Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification
Multichannel electroencephalography (EEG) is widely used in typical brain-computer interface (BCI) systems. In general, a number of parameters are essential for a EEG classification algorithm due to redundant features involved in EEG signals. However, the generalization of the EEG method is often ad...
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2018-05-01
|
| Series: | Frontiers in Neuroscience |
| Subjects: | |
| Online Access: | http://journal.frontiersin.org/article/10.3389/fnins.2018.00272/full |
| _version_ | 1819065108177879040 |
|---|---|
| author | Yuwei Zhao Jiuqi Han Yushu Chen Hongji Sun Jiayun Chen Jiayun Chen Ang Ke Ang Ke Yao Han Yao Han Peng Zhang Peng Zhang Yi Zhang Jin Zhou Changyong Wang |
| author_facet | Yuwei Zhao Jiuqi Han Yushu Chen Hongji Sun Jiayun Chen Jiayun Chen Ang Ke Ang Ke Yao Han Yao Han Peng Zhang Peng Zhang Yi Zhang Jin Zhou Changyong Wang |
| author_sort | Yuwei Zhao |
| collection | DOAJ |
| description | Multichannel electroencephalography (EEG) is widely used in typical brain-computer interface (BCI) systems. In general, a number of parameters are essential for a EEG classification algorithm due to redundant features involved in EEG signals. However, the generalization of the EEG method is often adversely affected by the model complexity, considerably coherent with its number of undetermined parameters, further leading to heavy overfitting. To decrease the complexity and improve the generalization of EEG method, we present a novel l1-norm-based approach to combine the decision value obtained from each EEG channel directly. By extracting the information from different channels on independent frequency bands (FB) with l1-norm regularization, the method proposed fits the training data with much less parameters compared to common spatial pattern (CSP) methods in order to reduce overfitting. Moreover, an effective and efficient solution to minimize the optimization object is proposed. The experimental results on dataset IVa of BCI competition III and dataset I of BCI competition IV show that, the proposed method contributes to high classification accuracy and increases generalization performance for the classification of MI EEG. As the training set ratio decreases from 80 to 20%, the average classification accuracy on the two datasets changes from 85.86 and 86.13% to 84.81 and 76.59%, respectively. The classification performance and generalization of the proposed method contribute to the practical application of MI based BCI systems. |
| first_indexed | 2024-12-21T15:41:13Z |
| format | Article |
| id | doaj.art-09899d003a514ad7aac4e52a8eede0f4 |
| institution | Directory Open Access Journal |
| issn | 1662-453X |
| language | English |
| last_indexed | 2024-12-21T15:41:13Z |
| publishDate | 2018-05-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Neuroscience |
| spelling | doaj.art-09899d003a514ad7aac4e52a8eede0f42022-12-21T18:58:30ZengFrontiers Media S.A.Frontiers in Neuroscience1662-453X2018-05-011210.3389/fnins.2018.00272347848Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery ClassificationYuwei Zhao0Jiuqi Han1Yushu Chen2Hongji Sun3Jiayun Chen4Jiayun Chen5Ang Ke6Ang Ke7Yao Han8Yao Han9Peng Zhang10Peng Zhang11Yi Zhang12Jin Zhou13Changyong Wang14Department of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaCollege of Life Science and Technology, Huazhong Agricultural UniversityWuhan, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaNeural Interface & Rehabilitation Technology Research Center, Huazhong University of Science and TechnologyWuhan, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaStem Cell and Tissue Engineering Lab, Beijing Institute of Transfusion MedicineBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaNeural Interface & Rehabilitation Technology Research Center, Huazhong University of Science and TechnologyWuhan, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaDepartment of Neural Engineering and Biological Interdisciplinary Studies, Institute of Military Cognition and Brain Sciences, Academy of Military Medical SciencesBeijing, ChinaMultichannel electroencephalography (EEG) is widely used in typical brain-computer interface (BCI) systems. In general, a number of parameters are essential for a EEG classification algorithm due to redundant features involved in EEG signals. However, the generalization of the EEG method is often adversely affected by the model complexity, considerably coherent with its number of undetermined parameters, further leading to heavy overfitting. To decrease the complexity and improve the generalization of EEG method, we present a novel l1-norm-based approach to combine the decision value obtained from each EEG channel directly. By extracting the information from different channels on independent frequency bands (FB) with l1-norm regularization, the method proposed fits the training data with much less parameters compared to common spatial pattern (CSP) methods in order to reduce overfitting. Moreover, an effective and efficient solution to minimize the optimization object is proposed. The experimental results on dataset IVa of BCI competition III and dataset I of BCI competition IV show that, the proposed method contributes to high classification accuracy and increases generalization performance for the classification of MI EEG. As the training set ratio decreases from 80 to 20%, the average classification accuracy on the two datasets changes from 85.86 and 86.13% to 84.81 and 76.59%, respectively. The classification performance and generalization of the proposed method contribute to the practical application of MI based BCI systems.http://journal.frontiersin.org/article/10.3389/fnins.2018.00272/fullmotor imageryelectroencephalography (EEG)classificationl1-norm regularizationgeneralization |
| spellingShingle | Yuwei Zhao Jiuqi Han Yushu Chen Hongji Sun Jiayun Chen Jiayun Chen Ang Ke Ang Ke Yao Han Yao Han Peng Zhang Peng Zhang Yi Zhang Jin Zhou Changyong Wang Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification Frontiers in Neuroscience motor imagery electroencephalography (EEG) classification l1-norm regularization generalization |
| title | Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification |
| title_full | Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification |
| title_fullStr | Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification |
| title_full_unstemmed | Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification |
| title_short | Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification |
| title_sort | improving generalization based on l1 norm regularization for eeg based motor imagery classification |
| topic | motor imagery electroencephalography (EEG) classification l1-norm regularization generalization |
| url | http://journal.frontiersin.org/article/10.3389/fnins.2018.00272/full |
| work_keys_str_mv | AT yuweizhao improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT jiuqihan improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT yushuchen improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT hongjisun improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT jiayunchen improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT jiayunchen improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT angke improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT angke improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT yaohan improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT yaohan improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT pengzhang improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT pengzhang improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT yizhang improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT jinzhou improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification AT changyongwang improvinggeneralizationbasedonl1normregularizationforeegbasedmotorimageryclassification |