Research on weighted hypergraph attention neural network for the diagnosis of psychiatric disorders using brain functional connectivity networks_Journal of Biomedical Engineering

Authors：

LIN Xiuwei ^1,2 , WANG Zhifeng ^1,2 , YAN Haotao ^1,2 , ZENG Siao ^1,2 ,  WANG Peizhou ³ , WU Zetao ^1,2 , YU Xin ^1,2 , HAN Junxi ^1,2

1. School of Mechatronic Engineering and Automation, Foshan University, Foshan, Guangdong 528000, P. R. China;
2. Guangdong Provincial Key Laboratory of Industrial Intelligent Inspection Technology, Foshan, Guangdong 528000, P. R. China;
3. Dermatology Hospital of Southern Medical University, Guangzhou 510091, P. R. China;

Corresponding?author：

WANG Peizhou, Email: 16676756604@163.com

Keywords：

Brain functional connectivity networks; Weighted hyperedge; Hypergraph neural network; Phase-amplitude coupling; Psychiatric diseases classification; Message passing mechanism

DOI：

10.7507/1001-5515.202510012

Video：

Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables Video References Cited by

The hypergraph neural network (HGNN) has demonstrated efficacy in modeling high-order interactions among brain regions, thus providing a promising framework for analyzing brain functional connectivity networks in the context of psychiatric research. The present study proposes a phase-amplitude coupling-weighted hypergraph attention neural network (PAC-HyperGAT) model for the diagnosis of psychiatric diseases. The proposed methodology first constructs a functional hypergraph using elastic net-based sparse regression and then assigns physiologically meaningful weights to hyperedges by quantifying the phase-amplitude coupling strength among nodes within each hyperedge. In light of these findings, the present study proposes a novel hypergraph attention convolution kernel. The efficacy of this approach is evidenced by its enhancement of the node-level message passing mechanism, a feat that facilitates the integration of hyperedge weight information. This phenomenon, in turn, results in an enhancement of the discriminative ability of brain functional connectivity network representations. The proposed model is systematically evaluated on publicly available electroencephalogram datasets for attention deficit hyperactivity disorder (ADHD) and major depressive disorder (MDD). The experimental results demonstrate that PAC-HyperGAT attains an accuracy of (72.14 ± 9.19) % in ADHD classification, surpassing the performance of existing brain functional connectivity network methods across a range of evaluation metrics. The model exhibits notable efficacy in MDD classification, signifying substantial cross-disorder generalization capabilities. Furthermore, PAC-HyperGAT has demonstrated efficacy in identifying brain regions associated with these disorders. In summary, the proposed model demonstrates excellent generalizability, robustness, and neurobiological interpretability, providing a reliable analytical framework for objective diagnosis and mechanistic investigation of psychiatric diseases.

Citation： LIN Xiuwei, WANG Zhifeng, YAN Haotao, ZENG Siao, WANG Peizhou, WU Zetao, YU Xin, HAN Junxi. Research on weighted hypergraph attention neural network for the diagnosis of psychiatric disorders using brain functional connectivity networks. Journal of Biomedical Engineering, 2026, 43(2): 319-327. doi: 10.7507/1001-5515.202510012 Copy

