基于多头交叉注意力机制融合多域特征的癫痫分类识别研究

常文文; 王亚俊; 郭晋成; 舒康; 马煜

doi:10.12263/DZXB.20250911

您当前的位置：

首页 >

文章列表页 >

基于多头交叉注意力机制融合多域特征的癫痫分类识别研究

学术论文 | 更新时间：2026-04-24

- 基于多头交叉注意力机制融合多域特征的癫痫分类识别研究
- Research on Epilepsy Classification and Recognition Based on Multi-Domain Feature Fusion via Multi-Head Cross-Attention Mechanism
- 电子学报 2025年53卷第12期页码：4460-4473
- 作者机构：
  
  兰州交通大学电子与信息工程学院，甘肃兰州 730070
- 作者简介：
  
  [ "常文文男，1987年3月生，甘肃通渭人．2019年于东北大学获得工学博士学位.瑞士洛桑联邦理工学院神经义肢技术中心访问学者.现为兰州交通大学副教授、硕士生导师.主要研究方向为脑-机交互、脑电信号处理和模式识别.中国电子学会会员编号：E190085440M.E-mail: changww2013@126.com" ]
  [ "王亚俊女，2002年10月生，河南南阳人.现为兰州交通大学计算机技术专业硕士研究生.主要研究方向为癫痫脑电检测与分类识别.E-mail: 155837736799@163.com" ]
  [ "郭晋成男，2002年4月生，甘肃兰州人.现为兰州交通大学计算机技术专业硕士研究生.主要研究方向为癫痫脑电研究.E-mail: 193942768@126.com" ]
- 基金信息：
  
  国家自然科学基金(62366028;W2421090);甘肃省科技重大专项(23ZDFA012);甘肃省自然科学基金(24JRRA256)
- DOI：10.12263/DZXB.20250911
  中图分类号： TP391;
- 收稿：2025-11-27，
  
  录用：2025-12-12，
  
  纸质出版：2025-12-25
- 稿件说明：
移动端阅览
常文文, 王亚俊, 郭晋成, 等. 基于多头交叉注意力机制融合多域特征的癫痫分类识别研究[J]. 电子学报, 2025, 53(12): 4460-4473.

CHANG Wen-wen, WANG Ya-jun, GUO Jin-cheng, et al. Research on Epilepsy Classification and Recognition Based on Multi-Domain Feature Fusion via Multi-Head Cross-Attention Mechanism[J]. Acta Electronica Sinica, 2025, 53(12): 4460-4473.
常文文, 王亚俊, 郭晋成, 等. 基于多头交叉注意力机制融合多域特征的癫痫分类识别研究[J]. 电子学报, 2025, 53(12): 4460-4473. DOI：10.12263/DZXB.20250911

CHANG Wen-wen, WANG Ya-jun, GUO Jin-cheng, et al. Research on Epilepsy Classification and Recognition Based on Multi-Domain Feature Fusion via Multi-Head Cross-Attention Mechanism[J]. Acta Electronica Sinica, 2025, 53(12): 4460-4473. DOI：10.12263/DZXB.20250911

摘要

针对癫痫脑电信号（ElectroEncephaloGram，EEG）发作前期（preictal）、发作期（ictal）和发作间期（interictal）三种状态下的检测和识别问题，提出了一种基于多头交叉注意力机制（Multi-Head Cross-Attention Mechanism，MHCA）融合脑电多域特征的分类识别模型.该模型通过连续小波变换（Continuous Wavelet Transform，CWT）对癫痫脑电信号生成的二维图像按照通道进行有序拼接，并利用浅层卷积神经网络（Convolutional Neural Network，CNN）对拼接后的时频图像进行特征提取以有效挖掘癫痫脑电信号的时频域特征.同时通过构建脑功能连接矩阵，刻画不同脑区之间的功能连接关系以捕捉癫痫发作过程中潜在的空域特征，最终采用MHCA实现时频特征与空域特征之间的全局交互与自适应融合，充分建模多维特征间的关联性与互补性，进而构建癫痫脑电信号在时域、频域与空域三个维度上的完整且统一的特征表征.实验结果表明，该模型在癫痫发作前期、发作期和发作间期的多被试分类中最高分类准确率可达92.49%，灵敏度可达92.48%，体现了其在跨被试场景下良好的泛化能力与稳定性；单被试分类中模型最高准确率为98.39%，灵敏度为98.04%，充分验证了该方法在个体化癫痫脑电识别任务中的有效性.消融实验最终也进一步证实了脑功能连接矩阵所表征的空域信息和多头交叉注意力机制在多域特征融合与判别特征增强中的关键作用，对模型的性能提升均具有正向贡献.本文针对癫痫脑电分类识别的有效性验证，不仅为临床脑电检测和识别提供了一种可靠且可行的技术手段，也为癫痫脑电信号中关键特征的提取、表征与建模提供了新的研究思路和方法参考.

