浏览全部资源
扫码关注微信
纸质出版日期： 2024-11-16 ，
移动端阅览
龙丹， 章梦达， 应仁辉， 陈丰农， 邵岚， 谢璩， 罗聪. 2024. 融合遗传算法和图神经网络的抑郁症智能诊断. 中国图象图形学报， 29(11):3476-3486
Long Dan， Zhang Mengda， Ying Renhui， Chen Fengnong， Shao Lan， Xie Qu， Luo Cong. 2024. Intelligent diagnosis of depression by integrating genetic algorithm and graph neural network. Journal of Image and Graphics， 29(11):3476-3486
目的
2
构建基于脑网络的抑郁症智能诊断模型是一项具有挑战性的任务。近年来，图神经网络（graph neural network，GNN）越来越多地应用于图的分类任务中，大部分GNN研究都只是对单一空间（样本空间或者特征空间）进行建模，导致模型分类性能不够好，本文提出一种基于遗传算法（genetic algorithm，GA）和GNN的多空间融合算法来对抑郁症患者进行智能诊断。
方法
2
模型采用留一站点交叉验证来确保模型的泛化性。脑网络的构建是基于Pearson相关的功能连接方法。整个算法以遗传算法作为主要框架，其中适应度函数是基于图卷积网络（graph convolutional network， GCN）分类算法，通过搜索个体间相似性阈值来找到具有最高分辨力的GCN。GCN由两个网络串联组成，一个网络获取受试者特征空间信息，另一个提取受试者之间样本空间的信息，最后通过两级GCN的联合学习实现分类。
结果
2
所有数据来源于The REST-meta-MDD项目，一共有来自10个站点1 160个受试者功能磁共振数据纳入本实验（男434、女726）。实验结果显示，本文提出的分类器准确性、精度和受试者特征曲线（receiver operating characteristic，ROC）下面积分别为64.72%、69.69%和64.58%，优于其他主流算法。
结论
2
与其他算法相比，本文提出的算法融合了传统模型和深度学习模型的优点，获得了最佳的分类性能，未来很有可能为临床抑郁症诊断提供重要依据。
Objective
2
Depression is currently one of the most common neuropsychiatric disorders in the world. However， its pathophysiological mechanisms are still unclear. The diagnosis of depression in clinical practice typically depends on neuropsychological scores and treatment responses， lacking objective evaluation tools， resulting in low consistency in diagnosis. In recent years， an increasing number of people have begun to use machine learning technology to extract imaging biomarkers for the intelligent diagnosis of depression due to the capability of functional magnetic resonance imaging to provide in vivo brain function and structural information. The brain network-based model has remarkable potential as an imaging marker for effectively distinguishing depression from normal controls. Graph neural networks （GNNs） are highly suitable for graph classification tasks because they directly acquire graph structure information and maintain the topological characteristics of the graph during task execution. However， most GNN studies only model a single space （sample or feature space）， and the aggregation of GNN information can lead to over-smooth effects， resulting in poor model classification performance. This study aims to integrate multiple feature space information and propose a multispace fusion algorithm for the intelligent diagnosis of depression patients.
Method
2
Leave-one-site cross-validation （LOSCV） is used to ensure the generalization of the model. The data are first preprocessed， and then a brain network is constructed using Pearson-related functional connectivity methods. The entire algorithm is mainly based on a genetic algorithm （GA）， where the fitness function is a classification algorithm based on a graph convolutional network （GCN）. The solution space searched by GA is the similarity between the subject networks. The main steps of GA are as follows： 1） Set the search range of the solution space ［0.05， 0.7］； 2） Generate an initial population； 3） Based on LOSCV， GCN is used to classify the data， with the F1 value as the target value of the fitness function， and the threshold with the best fitness is finally retained （representing A *）； 4） Generate new populations through selection， crossover， and mutation operations （representing A）； 5） Compare A with A *. If the fitness value of A is better than A *， then A replaces A *； 6） Determine whether the number of iterations for updating the population has reached the preset value. If not， then proceed to step 3 and continue executing the algorithm； if the threshold is reached， then the algorithm ends. The GA has a chromosome length of 8 bits and a threshold of 20 iterations. This paper aims to determine the similarity threshold between individuals with the highest classification capability in the population network. The GCN module comprises two networks connected in series： one mainly obtains information regarding the feature space of the brain network of a single subject， while the other network takes the subject as a node in the network. All subjects form a network to extract information from the sample space. The classification of a certain subject can be achieved through the joint learning of two levels of GCN. The two-level GCN architecture mainly includes f-GCN and p-GCN， and the basic ideas for constructing each architecture are as follows： f-GCN is a potential information representation for learning the connectivity relationships of each brain region and transforming it into a highly efficient information representation for each brain network. F-GCN uses GCN to learn the embedding representation of a single brain region and then uses Eigenpooling to embed all brain region nodes into a single supernode to represent the information representation of the entire brain network. Eigenpooling is a pooling method in graph convolution neural network （GCN）， which uses the eigenvectors of the Laplacian matrix to represent the information of nodes， transforms the original graph nodes into coordinates in the feature space， and associates each node with a specific number of high-energy eigenvectors， which are determined by the eigenvalues of the Laplacian matrix. The feature vector represents the position of a node in the feature space， and its corresponding feature values indicate node importance. P-GCN constructs a topological structure based on the relationship between subject brain networks and the representation of graph information acquired by f-GCN. The graph convolutional kernel aggregates the information representations of adjacent node entities of the subject and further reduces the dimensionality of the node information representation through graph pooling to generate the current supernode information representation. In this case， the hypernode represents the information of the entity as a whole. The graph information of the entire subject can be accurately represented through this super node， and the parameters of the f-GCN and p-GCN can be jointly updated through backpropagation to improve recognition accuracy. A scaled exponential similarity kernel is used for p-GCN to determine the similarity between samples.