1.	Vigo D, Thornicroft G, Atun R. Estimating the true global burden of mental illness. Lancet Psychiatry, 2016, 3(2): 171-178.
2.	Wang Z, Dou Y, Yang X, et al. Global, regional, and national burden of mental disorders among adolescents and young adults, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Transl Psychiatry, 2025, 15(1): 397.
3.	Rahimi Saryazdi A, Bayani A, Ghassemi F, et al. Brain functional connectivity network during deception: a visibility graph approach. Eur Phys J Spec Top, 2025, 234(6): 937-950.
4.	Khaleghi N, Hashemi S, Peivandi M, et al. EEG-based functional connectivity analysis of brain abnormalities: a systematic review study. Inform Med Unlocked, 2024, 47: 101476.
5.	劉晨旭. 基于圖卷積神經網絡的腦網絡特征分析及應用. 太原: 太原師范學院, 2023.
6.	Ling Q, Liu A, Li Y, et al. High-order graphical topology analysis of brain functional connectivity networks using fMRI. IEEE Trans Neural Syst Rehabil Eng, 2025, 33: 1611-1620.
7.	Scarselli F, Gori M, Tsoi A C, et al. The graph neural network model. IEEE Trans Neural Netw, 2009, 20(1): 61-80.
8.	李松, 伏云發, 張妍, 等. 基于圖卷積神經網絡和腦電信號的疲勞識別研究. 生物醫學工程學雜志, 2025, 42(4): 686-692.
9.	侯濤, 丁衛平, 黃嘉爽, 等. DE-NNs: 基于動態證據神經網絡的腦網絡分析算法. 計算機研究與發展, 2025, 62(4): 888-904.
10.	Li X, Zhou Y, Dvornek N, et al. BrainGNN: interpretable brain graph neural network for fMRI analysis. Med Image Anal, 2021, 74: 102233.
11.	Gamgam G, Kabakcioglu A, Yüksel Dal D, et al. Disentangled attention graph neural network for Alzheimer’s disease diagnosis//International Conference on Medical Image Computing and Computer-Assisted Intervention. Marrakesh: MICCAI, 2024: 219-228.
12.	Wang C, Xiao Z, Xu Y, et al. A novel approach for ASD recognition based on graph attention networks. Front Comput Neurosci, 2024, 18: 1388083.
13.	Li Y, Liu J, Gao X, et al. Multimodal hyper-connectivity of functional networks using functionally-weighted LASSO for MCI classification. Med Image Anal, 2019, 52: 80-96.
14.	Eavani H, Satterthwaite T D, Filipovych R, et al. Identifying sparse connectivity patterns in the brain using resting-state fMRI. NeuroImage, 2015, 105: 286-299.
15.	Bai S, Zhang F, Torr P H S. Hypergraph convolution and hypergraph attention. Pattern Recognit, 2021, 110: 107637.
16.	Gao Y, Zhang Z, Lin H, et al. Hypergraph learning: methods and practices. IEEE Trans Pattern Anal Mach Intell, 2022, 44(5): 2548-2566.
17.	孔文文. 基于靜息態腦電數據的高階腦網絡分析和抑郁識別研究. 蘭州: 蘭州大學, 2023.
18.	Liu J, Cui W, Chen Y, et al. Deep fusion of multi-template using spatio-temporal weighted multi-hypergraph convolutional networks for brain disease analysis. IEEE Trans Med Imaging, 2024, 43(2): 860-873.
19.	Jie B, Wee C Y, Shen D, et al. Hyper-connectivity of functional networks for brain disease diagnosis. Med Image Anal, 2016, 32: 84-100.
20.	Guo H, Li Y, Xu Y, et al. Resting-state brain functional hyper-network construction based on elastic net and group lasso methods. Front Neuroinform, 2018, 12: 25.
21.	Lostar M, Rekik I. Deep hypergraph u-net for brain graph embedding and classification. arXiv, 2020: 2008.13118.
22.	Wang J, Li H, Qu G, et al. Dynamic weighted hypergraph convolutional network for brain functional connectome analysis. Med Image Anal, 2023, 87: 102828.
23.	Xiao L, Wang J, Kassani P H, et al. Multi-hypergraph learning-based brain functional connectivity analysis in fMRI data. IEEE Trans Med Imaging, 2020, 39(5): 1746-1758.
24.	Xu B, Dall’Aglio L, Flournoy J, et al. Limited generalizability of multivariate brain-based dimensions of child psychiatric symptoms. Commun Psychol, 2024, 2(1): 16.
25.	Lu J, Li A, Li K, et al. An EEG study on β?γ phase-amplitude coupling-based functional brain network in epilepsy patients. IEEE J Biomed Health Inform, 2024, 28(6): 3446-3456.
26.	Alim A, Imtiaz M H. Automatic identification of children with ADHD from EEG brain waves. Signals, 2023, 4(1): 193-205.
27.	Cai H, Gao Y, Sun S, et al. Modma dataset: a multi-modal open dataset for mental-disorder analysis. arXiv, 2020: 2002.09283.
28.	Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods, 2004, 134(1): 9-21.
29.	Tadel F, Baillet S, Mosher J C, et al. Brainstorm: a user-friendly application for MEG/EEG analysis. Comput Intell Neurosci, 2011, 2011: 879716.
30.	Desikan R S, Ségonne F, Fischl B, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 2006, 31(3): 968-980.
31.	Brookshire G, Kasper J, Blauch N M, et al. Data leakage in deep learning studies of translational EEG. Front Neurosci, 2024, 18: 1373515.
32.	Jakkula V. Tutorial on support vector machine (SVM). Pullman: Washington State University, 2006.
33.	Moon S E, Jang S, Lee J S. Convolutional neural network approach for EEG-based emotion recognition using brain connectivity and its spatial information//2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Calgary: IEEE, 2018: 2556-2560.
34.	Veli?kovi? P, Cucurull G, Casanova A, et al. Graph attention networks. arXiv, 2017: 1710.10903.
35.	Selvaraju R R, Cogswell M, Das A, et al. Grad-cam: visual explanations from deep networks via gradient-based localization//Proceedings of the IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 618-626.
36.	Latifi B, Amini A, Nasrabadi A M. Siamese based deep neural network for ADHD detection using EEG signal. Comput Biol Med, 2024, 182: 109092.
37.	Hu Y, Ran J, Qiao R, et al. Identifying ADHD-related abnormal functional connectivity with a graph convolutional neural network. Neural Plast, 2024, 2024: 8862647.
38.	Scalabrini A, Poletti S, Vai B, et al. Abnormally slow dynamics in occipital cortex of depression. J Affect Disord, 2025, 374: 523-530.
39.	Chao Z C, Dillon D G, Liu Y H, et al. Altered coordination between frontal delta and parietal alpha networks underlies anhedonia and depressive rumination in major depressive disorder. J Psychiatry Neurosci, 2022, 47(6): E367-E378.
40.	Zhao Y, Chen L, Zhang W, et al. Gray matter abnormalities in non-comorbid medication-naive patients with major depressive disorder or social anxiety disorder. EBioMedicine, 2017, 21: 228-235.