Abstract

In this paper

we propose a classification and recognition model based on the multi-head cross-attention mechanism (MHCA) fusing EEG multi-domain features for the detection and recognition of epileptic electroencephalographic (EEG) signals in preictal

ictal and interictal states. The model is developed by sequentially splicing the two-dimensional images generated by continuous wavelet transform (CWT) of epileptic EEG signals according to the channels

and utilizing shallow convolutional neural network (CNN) to extract features from the spliced time-frequency images in order to effectively extract the time-frequency domain features of epileptic EEG signals. At the same time

the brain functional connectivity matrix is constructed to depict the functional connectivity between different brain regions to capture the potential spatial features during epileptic seizures

and finally

MHCA is used to realize the global interaction and adaptive fusion between time-frequency and spatial features to fully model the correlation and complementarity between multidimensional features

so as to construct a complete and unified feature characterization of epileptic EEG signals in the three dimensions of the time domain

the frequency domain

and the spatial domain. The experimental results show that the model can reach a maximum classification accuracy of 92.49% and a sensitivity of 92.48% in multi-subject classification of preictal

ictal and interictal phases

which reflects its good generalization ability and stability in cross-subject scenarios; and the model can reach a maximum accuracy of 98.39% and a sensitivity of 98.04% in single-subject classification

which fully verifies the efficacy of the method in the task of individualized epilepsy EEG recognition. The ablation experiments ultimately further confirmed the critical role of the spatial information represented by the brain functional connectivity matrix and the multi-head cross-attention mechanism in multi-domain feature fusion and discriminative feature enhancement

both of which positively contributed to the model's performance improvement. This paper validates the efficacy of epileptic EEG classification and recognition

which not only provides a reliable and feasible technical means for clinical EEG detection and recognition

but also provides new research ideas and methodological references for the extraction

characterization and modeling of key features in epileptic EEG signals.

关键词

Keywords

references

BANDOPADHYAY R , SINGH T , GHONEIM M M , et al . Recent developments in diagnosis of epilepsy: Scope of microRNA and technological advancements [J ] . Biology , 2021 , 10 ( 11 ): 1097 .

RIDSDALE L , CHARLTON J , ASHWORTH M , et al . Epilepsy mortality and risk factors for death in epilepsy: A population-based study [J ] . The British Journal of General Practice , 2011 , 61 ( 586 ): e271 - e278 .

FISHER R S , VICKREY B G , GIBSON P , et al . The impact of epilepsy from the patient’s perspective I. Descriptions and subjective perceptions [J ] . Epilepsy Research , 2000 , 41 ( 1 ): 39 - 51 .

WHO . Epilepsy [EB/OL ] . ( 2024-02-07 )[ 2025-11-20 ] . https://www.who.int/news-room/fact-sheets/detail/epilepsy https://www.who.int/news-room/fact-sheets/detail/epilepsy .

韩长明 , 彭福来 , 陈财 , 等 . 基于脑电信号的癫痫发作预测研究进展 [J ] . 生物医学工程学杂志 , 2021 , 38 ( 6 ): 1193 - 1202 .

