一种融合连续小波卷积与图嵌入的注意力网络

张璐; 李明爱

doi:10.12263/DZXB.20250435

您当前的位置：

首页 >

文章列表页 >

一种融合连续小波卷积与图嵌入的注意力网络

学术论文 | 更新时间：2026-04-24

- 一种融合连续小波卷积与图嵌入的注意力网络
- An Attention Network with Continuous Wavelet Convolution and Graph Embedding
- 电子学报 2025年53卷第12期页码：4337-4348
- 作者机构：
  
  北京工业大学信息科学技术学院，北京 100124
- 作者简介：
  
  [ "张璐男，2001年10月生，重庆开州人.现为北京工业大学信息科学技术学院硕士研究生.主要研究方向为脑机接口技术、信号处理与模式识别.E-mail: zl054490@163.com" ]
  [ "李明爱女，1966年7月生，河南鹤壁人.2006年于北京工业大学获得博士学位，现为北京工业大学教授.主要研究方向为脑机接口技术、人工智能与智能康复.E-mail: limingai@bjut.edu.cn" ]
- 基金信息：
  
  国家自然科学基金(62173010)
- DOI：10.12263/DZXB.20250435
  中图分类号： TP391;
- 收稿：2025-05-28，
  
  录用：2025-12-10，
  
  纸质出版：2025-12-25
- 稿件说明：
移动端阅览
张璐, 李明爱. 一种融合连续小波卷积与图嵌入的注意力网络[J]. 电子学报, 2025, 53(12): 4337-4348.

ZHANG Lu, LI Ming-ai. An Attention Network with Continuous Wavelet Convolution and Graph Embedding[J]. Acta Electronica Sinica, 2025, 53(12): 4337-4348.
张璐, 李明爱. 一种融合连续小波卷积与图嵌入的注意力网络[J]. 电子学报, 2025, 53(12): 4337-4348. DOI：10.12263/DZXB.20250435

ZHANG Lu, LI Ming-ai. An Attention Network with Continuous Wavelet Convolution and Graph Embedding[J]. Acta Electronica Sinica, 2025, 53(12): 4337-4348. DOI：10.12263/DZXB.20250435

摘要

基于深度学习模型解码运动想象脑电信号（Motor Imagery ElectroEncephaloGram，MI-EEG）是脑机接口（Brain-Computer Interface，BCI）技术领域的热点研究问题之一.针对MI-EEG的时频特点和个体差异性，众多研究对其进行时频分析并将其时频表示广泛用于MI-EEG解码.然而，现有方法大多忽略了MI-EEG的多电极空间分布特性，未能很好地挖掘和利用不同电极间的拓扑关系，从而影响了特征信息的完整性，限制了解码性能的改善.为自适应学习多电极MI-EEG间的拓扑信息，并有效增强其时频空特征信息，本文提出一种融合连续小波卷积与图嵌入的注意力网络（Attention Network with Continuous Wavelet Convolution and Graph Embedding，CWC-GEAN）.该网络包含5个模块：多分支连续小波卷积模块（Multi-branch Continuous Wavelet Convolution Module，MCWCM）、多分支动态图嵌入模块（Multi-branch dynamic Graph Embedding Module，MGEM）、多分支特征通道注意力模块（Multi-branch Feature Channel Attention Module，MFCAM）、多分支特征通道-时间注意力模块（Multi-branch Feature Channel-Time Attention Module，MFCTAM）及特征融合与分类模块（Feature Fusion and Classification Block，FFCB）.首先，将原始多电极MI-EEG信号输入多分支连续小波卷积模块，在4个分支中分别基于4个子频带（1~8 Hz、9~16 Hz、17~24 Hz、25~32 Hz）进行连续小波卷积，通过对尺度因子的动态学习获得最优多尺度频-空-时特征表示；其次，基于互信息构建包含电极间拓扑信息的先验邻接矩阵，并经由MGEM，从不同子频带对先验邻接矩阵进行自适应学习调整，并嵌入对应分支频-空-时特征表示中，得到蕴含电极间拓扑信息的图结构特征；再次，由MFCAM和MFCTAM针对各分支图结构特征进一步提取深层特征，并相继完成特征通道注意力向量和特征通道-时间注意力矩阵的自动获取与特征加权，得到多分支判别性特征；最后，由FFCB对多分支判别性特征进行融合，得到最终的分类结果.本文基于公开的BCI Competition IV 2a数据集和High-Gamma Dataset数据集对CWC-GEAN进行性能评估，平均分类准确率分别为85.45%和95.09%，平均Kappa值分别为0.806和0.934.结果表明，CWC-GEAN具有自适应学习、捕捉MI-EEG时频信息和电极拓扑信息，以及增强时-频-空特征的能力，且展现出较好的模型鲁棒性和分类结果的一致性，相对流行方法具有一定性能优势.

