1
|
Nam H, Kim JM, Choi W, Bak S, Kam TE. The effects of layer-wise relevance propagation-based feature selection for EEG classification: a comparative study on multiple datasets. Front Hum Neurosci 2023; 17:1205881. [PMID: 37342822 PMCID: PMC10277566 DOI: 10.3389/fnhum.2023.1205881] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/14/2023] [Accepted: 05/17/2023] [Indexed: 06/23/2023] Open
Abstract
Introduction The brain-computer interface (BCI) allows individuals to control external devices using their neural signals. One popular BCI paradigm is motor imagery (MI), which involves imagining movements to induce neural signals that can be decoded to control devices according to the user's intention. Electroencephalography (EEG) is frequently used for acquiring neural signals from the brain in the fields of MI-BCI due to its non-invasiveness and high temporal resolution. However, EEG signals can be affected by noise and artifacts, and patterns of EEG signals vary across different subjects. Therefore, selecting the most informative features is one of the essential processes to enhance classification performance in MI-BCI. Methods In this study, we design a layer-wise relevance propagation (LRP)-based feature selection method which can be easily integrated into deep learning (DL)-based models. We assess its effectiveness for reliable class-discriminative EEG feature selection on two different publicly available EEG datasets with various DL-based backbone models in the subject-dependent scenario. Results and discussion The results show that LRP-based feature selection enhances the performance for MI classification on both datasets for all DL-based backbone models. Based on our analysis, we believe that it can broad its capability to different research domains.
Collapse
Affiliation(s)
| | | | | | | | - Tae-Eui Kam
- Department of Artificial Intelligence, Korea University, Seoul, Republic of Korea
| |
Collapse
|
2
|
Arı E, Taçgın E. Input Shape Effect on Classification Performance of Raw EEG Motor Imagery Signals with Convolutional Neural Networks for Use in Brain-Computer Interfaces. Brain Sci 2023; 13:brainsci13020240. [PMID: 36831784 PMCID: PMC9954790 DOI: 10.3390/brainsci13020240] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/01/2023] [Revised: 01/26/2023] [Accepted: 01/28/2023] [Indexed: 02/04/2023] Open
Abstract
EEG signals are interpreted, analyzed and classified by many researchers for use in brain-computer interfaces. Although there are many different EEG signal acquisition methods, one of the most interesting is motor imagery signals. Many different signal processing methods, machine learning and deep learning models have been developed for the classification of motor imagery signals. Among these, Convolutional Neural Network models generally achieve better results than other models. Because the size and shape of the data is important for training Convolutional Neural Network models and discovering the right relationships, researchers have designed and experimented with many different input shape structures. However, no study has been found in the literature evaluating the effect of different input shapes on model performance and accuracy. In this study, the effects of different input shapes on model performance and accuracy in the classification of EEG motor imagery signals were investigated, which had not been specifically studied before. In addition, signal preprocessing methods, which take a long time before classification, were not used; rather, two CNN models were developed for training and classification using raw data. Two different datasets, BCI Competition IV 2A and 2B, were used in classification processes. For different input shapes, 53.03-89.29% classification accuracy and 2-23 s epoch time were obtained for 2A dataset, 64.84-84.94% classification accuracy and 4-10 s epoch time were obtained for 2B dataset. This study showed that the input shape has a significant effect on the classification performance, and when the correct input shape is selected and the correct CNN architecture is developed, feature extraction and classification can be done well by the CNN architecture without any signal preprocessing.