1. Vigo D, Thornicroft G, Atun R. Estimating the true global burden of mental illness. Lancet Psychiatry, 2016, 3(2): 171-178.
2. Wang Z, Dou Y, Yang X, et al. Global, regional, and national burden of mental disorders among adolescents and young adults, 1990–2021: a systematic analysis for the Global Burden of Disease Study 2021. Transl Psychiatry, 2025, 15(1): 397.
3. Rahimi Saryazdi A, Bayani A, Ghassemi F, et al. Brain functional connectivity network during deception: a visibility graph approach. Eur Phys J Spec Top, 2025, 234(6): 937-950.
4. Khaleghi N, Hashemi S, Peivandi M, et al. EEG-based functional connectivity analysis of brain abnormalities: a systematic review study. Inform Med Unlocked, 2024, 47: 101476.
5. 劉晨旭. 基于圖卷積神經網絡的腦網絡特征分析及應用. 太原: 太原師范學院, 2023.
6. Ling Q, Liu A, Li Y, et al. High-order graphical topology analysis of brain functional connectivity networks using fMRI. IEEE Trans Neural Syst Rehabil Eng, 2025, 33: 1611-1620.
7. Scarselli F, Gori M, Tsoi A C, et al. The graph neural network model. IEEE Trans Neural Netw, 2009, 20(1): 61-80.
8. 李松, 伏云發, 張妍, 等. 基于圖卷積神經網絡和腦電信號的疲勞識別研究. 生物醫學工程學雜志, 2025, 42(4): 686-692.
9. 侯濤, 丁衛平, 黃嘉爽, 等. DE-NNs: 基于動態證據神經網絡的腦網絡分析算法. 計算機研究與發展, 2025, 62(4): 888-904.
10. Li X, Zhou Y, Dvornek N, et al. BrainGNN: interpretable brain graph neural network for fMRI analysis. Med Image Anal, 2021, 74: 102233.
11. Gamgam G, Kabakcioglu A, Yüksel Dal D, et al. Disentangled attention graph neural network for Alzheimer’s disease diagnosis//International Conference on Medical Image Computing and Computer-Assisted Intervention. Marrakesh: MICCAI, 2024: 219-228.
12. Wang C, Xiao Z, Xu Y, et al. A novel approach for ASD recognition based on graph attention networks. Front Comput Neurosci, 2024, 18: 1388083.
13. Li Y, Liu J, Gao X, et al. Multimodal hyper-connectivity of functional networks using functionally-weighted LASSO for MCI classification. Med Image Anal, 2019, 52: 80-96.
14. Eavani H, Satterthwaite T D, Filipovych R, et al. Identifying sparse connectivity patterns in the brain using resting-state fMRI. NeuroImage, 2015, 105: 286-299.
15. Bai S, Zhang F, Torr P H S. Hypergraph convolution and hypergraph attention. Pattern Recognit, 2021, 110: 107637.
16. Gao Y, Zhang Z, Lin H, et al. Hypergraph learning: methods and practices. IEEE Trans Pattern Anal Mach Intell, 2022, 44(5): 2548-2566.
17. 孔文文. 基于靜息態腦電數據的高階腦網絡分析和抑郁識別研究. 蘭州: 蘭州大學, 2023.
18. Liu J, Cui W, Chen Y, et al. Deep fusion of multi-template using spatio-temporal weighted multi-hypergraph convolutional networks for brain disease analysis. IEEE Trans Med Imaging, 2024, 43(2): 860-873.
19. Jie B, Wee C Y, Shen D, et al. Hyper-connectivity of functional networks for brain disease diagnosis. Med Image Anal, 2016, 32: 84-100.
20. Guo H, Li Y, Xu Y, et al. Resting-state brain functional hyper-network construction based on elastic net and group lasso methods. Front Neuroinform, 2018, 12: 25.
21. Lostar M, Rekik I. Deep hypergraph u-net for brain graph embedding and classification. arXiv, 2020: 2008.13118.
22. Wang J, Li H, Qu G, et al. Dynamic weighted hypergraph convolutional network for brain functional connectome analysis. Med Image Anal, 2023, 87: 102828.
