A composite improved attention convolutional network for motor imagery EEG classification
IntroductionA brain-computer interface (BCI) is an emerging technology that aims to establish a direct communication pathway between the human brain and external devices. Motor imagery electroencephalography (MI-EEG) signals are analyzed to infer users’ intentions during motor imagery. These signals...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-02-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2025.1543508/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832096540199485440 |
|---|---|
| author | Wenzhe Liao Zipeng Miao Shuaibo Liang Linyan Zhang Chen Li |
| author_facet | Wenzhe Liao Zipeng Miao Shuaibo Liang Linyan Zhang Chen Li |
| author_sort | Wenzhe Liao |
| collection | DOAJ |
| description | IntroductionA brain-computer interface (BCI) is an emerging technology that aims to establish a direct communication pathway between the human brain and external devices. Motor imagery electroencephalography (MI-EEG) signals are analyzed to infer users’ intentions during motor imagery. These signals hold potential for applications in rehabilitation training and device control. However, the classification accuracy of MI-EEG signals remains a key challenge for the development of BCI technology.MethodsThis paper proposes a composite improved attention convolutional network (CIACNet) for MI-EEG signals classification. CIACNet utilizes a dual-branch convolutional neural network (CNN) to extract rich temporal features, an improved convolutional block attention module (CBAM) to enhance feature extraction, temporal convolutional network (TCN) to capture advanced temporal features, and multi-level feature concatenation for more comprehensive feature representation.ResultsThe CIACNet model performs well on both the BCI IV-2a and BCI IV-2b datasets, achieving accuracies of 85.15 and 90.05%, respectively, with a kappa score of 0.80 on both datasets. These results indicate that the CIACNet model’s classification performance exceeds that of four other comparative models.ConclusionExperimental results demonstrate that the proposed CIACNet model has strong classification capabilities and low time cost. Removing one or more blocks results in a decline in the overall performance of the model, indicating that each block within the model makes a significant contribution to its overall effectiveness. These results demonstrate the ability of the CIACNet model to reduce time costs and improve performance in motor imagery brain-computer interface (MI-BCI) systems, while also highlighting its practical applicability. |
| format | Article |
| id | doaj-art-cba66e87fc0946a386d693cc212582d9 |
| institution | Kabale University |
| issn | 1662-453X |
| language | English |
| publishDate | 2025-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj-art-cba66e87fc0946a386d693cc212582d92025-02-05T14:53:33ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2025-02-011910.3389/fnins.2025.15435081543508A composite improved attention convolutional network for motor imagery EEG classificationWenzhe Liao0Zipeng Miao1Shuaibo Liang2Linyan Zhang3Chen Li4School of Artificial Intelligence, Hebei University of Technology, Tianjin, ChinaSchool of Artificial Intelligence, Hebei University of Technology, Tianjin, ChinaSchool of Artificial Intelligence, Hebei University of Technology, Tianjin, ChinaSchool of Artificial Intelligence, Hebei University of Technology, Tianjin, ChinaTianjin Key Laboratory of Environment, Nutrition and Public Health, Department of Occupational and Environmental Health, School of Public Health, Tianjin Medical University, Tianjin, ChinaIntroductionA brain-computer interface (BCI) is an emerging technology that aims to establish a direct communication pathway between the human brain and external devices. Motor imagery electroencephalography (MI-EEG) signals are analyzed to infer users’ intentions during motor imagery. These signals hold potential for applications in rehabilitation training and device control. However, the classification accuracy of MI-EEG signals remains a key challenge for the development of BCI technology.MethodsThis paper proposes a composite improved attention convolutional network (CIACNet) for MI-EEG signals classification. CIACNet utilizes a dual-branch convolutional neural network (CNN) to extract rich temporal features, an improved convolutional block attention module (CBAM) to enhance feature extraction, temporal convolutional network (TCN) to capture advanced temporal features, and multi-level feature concatenation for more comprehensive feature representation.ResultsThe CIACNet model performs well on both the BCI IV-2a and BCI IV-2b datasets, achieving accuracies of 85.15 and 90.05%, respectively, with a kappa score of 0.80 on both datasets. These results indicate that the CIACNet model’s classification performance exceeds that of four other comparative models.ConclusionExperimental results demonstrate that the proposed CIACNet model has strong classification capabilities and low time cost. Removing one or more blocks results in a decline in the overall performance of the model, indicating that each block within the model makes a significant contribution to its overall effectiveness. These results demonstrate the ability of the CIACNet model to reduce time costs and improve performance in motor imagery brain-computer interface (MI-BCI) systems, while also highlighting its practical applicability.https://www.frontiersin.org/articles/10.3389/fnins.2025.1543508/fullelectroencephalographyconvolution neural networkattention mechanismtemporal convolution networkmotor imageryclassification |
| spellingShingle | Wenzhe Liao Zipeng Miao Shuaibo Liang Linyan Zhang Chen Li A composite improved attention convolutional network for motor imagery EEG classification Frontiers in Neuroscience electroencephalography convolution neural network attention mechanism temporal convolution network motor imagery classification |
| title | A composite improved attention convolutional network for motor imagery EEG classification |
| title_full | A composite improved attention convolutional network for motor imagery EEG classification |
| title_fullStr | A composite improved attention convolutional network for motor imagery EEG classification |
| title_full_unstemmed | A composite improved attention convolutional network for motor imagery EEG classification |
| title_short | A composite improved attention convolutional network for motor imagery EEG classification |
| title_sort | composite improved attention convolutional network for motor imagery eeg classification |
| topic | electroencephalography convolution neural network attention mechanism temporal convolution network motor imagery classification |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2025.1543508/full |
| work_keys_str_mv | AT wenzheliao acompositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT zipengmiao acompositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT shuaiboliang acompositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT linyanzhang acompositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT chenli acompositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT wenzheliao compositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT zipengmiao compositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT shuaiboliang compositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT linyanzhang compositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification AT chenli compositeimprovedattentionconvolutionalnetworkformotorimageryeegclassification |