Collapse
Affiliation(s)
- Emre Arı
- Department of Mechanical Engineering, Faculty of Engineering, Marmara University, Istanbul 34840, Turkey
- Department of Mechanical Engineering, Faculty of Engineering, Dicle University, Diyarbakır 21280, Turkey
- Correspondence:
| | - Ertuğrul Taçgın
- Department of Mechanical Engineering, Faculty of Engineering, Doğuş University, Istanbul 34775, Turkey
| |
Collapse
|
3
|
Wang G, Cerf M. Brain-Computer Interface using neural network and temporal-spectral features. Front Neuroinform 2022; 16:952474. [PMID: 36277476 PMCID: PMC9580359 DOI: 10.3389/fninf.2022.952474] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/25/2022] [Accepted: 08/24/2022] [Indexed: 11/24/2022] Open
Abstract
Brain-Computer Interfaces (BCIs) are increasingly useful for control. Such BCIs can be used to assist individuals who lost mobility or control over their limbs, for recreational purposes such as gaming or semi-autonomous driving, or as an interface toward man-machine integration. Thus far, the performance of algorithms used for thought decoding has been limited. We show that by extracting temporal and spectral features from electroencephalography (EEG) signals and, following, using deep learning neural network to classify those features, one can significantly improve the performance of BCIs in predicting which motor action was imagined by a subject. Our movement prediction algorithm uses Sequential Backward Selection technique to jointly choose temporal and spectral features and a radial basis function neural network for the classification. The method shows an average performance increase of 3.50% compared to state-of-the-art benchmark algorithms. Using two popular public datasets our algorithm reaches 90.08% accuracy (compared to an average benchmark of 79.99%) on the first dataset and 88.74% (average benchmark: 82.01%) on the second dataset. Given the high variability within- and across-subjects in EEG-based action decoding, we suggest that using features from multiple modalities along with neural network classification protocol is likely to increase the performance of BCIs across various tasks.
Collapse
Affiliation(s)
- Gan Wang
- School of Mechanical and Electrical Engineering, Soochow University, Suchow, China
| | - Moran Cerf
- Interdepartmental Neuroscience Program, Northwestern University, Evanston, IL, United States
- *Correspondence: Moran Cerf,
| |
Collapse
|
4
|
Luo J, Mao Q, Wang Y, Shi Z, Hei X. Algorithm Contest of Calibration-free Motor Imagery BCI in the BCI Controlled Robot Contest in World Robot Contest 2021: A survey. BRAIN SCIENCE ADVANCES 2022. [DOI: 10.26599/bsa.2022.9050011] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/08/2022] Open
Abstract
Objective: From September 10 to 13, 2021, the finals of the BCI Controlled Robot Contest in World Robot Contest 2021 were held in Beijing, China. Eleven teams participated in the Algorithm Contest of Calibration-free Motor Imagery BCI. The participants employed both traditional electroencephalograph (EEG) analysis methods and deep learning-based methods in the contest. In this paper, we reviewed the algorithms utilized by the participants, extracted the trends and highlighted interesting approaches from these methods to inform future contests and research recommendations. Method: First, we analyzed the algorithms in separate steps, including EEG channel and signal segment setup, prepossessing technology, and classification model. Then, we emphasized the highlights of each algorithm. Finally, we compared the competition algorithm with the SOTA algorithm. Results: The algorithm employed in the finals performed better than that of the SOTA algorithm. During the final stage of the contest, four of the top five teams used convolutional neural network models, suggesting that with the rapid development of deep learning, convolutional neural network-based models have been the most popular methods in the field of motor imagery BCI.
Collapse
Affiliation(s)
- Jing Luo
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710054, Shaanxi, China
| | - Qi Mao
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710054, Shaanxi, China
| | - Yaojie Wang
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710054, Shaanxi, China
| | - Zhenghao Shi
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710054, Shaanxi, China
| | - Xinhong Hei
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi’an University of Technology, Xi’an 710054, Shaanxi, China
| |
Collapse
|
5
|
An efficient multi-scale CNN model with intrinsic feature integration for motor imagery EEG subject classification in brain-machine interfaces. Biomed Signal Process Control 2022. [DOI: 10.1016/j.bspc.2022.103496] [Citation(s) in RCA: 17] [Impact Index Per Article: 8.5] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/30/2022]
