Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG
The quality and quantity of training data are crucial to the performance of a deep-learning-based brain-computer interface (BCI) system. However, it is not practical to record EEG data over several long calibration sessions. A promising time- and cost-efficient solution is artificial data generation...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-03-01
|
| Series: | Frontiers in Human Neuroscience |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnhum.2021.645952/full |
| _version_ | 1819140058499776512 |
|---|---|
| author | Yu Pei Yu Pei Zhiguo Luo Zhiguo Luo Ye Yan Ye Yan Huijiong Yan Huijiong Yan Jing Jiang Weiguo Li Liang Xie Liang Xie Erwei Yin Erwei Yin |
| author_facet | Yu Pei Yu Pei Zhiguo Luo Zhiguo Luo Ye Yan Ye Yan Huijiong Yan Huijiong Yan Jing Jiang Weiguo Li Liang Xie Liang Xie Erwei Yin Erwei Yin |
| author_sort | Yu Pei |
| collection | DOAJ |
| description | The quality and quantity of training data are crucial to the performance of a deep-learning-based brain-computer interface (BCI) system. However, it is not practical to record EEG data over several long calibration sessions. A promising time- and cost-efficient solution is artificial data generation or data augmentation (DA). Here, we proposed a DA method for the motor imagery (MI) EEG signal called brain-area-recombination (BAR). For the BAR, each sample was first separated into two ones (named half-sample) by left/right brain channels, and the artificial samples were generated by recombining the half-samples. We then designed two schemas (intra- and adaptive-subject schema) corresponding to the single- and multi-subject scenarios. Extensive experiments using the classifier of EEGnet were conducted on two public datasets under various training set sizes. In both schemas, the BAR method can make the EEGnet have a better performance of classification (p < 0.01). To make a comparative investigation, we selected two common DA methods (noise-added and flipping), and the BAR method beat them (p < 0.05). Further, using the proposed BAR for augmentation, EEGnet achieved up to 8.3% improvement than a typical decoding algorithm CSP-SVM (p < 0.01), note that both the models were trained on the augmented dataset. This study shows that BAR usage can significantly improve the classification ability of deep learning to MI-EEG signals. To a certain extent, it may promote the development of deep learning technology in the field of BCI. |
| first_indexed | 2024-12-22T11:32:31Z |
| format | Article |
| id | doaj.art-8ce8f7360dff42258fd4fc42a6aabdb7 |
| institution | Directory Open Access Journal |
| issn | 1662-5161 |
| language | English |
| last_indexed | 2024-12-22T11:32:31Z |
| publishDate | 2021-03-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Human Neuroscience |
| spelling | doaj.art-8ce8f7360dff42258fd4fc42a6aabdb72022-12-21T18:27:34ZengFrontiers Media S.A.Frontiers in Human Neuroscience1662-51612021-03-011510.3389/fnhum.2021.645952645952Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEGYu Pei0Yu Pei1Zhiguo Luo2Zhiguo Luo3Ye Yan4Ye Yan5Huijiong Yan6Huijiong Yan7Jing Jiang8Weiguo Li9Liang Xie10Liang Xie11Erwei Yin12Erwei Yin13School of Software, Beihang University, Beijing, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaUnmanned Systems Research Center, National Innovation Institute of Defense Technology, Academy of Military Sciences China, Beijing, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaUnmanned Systems Research Center, National Innovation Institute of Defense Technology, Academy of Military Sciences China, Beijing, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaUnmanned Systems Research Center, National Innovation Institute of Defense Technology, Academy of Military Sciences China, Beijing, ChinaNational Key Laboratory of Human Factors Engineering, China Astronaut Research and Training Center, Beijing, ChinaSchool of Software, Beihang University, Beijing, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaUnmanned Systems Research Center, National Innovation Institute of Defense Technology, Academy of Military Sciences China, Beijing, ChinaTianjin Artificial Intelligence Innovation Center (TAIIC), Tianjin, ChinaUnmanned Systems Research Center, National Innovation Institute of Defense Technology, Academy of Military Sciences China, Beijing, ChinaThe quality and quantity of training data are crucial to the performance of a deep-learning-based brain-computer interface (BCI) system. However, it is not practical to record EEG data over several long calibration sessions. A promising time- and cost-efficient solution is artificial data generation or data augmentation (DA). Here, we proposed a DA method for the motor imagery (MI) EEG signal called brain-area-recombination (BAR). For the BAR, each sample was first separated into two ones (named half-sample) by left/right brain channels, and the artificial samples were generated by recombining the half-samples. We then designed two schemas (intra- and adaptive-subject schema) corresponding to the single- and multi-subject scenarios. Extensive experiments using the classifier of EEGnet were conducted on two public datasets under various training set sizes. In both schemas, the BAR method can make the EEGnet have a better performance of classification (p < 0.01). To make a comparative investigation, we selected two common DA methods (noise-added and flipping), and the BAR method beat them (p < 0.05). Further, using the proposed BAR for augmentation, EEGnet achieved up to 8.3% improvement than a typical decoding algorithm CSP-SVM (p < 0.01), note that both the models were trained on the augmented dataset. This study shows that BAR usage can significantly improve the classification ability of deep learning to MI-EEG signals. To a certain extent, it may promote the development of deep learning technology in the field of BCI.https://www.frontiersin.org/articles/10.3389/fnhum.2021.645952/fullbrain-computer interfaceelectroencephalogrammotor imagerydeep learninginter-subject transfer learningpre-training |
| spellingShingle | Yu Pei Yu Pei Zhiguo Luo Zhiguo Luo Ye Yan Ye Yan Huijiong Yan Huijiong Yan Jing Jiang Weiguo Li Liang Xie Liang Xie Erwei Yin Erwei Yin Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG Frontiers in Human Neuroscience brain-computer interface electroencephalogram motor imagery deep learning inter-subject transfer learning pre-training |
| title | Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG |
| title_full | Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG |
| title_fullStr | Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG |
| title_full_unstemmed | Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG |
| title_short | Data Augmentation: Using Channel-Level Recombination to Improve Classification Performance for Motor Imagery EEG |
| title_sort | data augmentation using channel level recombination to improve classification performance for motor imagery eeg |
| topic | brain-computer interface electroencephalogram motor imagery deep learning inter-subject transfer learning pre-training |
| url | https://www.frontiersin.org/articles/10.3389/fnhum.2021.645952/full |
| work_keys_str_mv | AT yupei dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT yupei dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT zhiguoluo dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT zhiguoluo dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT yeyan dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT yeyan dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT huijiongyan dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT huijiongyan dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT jingjiang dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT weiguoli dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT liangxie dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT liangxie dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT erweiyin dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg AT erweiyin dataaugmentationusingchannellevelrecombinationtoimproveclassificationperformanceformotorimageryeeg |