HAN C M , PENG F L , CHEN C , et al . Research progress of epileptic seizure predictions based on electroencephalogram signals [J ] . Journal of Biomedical Engineering , 2021 , 38 ( 6 ): 1193 - 1202 . (in Chinese)

ZHANG T , CHEN W Z . LMD based features for the automatic seizure detection of EEG signals using SVM [J ] . IEEE Transactions on Neural Systems and Rehabilitation Engineering , 2017 , 25 ( 8 ): 1100 - 1108 .

JIANG L R , HE J W , PAN H Y , et al . Seizure detection algorithm based on improved functional brain network structure feature extraction [J ] . Biomedical Signal Processing and Control , 2023 , 79 : 104053 .

SI Y . Machine learning applications for electroencephalograph signals in epilepsy: A quick review [J ] . Acta Epileptologica , 2020 , 2 ( 1 ): 5 .

SIDDIQUI M K , MORALES-MENENDEZ R , HUANG X D , et al . A review of epileptic seizure detection using machine learning classifiers [J ] . Brain Informatics , 2020 , 7 ( 1 ): 5 .

张学军 , 景鹏 , 何涛 , 等 . 基于变分模态分解的癫痫脑电信号分类方法 [J ] . 电子学报 , 2020 , 48 ( 12 ): 2469 - 2475 .

ZHANG X J , JING P , HE T , et al . An epileptic electroencephalogram signal classification method based on variational mode decomposition [J ] . Acta Electronica Sinica , 2020 , 48 ( 12 ): 2469 - 2475 . (in Chinese)

王兴起 , 李明爱 . 应用基于注意力的多尺度残差网络实现癫痫脑电自动检测 [J ] . 生物医学工程学杂志 , 2024 , 41 ( 2 ): 253 - 261 .

WANG X Q , LI M A . Automatic epilepsy detection with an attention-based multiscale residual network [J ] . Journal of Biomedical Engineering , 2024 , 41 ( 2 ): 253 - 261 . (in Chinese)

AKOR P , ENEMALI G , MUHAMMAD U , et al . Hierarchical deep learning for comprehensive epileptic seizure analysis: From detection to fine-grained classification [J ] . Information , 2025 , 16 ( 7 ): 532 .

SHUKLA R , KUMAR B , GAURAV G , et al . Epileptic seizure detection using continuous wavelet transform and deep neural networks [M ] // Sensing Technology . Cham : Springer International Publishing , 2022 : 291 - 300 .

LI C Y , LI H T , DONG X C , et al . CNN-Informer: A hybrid deep learning model for seizure detection on long-term EEG [J ] . Neural Networks , 2025 , 181 : 106855 .

QIU X J , YAN F , LIU H H . A difference attention ResNet-LSTM network for epileptic seizure detection using EEG signal [J ] . Biomedical Signal Processing and Control , 2023 , 83 : 104652 .

CAO X S , ZHENG S J , ZHANG J C , et al . A hybrid CNN-Bi-LSTM model with feature fusion for accurate epilepsy seizure detection [J ] . BMC Medical Informatics and Decision Making , 2025 , 25 ( 1 ): 6 .

LIU G Y , TIAN L , WEN Y M , et al . Cosine convolutional neural network and its application for seizure detection [J ] . Neural Networks , 2024 , 174 : 106267 .

LI Q , CAO W , ZHANG A Y . Multi-stream feature fusion of vision transformer and CNN for precise epileptic seizure detection from EEG signals [J ] . Journal of Translational Medicine , 2025 , 23 ( 1 ): 871 .

WU X , YANG Z H , ZHANG T L , et al . An end-to-end seizure prediction approach using long short-term memory network [J ] . Frontiers in Human Neuroscience , 2023 , 17 : 1187794 .

LI H T , DONG X C , ZHONG X W , et al . End-to-end model for automatic seizure detection using supervised contrastive learning [J ] . Engineering Applications of Artificial Intelligence , 2024 , 133 : 108665 .

HUANG W Q , YAN G H , CHANG W W , et al . EEG-based classification combining Bayesian convolutional neural networks with recurrence plot for motor movement/imagery [J ] . Pattern Recognition , 2023 , 144 : 109838 .