|
6
|
Thenmozhi T, Helen R. Feature Selection Using Extreme Gradient Boosting Bayesian Optimization to upgrade the Classification Performance of Motor Imagery signals for BCI. J Neurosci Methods 2022; 366:109425. [PMID: 34838951 DOI: 10.1016/j.jneumeth.2021.109425] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/01/2021] [Revised: 11/17/2021] [Accepted: 11/20/2021] [Indexed: 11/16/2022]
Abstract
BACKGROUND A motor imagery (MI) based brain computer interface (BCI) is a challenging nonmuscular connection system used to independently perform movement-related tasks. It is gaining increasing prominence in helping paralyzed individuals communicate with the real world. Achieving better classification accuracy is the major concern in the field of motor imagery-based BCI. To upgrade the classification performance, relevant features play a vital role. The relevant features can be selected by the extreme gradient Bayesian optimization (XGBO) method. METHODS In this paper, a combination of time-, frequency-, and spatial-related MI features are extracted to design a reliable MI-BCI system. The proposed method incorporates the XGBO algorithm for feature selection and the random forest for the classification of EEG signals. The potency of the proposed system was investigated using two public EEG datasets (BCI Competition III dataset IIIa and dataset IVa). A novel XGBO algorithm increases the accuracy and reduces the time consumption by reducing the dimensionality of features. The proposed algorithm selects the minimum number of features that increase the computational efficacy for MI-based BCI systems. COMPARISON WITH EXISTING METHODS The proposed method is compared with ANOVA, sequential forward selection, recursive feature elimination, and LASSO methods and the accuracy rate is increased with the lowest computation time. RESULTS The proposed method achieves mean accuracies of 94.44% and 88.72% and classification errors of 5.56% and 11.28% for Datasets IIIa and IVa, respectively. It outperforms four state-of-art methods with 0.87% and 0.59% increases in the accuracy for Datasets IIIa and IVa, respectively.
Collapse
Affiliation(s)
- T Thenmozhi
- Department of Electrical and Electronics Engineering, Thiagarajar College of Engineering, Madurai 625015, India.
| | - R Helen
- Department of Electrical and Electronics Engineering, Thiagarajar College of Engineering, Madurai 625015, India.
| |
Collapse
|
7
|
Mohammadi E, Daneshmand PG, Khorzooghi SMSM. Electroencephalography-Based Brain-Computer Interface Motor Imagery Classification. JOURNAL OF MEDICAL SIGNALS & SENSORS 2022; 12:40-47. [PMID: 35265464 PMCID: PMC8804596 DOI: 10.4103/jmss.jmss_74_20] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/21/2020] [Revised: 12/25/2020] [Accepted: 01/25/2021] [Indexed: 11/04/2022]
Abstract
Background Advances in the medical applications of brain-computer interface, like the motor imagery systems, are highly contributed to making the disabled live better. One of the challenges with such systems is to achieve high classification accuracy. Methods A highly accurate classification algorithm with low computational complexity is proposed here to classify different motor imageries and execution tasks. An experimental study is performed on two electroencephalography datasets (Iranian Brain-Computer Interface competition [iBCIC] dataset and the world BCI Competition IV dataset 2a) to validate the effectiveness of the proposed method. For lower complexity, the common spatial pattern is applied to decrease the 64 channel signal to four components, in addition to increase the class separability. From these components, first, some features are extracted in the time and time-frequency domains, and next, the best linear combination of these is selected by adopting the stepwise linear discriminant analysis (LDA) method, which are then applied in training and testing the classifiers including LDA, random forest, support vector machine, and K nearest neighbors. The classification strategy is of majority voting among the results of the binary classifiers. Results The experimental results indicate that the proposed algorithm accuracy is much higher than that of the winner of the first iBCIC. As to dataset 2a of the world BCI competition IV, the obtained results for subjects 6 and 9 outperform their counterparts. Moreover, this algorithm yields a mean kappa value of 0.53, which is higher than that of the second winner of the competition. Conclusion The results indicate that this method is able to classify motor imagery and execution tasks in both effective and automatic manners.