Result
2
All data came from the REST-meta-MDD project， and a total of 1160 functional magnetic resonance imaging data from 10 sites （male 434， female 726） were included in this experiment. The experiment is a comparison of four representative algorithms of different types. The algorithm achieved the highest accuracy of 64.27%， which is 4.47% higher than the second-place support vector machine （SVM）. Based on the BrainNetCNN method， the accuracy is only 56.69%， demonstrating the worst classification performance. The accuracy of the Graphormer is 57.43%， and the hierarchical GCN also adopts the fusion of two networks， resulting in a classification accuracy of 58.28%. The sample similarity threshold also impacts the final result， with an interval of 0.4–0.5 during identification of the optimal solution.
Conclusion
2
The intelligent diagnosis framework for depression based on GA and GCN proposed in this article combines the advantages of traditional and deep learning models. The results show that the proposed algorithm is not only superior to traditional machine learning algorithms （such as SVM）， but also better than several mainstream GCN algorithms， with good generalization. This algorithm is likely to provide important information for clinical depression diagnosis in the future.
抑郁症图卷积网络（GCN）智能诊断融合算法个体相似性
major disorder depressiongraph convolutional network（GCN）intelligent diagnosisfusion algorithmindividual similarity
An Y， Qu Z， Xu N and Nima Z X. 2020. Automatic depression estimation using facial appearance. Journal of Image and Graphics， 25（11）： 2415-2427
安昳， 曲珍， 许宁， 尼玛扎西. 2020. 面部动态特征描述的抑郁症识别. 中国图象图形学报， 25（11）： 2415-2427 ［DOI： 10.11834/jig.200322http://dx.doi.org/10.11834/jig.200322］
Banka A， Buzi I and Rekik I. 2020. Multi-view brain HyperConnectome AutoEncoder for brain state classification//Proceedings of the 3rd International Workshop on Predictive Intelligence in Medicine. Lima， Peru： Springer： 101-110 ［DOI： 10.1007/978-3-030-59354-4_10http://dx.doi.org/10.1007/978-3-030-59354-4_10］
Banka A and Rekik I. 2019. Adversarial connectome embedding for mild cognitive impairment identification using cortical morphological networks//Proceedings of the 3rd International Workshop on Connectomics in Neuroimaging. Shenzhen， China： Springer： 74-82 ［DOI： 10.1007/978-3-030-32391-2_8http://dx.doi.org/10.1007/978-3-030-32391-2_8］
Bessadok A， Mahjoub M A and Rekik I. 2023. Graph neural networks in network neuroscience. IEEE Transactions on Pattern Analysis and Machine Intelligence， 45（5）： 5833-5848 ［DOI： 10.1109/TPAMI.2022.3209686http://dx.doi.org/10.1109/TPAMI.2022.3209686］
Drysdale A T， Grosenick L， Downar J， Dunlop K， Mansouri F， Meng Y， Fetcho R N， Zebley B， Oathes D J Etkin A， Schatzberg A F， Sudheimer K， Keller J， Mayberg H S， Gunning F M， Alexopoulos G S， Fox M D， Pascual-Leone A， Voss H U， Casey B J， Dubin M J and Liston C. 2017. Resting-state connectivity biomarkers define neurophysiological subtypes of depression. Nature Medicine， 23（1）： 28-38 ［DOI： 10.1038/nm.4246http://dx.doi.org/10.1038/nm.4246］
Fang P， Zeng L L， Shen H， Wang L B， Li B J， Liu L and Hu D W. 2012. Increased cortical-limbic anatomical network connectivity in major depression revealed by diffusion tensor imaging. PLoS One， 7（9）： #e45972 ［DOI： 10.1371/journal.pone.0045972http://dx.doi.org/10.1371/journal.pone.0045972］
