84 0.68
GC 1 1 0.97 0.95 1 0.64 0.65 0.58
GS 1 1 1 1 0.6 0.82 0.89 0.79
Patel 1 1 1 1 0.8 0.82 0.84 0.77
AIAEC 1 1 1 1 1 0.91 0.84 0.75
ACOEC 1 1 1 1 1 0.89 0.86 0.80
DFA-EC 1 1 1 1 1 0.93 0.82 0.77
Tab.6 Identification results of eight algorithms on Sim4-7
算法 Fc Fd
Sim8(a) Sim8(b) Sim9 Sim8(a) Sim8(b) Sim9
GES 1 1 0.91 0.6 1 0.55
LiNGAM 1 1 1 1 1 0.8
GC 0.89 1 0.91 0.89 0.8 0.36
GS 1 1 0.83 0.6 0.6 0.5
Patel 1 1 0.83 0.6 0.8 0.67
AIAEC 1 1 1 0.88 0.92 0.76
ACOEC 1 1 1 0.92 0.88 0.80
DFA-EC 1 1 1 1 1 0.92
Tab.7 Identification results of eight algorithms on Sim8, 9
Fig.5 Effective connectivity network constructed by DFA-EC on real dataset
脑叶 额叶 顶叶 枕叶 颞叶
额叶 9 25 12 29
顶叶 ? 15 14 38
枕叶 ? ? 11 25
颞叶 ? ? ? 40
Tab.8 Connection number in normal person EC networks
脑叶 额叶 顶叶 枕叶 颞叶
额叶 12 23 13 23
顶叶 ? 10 12 39
枕叶 ? ? 10 23
颞叶 ? ? ? 26
Tab.9 Connection number in AD patients EC networks
[1]   HEIN G, MORISHIMA Y, LEIBERG S, et al The brain's functional network architecture reveals human motives[J]. Science, 2016, 351 (6277): 1074
doi: 10.1126/science.aac7992
[2]   FRISTON K J Functional and effective connectivity in neuroimaging: a synthesis[J]. Human Brain Mapping, 2011, 1 (1): 13
[3]   GOLDENBERG D, GALVáN A The use of functional and effective connectivity techniques to understand the developing brain[J]. Developmental Cognitive Neuroscience, 2015, 12: 155- 164
doi: 10.1016/j.dcn.2015.01.011
[4]   颜志雄, 刘勋, 谭淑平, 等 发展认知神经科学: 人脑毕生发展的功能连接组学时代[J]. 科学通报, 2016, (7): 718- 727
YAN Zhi-xiong, LIU Xun, TAN Shu-ping, et al Developmental cognitive neurosciecne: functional connectom- ics agenda for human brain lifespan development (in Chinese)[J]. Chinese Science Bulletin, 2016, (7): 718- 727
[5]   CHIANG S, GUINDANI M, YEH H J, et al Bayesian vector autoregressive model for multi-subject effective connectivity inference using multi-modal neuroimaging data[J]. Human Brain Mapping, 2017, 38 (3): 1311- 1332
doi: 10.1002/hbm.23456
[6]   赵学武. 基于群智能算法的人脑功能划分方法研究[D]. 北京: 北京工业大学, 2018.
ZHAO Xue-wu. Brain functional parcellation on methods based on swarm intelligence algorithm [D]. Beijing: Beijing University of Technology, 2018.
[7]   IDE J S, ZHANG S, LI C S R Bayesian network models in brain functional connectivity analysis[J]. International Journal of Approximate Reasoning Official Publication of the North American Fuzzy Information Processing Society, 2014, 55 (1): 23- 35
[8]   JI J, LIU J, LIANG P, et al Learning effective connectivity network structure from fMRI data based on artificial immune algorithm[J]. Plos One, 2016, 11 (4): e0152600
doi: 10.1371/journal.pone.0152600
[9]   LIU J, JI J, ZHANG A, et al. An ant colony optimization algorithm for learning brain effective connectivity network from fMRI data [C] // IEEE International Conference on Bioinformatics and Biomedicine. Shenzhen: IEEE, 2016: 360-367.
[10]   纪子龙, 冀俊忠, 刘金铎, 等. 基于萤火虫算法的脑效应连接网络学习方法[J]. 哈尔滨工业大学学报, 2019, 51(5): 82-90.