Abstract

Decoding motor imagery electroencephalogram (MI-EEG) signals based on deep learning models is one of the hot research topics in the field of brain-computer interface (BCI) technology. Aiming at the time-frequency characteristics and individual differences of MI-EEG

numerous studies have conducted time-frequency analysis on MI-EEG and widely applied its time-frequency representations to MI-EEG decoding. However

most existing methods ignore the spatial distribution characteristics of multi-electrode MI-EEG and fail to fully explore and utilize the topological relationships between different electrodes

thereby affecting the integrity of feature information and limiting the further improvement of decoding performance. To adaptively learn the topological information between multi-electrode MI-EEG and effectively enhance its time-frequency-spatial feature information

this paper proposes an attention network with continuous wavelet convolution and graph embedding (CWC-GEAN). The network consists of five modules: a multi-branch continuous wavelet convolution module (MCWCM)

a multi-branch dynamic graph embedding module (MGEM)

a multi-branch feature channel attention module (MFCAM)

a multi-branch feature channel-time attention module (MFCTAM)

and a feature fusion and classification block (FFCB). First

the original multi-electrode MI-EEG signals are input into the MCWCM

where continuous wavelet convolution is performed based on four sub-bands(1 Hz to 8 Hz

9 Hz to 16 Hz

17 Hz to 24 Hz

25 Hz to 32 Hz) in four branches respectively

and the optimal multi-scale frequency-spatial-temporal feature representations are obtained through dynamic learning of scale factors. Then

a prior adjacency matrix containing topological information between electrodes is constructed based on mutual information

and the prior adjacency matrix is adaptively learned and adjusted from different sub-bands via the MGEM

which is embedded into the frequency-spatial-temporal feature representations of corresponding branches to obtain graph structure features containing topological information between electrodes. Furthermore

the MFCAM and the MFCTAM further extract deep features from the graph structure features of each branch

and successively complete the automatic acquisition of feature channel attention vectors and feature channel-time attention matrices as well as feature weighting to obtain multi-branch discriminative features. Finally

the FFCB fuses the multi-branch discriminative features to obtain the final classification results. In this paper

the performance of CWC-GEAN is evaluated based on the public BCI Competition IV 2a dataset and High-Gamma Dataset

with average classification accuracies of 85.45% and 95.09%

and average Kappa values of 0.806 and 0.934

respectively. The results show that CWC-GEAN has the ability to adaptively learn and capture the time-frequency information and electrode topological information of MI-EEG

as well as enhance time-frequency-spatial features

and exhibits good model robustness and consistency of classification results

with certain performance advantages over popular methods.

关键词

Keywords

references

李明爱 , 张圆圆 . 基于连续小波变换和符号传递熵的脑功能网络构建方法 [J ] . 电子学报 , 2022 , 50 ( 7 ): 1600 - 1608 .

LI M A , ZHANG Y Y . A brain functional network based on continuous wavelet transform and symbolic transfer entropy [J ] . Acta Electronica Sinica , 2022 , 50 ( 7 ): 1600 - 1608 . (in Chinese)

于城 , 戴亚康 , 刘燕 , 等 . 上肢康复运动想象脑机接口的研究进展 [J ] . 北京生物医学工程 , 2025 , 44 ( 2 ): 198 - 204 .

YU C , DAI Y K , LIU Y , et al . Research progress of motor imagery-based brain-computer interface for upper limb rehabilitation [J ] . Beijing Biomedical Engineering , 2025 , 44 ( 2 ): 198 - 204 . (in Chinese)

张瑞 , 李远清 . 多模态脑机接口游戏系统的设计与应用 [J ] . 计算机工程与应用 , 2012 , 48 ( 22 ): 65 - 69 .

ZHANG R , LI Y Q . Design and application of multi-model BCI game system [J ] . Computer Engineering and Applications , 2012 , 48 ( 22 ): 65 - 69 . (in Chinese)

LV D M , DANG W D , FENG J H , et al . Attention-based CNN model for motor imagery classification from nonlinear EEG signals [J ] . Physica A: Statistical Mechanics and its Applications , 2024 , 655 : 130191 .