Gao S， Calhoun V D and Sui J. 2018. Machine learning in major depression： from classification to treatment outcome prediction. CNS Neuroscience and Therapeutics， 24（11）： 1037-1052 ［DOI： 10.1111/cns.13048http://dx.doi.org/10.1111/cns.13048］
GBD 2019 Stroke Collaborators. 2021. Global， regional， and national burden of stroke and its risk factors， 1990-2019： a systematic analysis for the global burden of disease study 2019. The Lancet Neurology， 20（10）： 795-820 ［DOI： 10.1016/S1474-4422（21）00252-0http://dx.doi.org/10.1016/S1474-4422（21）00252-0］
He H， Sui J， Du Y H， Yu Q B， Lin D D， Drevets W C， Savitz J B， Yang J， Victor T A and Calhoun V D. 2017. Co-altered functional networks and brain structure in unmedicated patients with bipolar and major depressive disorders. Brain Structure and Function， 222（9）： 4051-4064 ［DOI： 10.1007/s00429-017-1451-xhttp://dx.doi.org/10.1007/s00429-017-1451-x］
Huang J S， Zhou L P， Wang L and Zhang D Q. 2020. Attention-diffusion-bilinear neural network for brain network analysis. IEEE Transactions on Medical Imaging， 39（7）： 2541-2552 ［DOI： 10.1109/TMI.2020.2973650http://dx.doi.org/10.1109/TMI.2020.2973650］
Kahali S， Raichle M E and Yablonskiy D A. 2021. The role of the human brain neuron-glia-synapse composition in forming resting-state functional connectivity networks. Brain Sciences， 11（12）： #1565 ［DOI： 10.3390/brainsci11121565http://dx.doi.org/10.3390/brainsci11121565］
Kawahara J， Brown C J， Miller S P， Booth B G， Chau V， Grunau R E， Zwicker J G and Hamarneh G. 2017. BrainNetCNN： convolutional neural networks for brain networks； towards predicting neurodevelopment. NeuroImage， 146： 1038-1049 ［DOI： 10.1016/j.neuroimage.2016.09.046http://dx.doi.org/10.1016/j.neuroimage.2016.09.046］
Kipf T N and Welling M. 2017. Semi-supervised classification with graph convolutional networks//Proceedings of the 5th International Conference on Learning Representations. Toulon， France： ICLR
Parisot S， Ktena S I， Ferrante E， Lee M， Moreno R G， Glocker B and Rueckert D. 2017. Spectral graph convolutions for population-based disease prediction//Proceedings of the 20th International Conference on Medical Image Computing and Computer-Assisted Intervention. Quebec City， Canada： Springer： 177-185 ［DOI： 10.1007/978-3-319-66179-7_21http://dx.doi.org/10.1007/978-3-319-66179-7_21］
Regier D A， Narrow W E， Clarke D E， Kraemer H C， Kuramoto S J， Kuhl E A and Kupfer D J. 2013. DSM-5 field trials in the United States and Canada， Part II： test-retest reliability of selected categorical diagnoses. The American Journal of Psychiatry， 170（1）： 59-70 ［DOI： 10.1176/appi.ajp.2012.12070999http://dx.doi.org/10.1176/appi.ajp.2012.12070999］
Tan K W， Huang W X， Liu X F， Hu J L and Dong S B. 2021. A hierarchical graph convolution network for representation learning of gene expression data. IEEE Journal of Biomedical and Health Informatics， 25（8）： 3219-3229 ［DOI： 10.1109/JBHI.2021.3052008http://dx.doi.org/10.1109/JBHI.2021.3052008］
Wang B， Mezlini A M， Demir F， Fiume M， Tu Z W， Brudno M， Haibe-Kains B and Goldenberg A. 2014. Similarity network fusion for aggregating data types on a genomic scale. Nature Methods， 11（3）： 333-337 ［DOI： 10.1038/nmeth.2810http://dx.doi.org/10.1038/nmeth.2810］
Wang X， Ren Y S and Zhang W S. 2017. Depression disorder classification of fMRI data using sparse low-rank functional brain network and graph-based features. Computational and Mathematical Methods in Medicine， 2017： #3609821 ［DOI： 10.1155/2017/3609821http://dx.doi.org/10.1155/2017/3609821］
