Research on Epilepsy Detection Based on Pyramid Graph Convolution Network for Brain Electrical Activity Spatial Topological Features
DOI:
https://doi.org/10.54097/hns90p36Keywords:
EEG, Class Imbalance, Pyramid Graph Convolutional Network, Feature ExtractionAbstract
To address the high cost of spatio-temporal modeling for epilepsy detection, this paper proposes a lightweight pyramid graph convolutional network model. This model abandons complex temporal mechanisms and focuses on mining the spatial topology of the EEG frequency domain: extracting multi-band power spectra and statistical moments as node features, and constructing a Gaussian adjacency matrix based on the 10-20 system to simulate functional connections; the core adopts multi-scale pyramid graph convolutional blocks to capture multi-receptive field spatial dependencies and fuse features. Additionally, EEG-specific augmentation, Borderline-SMOTE, and Focal Loss are combined to address class imbalance. On the CHB-MIT dataset, the model achieves a sensitivity of 97.53%, specificity of 96.50%, accuracy of 97.02%, F1 score of 97.04%, and AUC of 99.13%. Experiments confirm the sufficiency of spatial topological features in epilepsy discrimination, with performance superior to or comparable to mainstream methods, providing an efficient solution for portable clinical monitoring.
Downloads
References
[1] Elger Christian E, Hoppe Christian. Diagnostic challenges in epilepsy: seizure under-reporting and seizure detection [J]. The Lancet Neurology, 2018, 17(3): 204-205.
[2] Tzallas Alexandros T, Tsipouras Markos G, Tsalikakis Dimitrios G, et al.Automated Epileptic Seizure Detection Methods: A Review Study[J].Epilepsy - Histological, Electroencephalographic and Psychological Aspects,2012,9:2027-2036.
[3] Yuan Qi, Zhou Weidong, Liu Yinxia, et al.Epileptic seizure detection with linear and nonlinear features[J].Epilepsy & Behavior,2012,24(4):415-421.
[4] Kemal Akyol.Stacking ensemble based deep neural networks modeling for effective epileptic seizure detection[J]. Expert Systems with Applications,2020,148:113239.
[5] Dash Deba Prasad, Kolekar Maheshkumar H, Chakraborty Chinmay, et al.Review of Machine and Deep Learning Techniques in Epileptic Seizure Detection using Physiological Signals and Sentiment Analysis[J].ACM Transactions on Asian and Low-Resource Language Information Processing,2024,23(1):1-29.
[6] Liu Mingze, Liu Jie, Xu Mengna, et al.Combining meta and ensemble learning to classify EEG for seizure detection[J]. Scientific Reports,2025,15(1):10755.
[7] Feng Hailing, Wang Shuai, Lv Hongbing, et al.Cross-subject seizure detection with vision transformer and unsupervised domain adaptation[J].Biomedical Signal Processing and Control,2026,111:108341.
[8] Zhou Mengni, Tian Cheng, Cao Rui, et al.Epileptic seizure detection based on EEG signals and CNN[J]. Frontiers in Neuroinformatics,2018,12:425101.
[9] Hossain M S, Amin S U, Alsulaiman M, et al.Applying deep learning for epilepsy seizure detection and brain mapping visualization[J].ACM Transactions on Multimedia Computing, Communications and Applications, 2019, 15(1s):1-17.
[10] Khan Gul Hameed, Khan Nadeem Ahmad, Altaf Muhammad Awais Bin, et al.A Shallow Autoencoder Framework for Epileptic Seizure Detection in EEG Signals[J].Sensors,2023,2023,23(8):4112.
[11] Huang Chengbin, Chen Weiting, Cao Guitao.Automatic Epileptic Seizure Detection via Attention-Based CNN-BiRNN[J]. Proceedings-2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019,2019:660-663.
[12] Zabihi Morteza, Kiranyaz Serkan, Jäntti Ville, et al. Patient-specific seizure detection using nonlinear dynamics and nullclines[J]. IEEE Journal of Biomedical and Health Informatics, 2020, 24(2): 543-555..
[13] Guo Yao, Jiang Xinyu, Tao Linkai, et al. Epileptic Seizure Detection by Cascading Isolation Forest-Based Anomaly Screening and EasyEnsemble[J].IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022, 30: 915-924.
[14] Yang Yong, Qin Xiaolin, Lin Xiaoguang, et al. Epilepsy detection and analysis method for specific patient based on data augmentation and deep learning[J].Journal of Biomedical Engineering, 2022, 39(2): 293-300.
[15] Tao Tianli, Guo Lianghu, He Qiang, Zhang Han, et al. Seizure detection by brain connectivity analysis using dynamic graph isomorphism network[C]//Proceedings of the 44th Annual International Conference of the IEEE Engineering in Medicine and Biology Society. 2022: 2302-2305.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Advanced Engineering and Technology Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