WANG X F , WANG Y , QI W X , et al . BrainGridNet: A two-branch depthwise CNN for decoding EEG-based multi-class motor imagery [J ] . Neural Networks , 2024 , 170 : 312 - 324 .

LI F , HE F , WANG F , et al . A novel simplified convolutional neural network classification algorithm of motor imagery EEG signals based on deep learning [J ] . Applied Sciences , 2020 , 10 ( 5 ): 1605 .

WANG C , WU Y , WANG C , et al . MI-EEG classification using Shannon complex wavelet and convolutional neural networks [J ] . Applied Soft Computing , 2022 , 130 : 109685 .

KIM J M , HEO K S , SHIN D H , et al . A learnable continuous wavelet-based multi-branch attentive convolutional neural network for spatio-spectral-temporal EEG signal decoding [J ] . Expert Systems with Applications , 2024 , 251 : 123975 .

ZHANG D L , CHEN K X , JIAN D B , et al . Motor imagery classification via temporal attention cues of graph embedded EEG signals [J ] . IEEE Journal of Biomedical and Health Informatics , 2020 , 24 ( 9 ): 2570 - 2579 .

SHI J , TANG J M , LU Z H , et al . A brain topography graph embedded convolutional neural network for EEG-based motor imagery classification [J ] . Biomedical Signal Processing and Control , 2024 , 95 : 106401 .

TANG X L , ZHANG J , QI Y D , et al . A spatial filter temporal graph convolutional network for decoding motor imagery EEG signals [J ] . Expert Systems with Applications , 2024 , 238 : 121915 .

DEMIR A , KOIKE-AKINO T , WANG Y , et al . EEG-GAT: Graph attention networks for classification of electroencephalogram (EEG) signals [C ] // 2022 44th Annual International Conference of the IEEE Engineering in Medicine & Biology Society . Piscataway : IEEE , 2022 : 30 - 35 .

唐向宏 , 李齐良 . 时频分析与小波变换 [M ] . 北京 : 科学出版社 , 2008 : 118 - 126 .

TANG X H , LI Q L . Time-Frequency Analysis and Wavelet Transform [M ] . Beijing : Science Press , 2008 : 118 - 126 . (in Chinese)

LI S Q , MA S Y , WANG S Q . Optimal complex morlet wavelet parameters for quantitative time-frequency analysis of molecular vibration [J ] . Applied Sciences , 2023 , 13 ( 4 ): 2734 .

SCHREIBER T . Measuring information transfer [J ] . Physical Review Letters , 2000 , 85 ( 2 ): 461 - 464 .

HU J , SHEN L , SUN G . Squeeze-and-excitation networks [C ] // 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2018 : 7132 - 7141 .

PARK J , WOO S , LEE J Y , et al . BAM: Bottleneck attention module [EB/OL ] . ( 2018-07-18 )[ 2025-10-10 ] . https://arXiv.org/abs/1807.06514 https://arXiv.org/abs/1807.06514 .

YU G L , LIN Y H . Spatio-temporal adaptive fusion network based on spatial-channel-axis attention mechanism [C ] // 2025 5th International Conference on Artificial Intelligence, Big Data and Algorithms . Piscataway : IEEE , 2025 : 225 - 229 .

LAWHERN V J , SOLON A J , WAYTOWICH N R , et al . EEGNet: A compact convolutional network for eeg-based brain-computer interfaces [EB/OL ] . ( 2018-05-16 )[ 2025-10-11 ] . https://arXiv.org/abs/1611.08024 https://arXiv.org/abs/1611.08024 .

INGOLFSSON T M , HERSCHE M , WANG X Y , et al . EEG-TCNet: An accurate temporal convolutional network for embedded motor-imagery brain-machine interfaces [C ] // 2020 IEEE International Conference on Systems, Man, and Cybernetics . Piscataway : IEEE , 2020 : 2958 - 2965 .

WANG H Y , YU H , WANG H X . EEG_GENet: A feature-level graph embedding method for motor imagery classification based on EEG signals [J ] . Biocybernetics and Biomedical Engineering , 2022 , 42 ( 3 ): 1023 - 1040 .

SCHIRRMEISTER R T , SPRINGENBERG J T , FIEDERER L D J , et al . Deep learning with convolutional neural networks for EEG decoding and visualization [J ] . Human Brain Mapping , 2017 , 38 ( 11 ): 5391 - 5420 .

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

一种基于对比学习大模型的视觉定位方法

统一全局空间表达的脑电信号跨被试情感识别

双注意力引导的细节和结构信息融合图像去雾网络

基于方面和胶囊网络的跨域评分预测模型