JI Zi-long, JI Jun-zhong, LIU Jin-duo, et al. Learning effective connectivity network structure based on firefly algorithm [J]. Journal of Harbin Institute of Technology, 2019, 51(5): 82-90.
[11]   SMITH S M, MILLER K L, SALIMI-KHORSHIDI G, et al Network modeling methods for FMRI[J]. Neuroimage, 2011, 54 (2): 875- 891
doi: 10.1016/j.neuroimage.2010.08.063
[12]   HUANG S, LI J, YE J, et al A sparse structure learning algorithm for Gaussian Bayesian network identification from high-dimensional data[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013, 35 (6): 1328- 1342
doi: 10.1109/TPAMI.2012.129
[13]   CHICKERING D M Optimal structure identification with greedy search[J]. The Journal of Machine Learning Research, 2003, 3 (3): 507- 554
[14]   XU L, FAN T, WU X, et al A pooling-LiNGAM algorithm for effective connectivity analysis of fMRI data[J]. Frontiers in Computational Neuroscience, 2014, 8 (8): 125
[15]   BARNETT L, SETH A K The MVGC multivariate Granger causality toolbox: a new approach to Granger-causal inference[J]. Journal of Neuroscience Methods, 2014, 223: 50- 68
doi: 10.1016/j.jneumeth.2013.10.018
[16]   DAUWELS J, VIALATTE F, MUSHA T, et al A comparative study of synchrony measures for the early diagnosis of Alzheimer's disease based on EEG[J]. Neuroimage, 2010, 49 (1): 668- 693
doi: 10.1016/j.neuroimage.2009.06.056
[17]   PATEL R S, BOWMAN D B, RILLING J K A Bayesian approach to determining connectivity of the human brain[J]. Human Brain Mapping, 2010, 27 (3): 267- 276
[18]   WU X, LI R, FLEISHER A S, et al Altered default mode network connectivity in Alzheimer’s disease: aresting functional MRI and Bayesian network study[J]. Human Brain Mapping, 2011, 32 (11): 1868- 1881
doi: 10.1002/hbm.21153
[19]   韩利坤, 姚婧璠, 李越秀, 等 阿尔茨海默病静息态默认模式网络研究[J]. 中国医刊, 2014, (9): 46- 50
HAN Li-kun, YAO Jing-pan, LI Yue-xiu, et al Research the default in patients with Alzheimer's disease at resting state[J]. Chinese Journal of Medicine, 2014, (9): 46- 50
doi: 10.3969/j.issn.1008-1070.2014.09.016
[20]   GROSSMAN M, KOENIG P, GLOSSER G, et al Neural basis for semantic memory difficulty in Alzheimer's disease: an fMRI study[J]. Brain A Journal of Neurology, 2003, 126 (2): 292- 311
doi: 10.1093/brain/awg027
[21]   周鹏, 朱浩泽, 倪红艳, 等 阿尔茨海默症患者额顶叶控制系统脑网络异常[J]. 科学通报, 2016, 61 (32): 3490
ZHOU Peng, ZHOU Hao-ze, NI Hong-yan, et al The abnorm- ality of frontal-parietal control system in Alzheimer’s disease[J]. Chinese Science Bulletin, 2016, 61 (32): 3490
doi: 10.1360/N972015-01273
[1] Hong-hui WANG,Xin FANG,De-jiang LI,Gui-jie LIU. Fatigue crack growth prediction method under variable amplitude load based on dynamic Bayesian network[J]. Journal of ZheJiang University (Engineering Science), 2021, 55(2): 280-288.
[2] Xue-qin ZHANG,Li ZHANG,Chun-hua GU. Quantitative assessment of social engineering threat in social network[J]. Journal of ZheJiang University (Engineering Science), 2019, 53(5): 837-842.
[3] XU Zhe, XIONG Xiao-feng, HONG Jia-ming, HE Bi-shi, CHEN Yun. Data-driven abnormal event detection method for urban water supply network[J]. Journal of ZheJiang University (Engineering Science), 2017, 51(11): 2222-2231.
[4] ZHAO Jian-jun, WANG Yi, YANG Li-bin. Threat assessment method based on time series forecast[J]. Journal of ZheJiang University (Engineering Science), 2014, 48(3): 398-403.
[5] GUO Tong,LIN Feng. Bayesian network structure learning based on hybrid genetic
and fish swarm algorithm[J]. Journal of ZheJiang University (Engineering Science), 2014, 48(1): 130-135.