A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation
IntroductionMotor imagery (MI)-based brain-computer interfaces (BCI) offers promising applications in rehabilitation. Traditional force-based MI-BCI paradigms generally require subjects to imagine constant force during static or dynamic state. It is challenging to meet the demands of dynamic interac...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2025-06-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2025.1591398/full |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849689195141922816 |
|---|---|
| author | Ankai Ying Ankai Ying Ankai Ying Jinwang Lv Jinwang Lv Jinwang Lv Junchen Huang Junchen Huang Junchen Huang Tian Wang Peixin Si Peixin Si Jiyu Zhang Guokun Zuo Guokun Zuo Guokun Zuo Guokun Zuo Jialin Xu Jialin Xu Jialin Xu Jialin Xu |
| author_facet | Ankai Ying Ankai Ying Ankai Ying Jinwang Lv Jinwang Lv Jinwang Lv Junchen Huang Junchen Huang Junchen Huang Tian Wang Peixin Si Peixin Si Jiyu Zhang Guokun Zuo Guokun Zuo Guokun Zuo Guokun Zuo Jialin Xu Jialin Xu Jialin Xu Jialin Xu |
| author_sort | Ankai Ying |
| collection | DOAJ |
| description | IntroductionMotor imagery (MI)-based brain-computer interfaces (BCI) offers promising applications in rehabilitation. Traditional force-based MI-BCI paradigms generally require subjects to imagine constant force during static or dynamic state. It is challenging to meet the demands of dynamic interaction with force intensity variation in MI-BCI systems.MethodsTo address this gap, we designed a novel MI paradigm inspired by daily life, where subjects imagined variations in force intensity during dynamic unilateral upper-limb movements. In a single trial, the subjects were required to complete one of three combinations of force intensity variations: large-to-small, large-to-medium, or medium-to-small. During the execution of this paradigm, electroencephalography (EEG) features exhibit dynamic coupling, with subtle variations in intensity, timing, frequency coverage, and spatial distribution, as the force intensity imagined by the subjects changed. To recognize these fine-grained features, we propose a feature fusion network with a spatial-temporal-enhanced strategy and an information reconstruction (FN-SSIR) algorithm. This model combines a multi-scale spatial-temporal convolution module with a spatial-temporal-enhanced strategy, a convolutional auto-encoder for information reconstruction, and a long short-term memory with self-attention, enabling the comprehensive extraction and fusion of EEG features across fine-grained time-frequency variations and dynamic spatial-temporal patterns.ResultsThe proposed FN-SSIR achieved a classification accuracy of 86.7% ± 6.6% on our force variation MI dataset, and 78.4% ± 13.0% on the BCI Competition IV 2a dataset.DiscussionThese findings highlight the potential of this paradigm and algorithm for advancing MI-BCI systems in rehabilitation training based on dynamic force interactions. |
| format | Article |
| id | doaj-art-9e4d21b8803447c095f1fe4d96cf6abd |
| institution | DOAJ |
| issn | 1662-453X |
| language | English |
| publishDate | 2025-06-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj-art-9e4d21b8803447c095f1fe4d96cf6abd2025-08-20T03:21:43ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2025-06-011910.3389/fnins.2025.15913981591398A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variationAnkai Ying0Ankai Ying1Ankai Ying2Jinwang Lv3Jinwang Lv4Jinwang Lv5Junchen Huang6Junchen Huang7Junchen Huang8Tian Wang9Peixin Si10Peixin Si11Jiyu Zhang12Guokun Zuo13Guokun Zuo14Guokun Zuo15Guokun Zuo16Jialin Xu17Jialin Xu18Jialin Xu19Jialin Xu20Cixi Biomedical Research Institute, Wenzhou Medical University, Ningbo, Zhejiang, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaCixi Biomedical Research Institute, Wenzhou Medical University, Ningbo, Zhejiang, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaCixi Biomedical Research Institute, Wenzhou Medical University, Ningbo, Zhejiang, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaHangzhou RoboCT Technology Development Co., Ltd., Hangzhou, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaHangzhou RoboCT Technology Development Co., Ltd., Hangzhou, ChinaCixi Biomedical Research Institute, Wenzhou Medical University, Ningbo, Zhejiang, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaNingbo College of Materials Engineering, University of Chinese Academy of Sciences, Beijing, ChinaCixi Biomedical Research Institute, Wenzhou Medical University, Ningbo, Zhejiang, ChinaNingbo Institute of Materials Technology and Engineering, Chinese Academy of Sciences, Ningbo, Zhejiang, ChinaNingbo Cixi Institute of Biomedical