Yan C G， Chen X， Li L， Castellanos F X， Bai T J， Bo Q J， Cao J， Chen G M， Chen N X， Chen W， Cheng C， Cheng Y Q， Cui X L， Duan J， Fang Y R， Gong Q Y， Guo W B， Hou Z H， Hu L， Kuang L， Li F， Li K M， Li T， Liu Y S， Liu Z N， Long Y C， Luo Q H， Meng H Q， Peng D H， Qiu H T， Qiu J， Shen Y D， Shi Y S， Wang C Y， Wang F， Wang K， Wang L， Wang X， Wang Y， Wu X P， Wu X R， Xie C M， Xie G R， Xie H Y， Xie P， Xu X F， Yang H， Yang J， Yao J S， Yao S Q， Yin Y Y， Yuan Y G， Zhang A X， Zhang H， Zhang K R， Zhang L， Zhang Z J， Zhou R B， Zhou Y T， Zhu J J， Zou C J， Si T M， Zuo X N， Zhao J P and Zang Y F. 2019. Reduced default mode network functional connectivity in patients with recurrent major depressive disorder. Proceedings of the National Academy of Sciences of the United States of America， 116（18）： 9078-9083 ［DOI： 10.1073/pnas.1900390116http://dx.doi.org/10.1073/pnas.1900390116］
Yang H Z， Li X X， Wu Y F， Li S Y， Lu S， Duncan J S， Gee J C and Gu S. 2019. Interpretable multimodality embedding of cerebral cortex using attention graph network for identifying bipolar disorder//Proceedings of the 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention. Shenzhen， China： Springer： 799-807 ［DOI： 10.1007/978-3-030-32248-9_89http://dx.doi.org/10.1007/978-3-030-32248-9_89］
Yin P Y. 1999. A fast scheme for optimal thresholding using genetic algorithms. Signal Processing， 72（2）： 85-95 ［DOI： 10.1016/S0165-1684（98）00167-4http://dx.doi.org/10.1016/S0165-1684（98）00167-4］
Ying C X， Cai T L， Luo S J， Zheng S X， Ke G L， He D， Shen Y M and Liu T Y. 2021. Do Transformers really perform badly for graph representation?//Proceedings of the 34th Annual Conference on Neural Information Processing Systems. 28877-28888
Yoshida K， Shimizu Y， Yoshimoto J， Takamura M， Okada G， Okamoto Y， Yamawaki and Doya K. 2017. Prediction of clinical depression scores and detection of changes in whole-brain using resting-state functional MRI data with partial least squares regression. PLoS One， 12（7）： #e0179638 ［DOI： 10.1371/journal.pone.0179638http://dx.doi.org/10.1371/journal.pone.0179638］
Zeng L L， Shen H， Liu L， Wang L B， Li B J， Fang P， Zhou Z T， Li Y M and Hu D W. 2012. Identifying major depression using whole-brain functional connectivity： a multivariate pattern analysis. Brain， 135（5）： 1498-1507 ［DOI： 10.1093/brain/aws059http://dx.doi.org/10.1093/brain/aws059］
Zhang Y F and Huang H. 2019. New graph-blind convolutional network for brain connectome data analysis//Proceedings of the 26th International Conference on Information Processing in Medical Imaging. Hong Kong， China： Springer： 669-681 ［DOI： 10.1007/978-3-030-20351-1_52http://dx.doi.org/10.1007/978-3-030-20351-1_52］
Zhong X， Shi H Q， Ming Q S， Dong D F， Zhang X C， Zeng L L and Yao S Q. 2017. Whole-brain resting-state functional connectivity identified major depressive disorder： a multivariate pattern analysis in two independent samples. Journal of Affective Disorders， 218： 346-352 ［DOI： 10.1016/j.jad.2017.04.040http://dx.doi.org/10.1016/j.jad.2017.04.040］
Zhou Z， Chen X B， Zhang Y， Hu D， Qiao L S， Yu R P， Yap P T， Pan G， Zhang H and Shen D G. 2020. A toolbox for brain network construction and classification （BrainNetClass）. Human Brain Mapping， 41（10）： 2808-2826 ［DOI： 10.1002/hbm.24979http://dx.doi.org/10.1002/hbm.24979］
相关作者
相关机构
京公网安备11010802024621