Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding
IntroductionThe time, frequency, and space information of electroencephalogram (EEG) signals is crucial for motor imagery decoding. However, these temporal-frequency-spatial features are high-dimensional small-sample data, which poses significant challenges for motor imagery decoding. Sparse regular...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2023-11-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnins.2023.1292724/full |
| _version_ | 1797639895603216384 |
|---|---|
| author | Shaorong Zhang Shaorong Zhang Qihui Wang Benxin Zhang Zhen Liang Li Zhang Linling Li Gan Huang Zhiguo Zhang Bao Feng Tianyou Yu |
| author_facet | Shaorong Zhang Shaorong Zhang Qihui Wang Benxin Zhang Zhen Liang Li Zhang Linling Li Gan Huang Zhiguo Zhang Bao Feng Tianyou Yu |
| author_sort | Shaorong Zhang |
| collection | DOAJ |
| description | IntroductionThe time, frequency, and space information of electroencephalogram (EEG) signals is crucial for motor imagery decoding. However, these temporal-frequency-spatial features are high-dimensional small-sample data, which poses significant challenges for motor imagery decoding. Sparse regularization is an effective method for addressing this issue. However, the most commonly employed sparse regularization models in motor imagery decoding, such as the least absolute shrinkage and selection operator (LASSO), is a biased estimation method and leads to the loss of target feature information.MethodsIn this paper, we propose a non-convex sparse regularization model that employs the Cauchy function. By designing a proximal gradient algorithm, our proposed model achieves closer-to-unbiased estimation than existing sparse models. Therefore, it can learn more accurate, discriminative, and effective feature information. Additionally, the proposed method can perform feature selection and classification simultaneously, without requiring additional classifiers.ResultsWe conducted experiments on two publicly available motor imagery EEG datasets. The proposed method achieved an average classification accuracy of 82.98% and 64.45% in subject-dependent and subject-independent decoding assessment methods, respectively.ConclusionThe experimental results show that the proposed method can significantly improve the performance of motor imagery decoding, with better classification performance than existing feature selection and deep learning methods. Furthermore, the proposed model shows better generalization capability, with parameter consistency over different datasets and robust classification across different training sample sizes. Compared with existing sparse regularization methods, the proposed method converges faster, and with shorter model training time. |
| first_indexed | 2024-03-11T13:24:06Z |
| format | Article |
| id | doaj.art-000d1a36709d408dafc0b7f0326a9da3 |
| institution | Directory Open Access Journal |
| issn | 1662-453X |
| language | English |
| last_indexed | 2024-03-11T13:24:06Z |
| publishDate | 2023-11-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj.art-000d1a36709d408dafc0b7f0326a9da32023-11-03T08:45:48ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2023-11-011710.3389/fnins.2023.12927241292724Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decodingShaorong Zhang0Shaorong Zhang1Qihui Wang2Benxin Zhang3Zhen Liang4Li Zhang5Linling Li6Gan Huang7Zhiguo Zhang8Bao Feng9Tianyou Yu10Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, School of Biomedical Engineering, Shenzhen University Medical School, Shenzhen, ChinaSchool of Electronic Information and Automation, Guilin University of Aerospace Technology, Guilin, ChinaSchool of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin, ChinaSchool of Electronic Engineering and Automation, Guilin University of Electronic Technology, Guilin, ChinaGuangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, School of Biomedical Engineering, Shenzhen University Medical School, Shenzhen, ChinaGuangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, School of Biomedical Engineering, Shenzhen University Medical School, Shenzhen, ChinaGuangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, School of Biomedical Engineering, Shenzhen University Medical School, Shenzhen, ChinaGuangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, National-Regional Key Technology Engineering Laboratory for Medical Ultrasound, School of Biomedical Engineering, Shenzhen University Medical School, Shenzhen, ChinaInstitute of Computing and Intelligence, Harbin Institute of Technology, Shenzhen, ChinaSchool of Electronic Information and Automation, Guilin University of Aerospace Technology, Guilin, ChinaSchool of Automation Science and Engineering, South China University of Technology, Guangzhou, ChinaIntroductionThe time, frequency, and space information of electroencephalogram (EEG) signals is crucial for motor imagery decoding. However, these temporal-frequency-spatial features are high-dimensional small-sample data, which poses significant challenges for motor imagery decoding. Sparse regularization is an effective method for addressing this issue. However, the most commonly employed sparse regularization models in motor imagery decoding, such as the least absolute shrinkage and selection operator (LASSO), is a biased estimation method and leads to the loss of target feature information.MethodsIn this paper, we propose a non-convex sparse regularization model that employs the Cauchy function. By designing a proximal gradient algorithm, our proposed model achieves closer-to-unbiased estimation than existing sparse models. Therefore, it can learn more accurate, discriminative, and effective feature information. Additionally, the proposed method can perform feature selection and classification simultaneously, without requiring additional classifiers.ResultsWe conducted experiments on two publicly available motor imagery EEG datasets. The proposed method achieved an average classification accuracy of 82.98% and 64.45% in subject-dependent and subject-independent decoding assessment methods, respectively.ConclusionThe experimental results show that the proposed method can significantly improve the performance of motor imagery decoding, with better classification performance than existing feature selection and deep learning methods. Furthermore, the proposed model shows better generalization capability, with parameter consistency over different datasets and robust classification across different training sample sizes. Compared with existing sparse regularization methods, the proposed method converges faster, and with shorter model training time.https://www.frontiersin.org/articles/10.3389/fnins.2023.1292724/fullmotor imageryEEG decodingfeature selectionnonconvex regularizationhigh-dimensional small-sample |
| spellingShingle | Shaorong Zhang Shaorong Zhang Qihui Wang Benxin Zhang Zhen Liang Li Zhang Linling Li Gan Huang Zhiguo Zhang Bao Feng Tianyou Yu Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding Frontiers in Neuroscience motor imagery EEG decoding feature selection nonconvex regularization high-dimensional small-sample |
| title | Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding |
| title_full | Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding |
| title_fullStr | Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding |
| title_full_unstemmed | Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding |
| title_short | Cauchy non-convex sparse feature selection method for the high-dimensional small-sample problem in motor imagery EEG decoding |
| title_sort | cauchy non convex sparse feature selection method for the high dimensional small sample problem in motor imagery eeg decoding |
| topic | motor imagery EEG decoding feature selection nonconvex regularization high-dimensional small-sample |
| url | https://www.frontiersin.org/articles/10.3389/fnins.2023.1292724/full |
| work_keys_str_mv | AT shaorongzhang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT shaorongzhang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT qihuiwang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT benxinzhang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT zhenliang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT lizhang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT linlingli cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT ganhuang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT zhiguozhang cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT baofeng cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding AT tianyouyu cauchynonconvexsparsefeatureselectionmethodforthehighdimensionalsmallsampleprobleminmotorimageryeegdecoding |