Engineering, Ningbo, Zhejiang, ChinaNingbo College of Materials Engineering, University of Chinese Academy of Sciences, Beijing, ChinaIntroductionMotor imagery (MI)-based brain-computer interfaces (BCI) offers promising applications in rehabilitation. Traditional force-based MI-BCI paradigms generally require subjects to imagine constant force during static or dynamic state. It is challenging to meet the demands of dynamic interaction with force intensity variation in MI-BCI systems.MethodsTo address this gap, we designed a novel MI paradigm inspired by daily life, where subjects imagined variations in force intensity during dynamic unilateral upper-limb movements. In a single trial, the subjects were required to complete one of three combinations of force intensity variations: large-to-small, large-to-medium, or medium-to-small. During the execution of this paradigm, electroencephalography (EEG) features exhibit dynamic coupling, with subtle variations in intensity, timing, frequency coverage, and spatial distribution, as the force intensity imagined by the subjects changed. To recognize these fine-grained features, we propose a feature fusion network with a spatial-temporal-enhanced strategy and an information reconstruction (FN-SSIR) algorithm. This model combines a multi-scale spatial-temporal convolution module with a spatial-temporal-enhanced strategy, a convolutional auto-encoder for information reconstruction, and a long short-term memory with self-attention, enabling the comprehensive extraction and fusion of EEG features across fine-grained time-frequency variations and dynamic spatial-temporal patterns.ResultsThe proposed FN-SSIR achieved a classification accuracy of 86.7% ± 6.6% on our force variation MI dataset, and 78.4% ± 13.0% on the BCI Competition IV 2a dataset.DiscussionThese findings highlight the potential of this paradigm and algorithm for advancing MI-BCI systems in rehabilitation training based on dynamic force interactions.https://www.frontiersin.org/articles/10.3389/fnins.2025.1591398/fullbrain-computer interfacesforce intensity variationspatial-temporal-enhanced strategymotor imagerydeep learning |
| spellingShingle | Ankai Ying Ankai Ying Ankai Ying Jinwang Lv Jinwang Lv Jinwang Lv Junchen Huang Junchen Huang Junchen Huang Tian Wang Peixin Si Peixin Si Jiyu Zhang Guokun Zuo Guokun Zuo Guokun Zuo Guokun Zuo Jialin Xu Jialin Xu Jialin Xu Jialin Xu A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation Frontiers in Neuroscience brain-computer interfaces force intensity variation spatial-temporal-enhanced strategy motor imagery deep learning |
| title | A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation |
| title_full | A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation |
| title_fullStr | A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation |
| title_full_unstemmed | A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation |
| title_short | A feature fusion network with spatial-temporal-enhanced strategy for the motor imagery of force intensity variation |
| title_sort | feature fusion network with spatial temporal enhanced strategy for the motor imagery of force intensity variation |
| topic | brain-computer interfaces force intensity variation spatial-temporal-enhanced strategy motor imagery deep learning |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2025.1591398/full |
| work_keys_str_mv | AT ankaiying afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT ankaiying afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT ankaiying afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT tianwang afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT peixinsi afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT peixinsi afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jiyuzhang afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu afeaturefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT ankaiying featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT ankaiying featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT ankaiying featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jinwanglv featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT junchenhuang featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT tianwang featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT peixinsi featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT peixinsi featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jiyuzhang featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT guokunzuo featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation AT jialinxu featurefusionnetworkwithspatialtemporalenhancedstrategyforthemotorimageryofforceintensityvariation |