23. Xiao L, Wang J, Kassani P H, et al. Multi-hypergraph learning-based brain functional connectivity analysis in fMRI data. IEEE Trans Med Imaging, 2020, 39(5): 1746-1758.
24. Xu B, Dall’Aglio L, Flournoy J, et al. Limited generalizability of multivariate brain-based dimensions of child psychiatric symptoms. Commun Psychol, 2024, 2(1): 16.
25. Lu J, Li A, Li K, et al. An EEG study on β?γ phase-amplitude coupling-based functional brain network in epilepsy patients. IEEE J Biomed Health Inform, 2024, 28(6): 3446-3456.
26. Alim A, Imtiaz M H. Automatic identification of children with ADHD from EEG brain waves. Signals, 2023, 4(1): 193-205.
27. Cai H, Gao Y, Sun S, et al. Modma dataset: a multi-modal open dataset for mental-disorder analysis. arXiv, 2020: 2002.09283.
28. Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis. J Neurosci Methods, 2004, 134(1): 9-21.
29. Tadel F, Baillet S, Mosher J C, et al. Brainstorm: a user-friendly application for MEG/EEG analysis. Comput Intell Neurosci, 2011, 2011: 879716.
30. Desikan R S, Ségonne F, Fischl B, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. NeuroImage, 2006, 31(3): 968-980.
31. Brookshire G, Kasper J, Blauch N M, et al. Data leakage in deep learning studies of translational EEG. Front Neurosci, 2024, 18: 1373515.
32. Jakkula V. Tutorial on support vector machine (SVM). Pullman: Washington State University, 2006.
33. Moon S E, Jang S, Lee J S. Convolutional neural network approach for EEG-based emotion recognition using brain connectivity and its spatial information//2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Calgary: IEEE, 2018: 2556-2560.
34. Veli?kovi? P, Cucurull G, Casanova A, et al. Graph attention networks. arXiv, 2017: 1710.10903.
35. Selvaraju R R, Cogswell M, Das A, et al. Grad-cam: visual explanations from deep networks via gradient-based localization//Proceedings of the IEEE International Conference on Computer Vision. Venice: IEEE, 2017: 618-626.
36. Latifi B, Amini A, Nasrabadi A M. Siamese based deep neural network for ADHD detection using EEG signal. Comput Biol Med, 2024, 182: 109092.
37. Hu Y, Ran J, Qiao R, et al. Identifying ADHD-related abnormal functional connectivity with a graph convolutional neural network. Neural Plast, 2024, 2024: 8862647.
38. Scalabrini A, Poletti S, Vai B, et al. Abnormally slow dynamics in occipital cortex of depression. J Affect Disord, 2025, 374: 523-530.
39. Chao Z C, Dillon D G, Liu Y H, et al. Altered coordination between frontal delta and parietal alpha networks underlies anhedonia and depressive rumination in major depressive disorder. J Psychiatry Neurosci, 2022, 47(6): E367-E378.
40. Zhao Y, Chen L, Zhang W, et al. Gray matter abnormalities in non-comorbid medication-naive patients with major depressive disorder or social anxiety disorder. EBioMedicine, 2017, 21: 228-235.

Journal of Biomedical Engineering

Research on weighted hypergraph attention neural network for the diagnosis of psychiatric disorders using brain functional connectivity networks

Abstract Full text Figures/Tables Video References Cited by

Previous Article

Next Article

Format

Content