Accurate recognition and classification of depression have important practical significance. Targeting the challenges of low accuracy and poor interpretability in depression identification, a machine recognition model for depression based on a multi-layer dynamic brain function network is proposed. This model overcomes the traditional single linear connection analysis method. By integrating the Pearson correlation coefficient and the maximum information coefficient and using the EEG sliding time window, it constructs a multi-layer dynamic brain function network with both linear and nonlinear dynamic connection characteristics and extracts network features. This model provides more comprehensive brain function feature information and significantly improves the accuracy of depression identification. The model was used to identify normal people and patients with mild, moderate, and severe depression, with an average recognition accuracy of 97.46%. By visualizing the brain network topology and analyzing the node degree, it is found that the occipital lobe, prefrontal lobe and temporal lobe of the brain in depression are abnormal. By analyzing the characteristics of the brain network, it can be seen that the brain beta rhythm plays a key role in the identification of depression, and the global efficiency, average shortest path length and local efficiency of the brain network of patients with mild depression are increased compared with normal people, and the clustering coefficient is reduced. The experimental results show that the model can improve the accuracy of depression identification, provide a new method for clinical diagnosis, depression classification and rapid screening, and provide a new reference for exploring the neurophysiological mechanism of depression.
KENNEDYS H, CENITIA K. Unpacking major depressive disorder: from classification to treatment selection[J]. Canadian Journal of Psychiatry Revue Canadienne de Psychiatrie, 2018, 63(5): 308-313.
SHANGZhaoyan, QIAOXiaoyan. Study on analysis and recognition of EEG characteristics of mild depression[J]. Journal of Test and Measurement Technology, 2022, 36(6): 498-505. (in Chinese)
[4]
LIX, LAR, WANGY, et al. EEG-based mild depression recognition using convolutional neural network[J]. Medical & Biological Engineering & Computing, 2019, 57(6): 1341-1352.
[5]
ZHOUX, WANGX, LIUW, et al. Classification model of depression based on the CNN-LSTM network[C]//2023 3rd International Conference on Frontiers of Electronics, Information and Computation Technologies (ICFEICT), 2023: 210-214.
SHANGZhaoyan, QIAOXiaoyan. Depression recognition with CNN based on feature fusion and attention mechanism[J]. Chinese Journal of Sensors and Actuators, 2024, 37(4): 658-664. (in Chinese)
[8]
ALMEIDAS, CAMACHOM, BARAHONA-CORRÊAJ B, et al. Criterion and construct validity of the Beck Depression Inventory (BDI-II) to measure depression in patients with cancer: The contribution of somatic items[J]. International Journal of Clinical and Health Psychology, 2023, 23(2): 100350.
[9]
HRAMOVA E, FROLOVN S, MAKSIMENKOV A, et al. Functional networks of the brain: from connectivity restoration to dynamic integration[J]. Physics-Uspekhi, 2021, 64(6): 584-616.
[10]
LEEY Y, HSIEHS. Classifying different emotional states by means of EEG-based functional connectivity patterns[J]. PLoS One, 2014, 9(4): e95415.
[11]
WANGZ, TONGY, HENGX. Emotional analysis based on dynamic functional brain network[C]//2019 International Conference on Networking and Network Applications (NaNA), 2019: 202-207.
EEG: Probabilistic Selection and Depression[EB/OL].
[14]
WANGC, CHENJ, ZENGR. An analysis and forecasts of online product sales based on BP Neural Network and Pearson Coefficient[C]//2020 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA), 2020: 559-564.
RENJia, SUNSiyu, BAOKe. Fault diagnosis method in industrial processes based on maximal information coefficient and depth residual graph convolution[J]. Journal of Chemical Engineering of Chinese Universities, 2023, 37(1): 111-119. (in Chinese)
[17]
EK B, VERSCHNEIDERC, NARAYAND A. Global efficiency of graphs[J]. AKCE International Journal of Graphs and Combinatorics, 2015, 12(1): 1-13.
[18]
WANGJ, ZUOX, HEY. Graph-based network analysis of resting-state functional MRI[J]. Frontiers in Systems Neuroscience, 2010, 4: 16.
[19]
GALLENC L, D’ESPOSITOM. Brain modularity: a biomarker of intervention-related plasticity[J]. Trends in Cognitive Sciences, 2019, 23(4): 293-304.
[20]
WANGB, MENGL. Functional brain network alterations in epilepsy: a magnetoencephalography study[J]. Epilepsy Research, 2016, 126: 62-69.
[21]
RUBINOVM, SPORNSO. Complex network measures of brain connectivity: Uses and interpretations[J]. NeuroImage, 2010, 52(3): 1059-1069.
[22]
XIAJ, SHENQ. Correlation analysis of different vegetable categories based on Spearman rank correlation coefficient[C]//2024 IEEE 6th Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC), 2024: 548-551.
[23]
KUMARIK, DEWANGANK K, SAHUR. Convolutional neural network based approach for depression detection through EEG signals[C]//2023 International Conference on Artificial Intelligence for Innovations in Healthcare Industries (ICAIIHI), 2023: 1-7.
WANGYixin, ZHUXiangru, YANGLijun. EEG depression recognition based on feature fusion of common spatial pattern and brain connectivity[J]. Computer Engineering and Applications, 2022, 58(22): 150-158. (in Chinese)
[26]
SUNS, YANC, LYU J, et al. EEG based depression recognition by employing static and dynamic network metrics[C]//2022 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 2022: 1740-1744.