SHANKAR A , DANDAPAT S , BARMA S . Seizure type classification using EEG based on gramian angular field transformation and deep learning [C ] // 2021 43rd Annual International Conference of the IEEE Engineering in Medicine & Biology Society . Piscataway : IEEE , 2021 : 3340 - 3343 .

OZCAN A R , ERTURK S . Seizure prediction in scalp EEG using 3D convolutional neural networks with an image-based approach [J ] . IEEE Transactions on Neural Systems and Rehabilitation Engineering , 2019 , 27 ( 11 ): 2284 - 2293 .

CHO K O , JANG H J . Comparison of different input modalities and network structures for deep learning-based seizure detection [J ] . Scientific Reports , 2020 , 10 : 122 .

BHANDAGE V , POKURI T , DESAI D , et al . Detection of epilepsy disorder using spectrogram images generated from brain EEG signals [J ] . IEEE Access , 2024 , 12 : 195054 - 195064 .

SHEN M K , YANG F W , WEN P , et al . A real-time epilepsy seizure detection approach based on EEG using short-time Fourier transform and Google-Net convolutional neural network [J ] . Heliyon , 2024 , 10 ( 11 ): e31827 .

DIŞLI F , GEDIKPıNAR M , FıRAT H , et al . Epilepsy diagnosis from EEG signals using continuous wavelet transform-based depthwise convolutional neural network model [J ] . Diagnostics , 2025 , 15 ( 1 ): 84 .

LIU G Y , ZHOU W D , GENG M X . Automatic seizure detection based on S-transform and deep convolutional neural network [J ] . International Journal of Neural Systems , 2020 , 30 ( 4 ): 1950024 .

吕仁杰 , 常文文 , 闫光辉 , 等 . 基于GAF和混合模型的运动想象分类研究 [J ] . 电子科技大学学报 , 2024 , 53 ( 6 ): 952 - 960 .

LYU R J , CHANG W W , YAN G H , et al . Research on motor imagery classification based on GAF and hybrid model [J ] . Journal of University of Electronic Science and Technology of China , 2024 , 53 ( 6 ): 952 - 960 . (in Chinese)

VASWANI A , SHAZEER N , PARMAR N , et al . Attention is all you need [EB/OL ] . ( 2023-08-02 )[ 2025-11-20 ] . https://arxiv.org/abs/1706.03762 https://arxiv.org/abs/1706.03762 .

SHARMA R , PACHORI R B . Classification of epileptic seizures in EEG signals based on phase space representation of intrinsic mode functions [J ] . Expert Systems with Applications , 2015 , 42 ( 3 ): 1106 - 1117 .

LV R J , CHANG W W , YAN G H , et al . Enhanced classification of motor imagery EEG signals using spatio-temporal representations [J ] . Information Sciences , 2025 , 714 : 122221 .

SUN C S , XU C , LI H W , et al . A novel multi-feature fusion attention neural network for the recognition of epileptic EEG signals [J ] . Frontiers in Computational Neuroscience , 2024 , 18 : 1393122 .

ZHAO J Q , ZHANG M , ZHOU Z H , et al . CFFormer: A cross-fusion transformer framework for the semantic segmentation of multisource remote sensing images [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2025 , 63 : 4401117 .

LI Y , LIU Y , CUI W G , et al . Epileptic seizure detection in EEG signals using a unified temporal-spectral squeeze-and-excitation network [J ] . IEEE Transactions on Neural Systems and Rehabilitation Engineering , 2020 , 28 ( 4 ): 782 - 794 .

IBRAHIM A K , ZHUANG H Q , TOGNOLI E , et al . Epileptic seizure prediction based on multiresolution convolutional neural networks [J ] . Frontiers in Signal Processing , 2023 , 3 : 1175305 .

SU J Y , HUANG Z T , MA Y H , et al . A model for epileptic EEG detection and recognition based on multi-attention mechanism and spatiotemporal [J ] . Scientific Reports , 2025 , 15 : 31993 .

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于混合时频域特征的卷积神经网络心律失常分类方法的研究

基于连续小波变换和符号传递熵的脑功能网络构建方法

利用脊的特征进行信号盲分离