Collapse
Affiliation(s)
- Ehsan Mohammadi
- Medical Image and Signal Processing Research Centre, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan, Iran
| | - Parisa Ghaderi Daneshmand
- Department of Biomedical Engineering, Student Research Committee, School of Advanced Technologies in Medicine, Isfahan University of Medical Sciences, Isfahan, Iran
| | | |
Collapse
|
8
|
Lashgari E, Ott J, Connelly A, Baldi P, Maoz U. An end-to-end CNN with attentional mechanism applied to raw EEG in a BCI classification task. J Neural Eng 2021; 18. [PMID: 34352734 DOI: 10.1088/1741-2552/ac1ade] [Citation(s) in RCA: 10] [Impact Index Per Article: 3.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/04/2021] [Accepted: 08/05/2021] [Indexed: 11/12/2022]
Abstract
Objective.Motor-imagery (MI) classification base on electroencephalography (EEG) has been long studied in neuroscience and more recently widely used in healthcare applications such as mobile assistive robots and neurorehabilitation. In particular, EEG-based MI classification methods that rely on convolutional neural networks (CNNs) have achieved relatively high classification accuracy. However, naively training CNNs to classify raw EEG data from all channels, especially for high-density EEG, is computationally demanding and requires huge training sets. It often also introduces many irrelevant input features, making it difficult for the CNN to extract the informative ones. This problem is compounded by a dearth of training data, which is particularly acute for MI tasks, because these are cognitively demanding and thus fatigue inducing.Approach.To address these issues, we proposed an end-to-end CNN-based neural network with attentional mechanism together with different data augmentation (DA) techniques. We tested it on two benchmark MI datasets, brain-computer interface (BCI) competition IV 2a and 2b. In addition, we collected a new dataset, recorded using high-density EEG, and containing both MI and motor execution (ME) tasks, which we share with the community.Main results.Our proposed neural-network architecture outperformed all state-of-the-art methods that we found in the literature, with and without DA, reaching an average classification accuracy of 93.6% and 87.83% on BCI 2a and 2b, respectively. We also directly compare decoding of MI and ME tasks. Focusing on MI classification, we find optimal channel configurations and the best DA techniques as well as investigate combining data across participants and the role of transfer learning.Significance.Our proposed approach improves the classification accuracy for MI in the benchmark datasets. In addition, collecting our own dataset enables us to compare MI and ME and investigate various aspects of EEG decoding critical for neuroscience and BCI.
Collapse
Affiliation(s)
- Elnaz Lashgari
- Schmid College of Science and Technology, Chapman University, Orange, CA, United States of America.,Institute for Interdisciplinary Brain and Behavioral Sciences, Chapman University, Orange, CA, United States of America
| | - Jordan Ott
- Department of Computer Science, University of California, Irvine, CA, United States of America
| | - Akima Connelly
- Schmid College of Science and Technology, Chapman University, Orange, CA, United States of America.,Institute for Interdisciplinary Brain and Behavioral Sciences, Chapman University, Orange, CA, United States of America
| | - Pierre Baldi
- Department of Computer Science, University of California, Irvine, CA, United States of America.,Center for Machine Learning and Intelligent Systems, University of California Irvine, Irvine, CA, United States of America.,Institute for Genomics and Bioinformatics, University of California Irvine, Irvine, CA, United States of America
| | - Uri Maoz
- Schmid College of Science and Technology, Chapman University, Orange, CA, United States of America.,Institute for Interdisciplinary Brain and Behavioral Sciences, Chapman University, Orange, CA, United States of America.,Computational Neuroscience and Psychology, Crean College of Health and Behavioral Sciences, Chapman University, Orange, CA, United States of America.,Fowler School of Engineering, Chapman University, Orange, CA, United States of America.,Anderson School of Management, University of California Los Angeles, Los Angeles, CA, United States of America.,Biology and Bioengineering, California Institute of Technology, Pasadena, CA, United States of America
| |
Collapse
|
9
|
Radman M, Chaibakhsh A, Nariman-zadeh N, He H. Feature fusion for improving performance of motor imagery brain-computer interface system. Biomed Signal Process Control 2021. [DOI: 10.1016/j.bspc.2021.102763] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/16/2022]
|
10
|
Kumar S, Tsunoda T, Sharma A. SPECTRA: a tool for enhanced brain wave signal recognition. BMC Bioinformatics 2021; 22:195. [PMID: 34078274 PMCID: PMC8170968 DOI: 10.1186/s12859-021-04091-x] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Abstract] [Key Words] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 03/15/2021] [Accepted: 03/21/2021] [Indexed: 12/31/2022] Open
Abstract
Background Brain wave signal recognition has gained increased attention in neuro-rehabilitation applications. This has driven the development of brain–computer interface (BCI) systems. Brain wave signals are acquired using electroencephalography (EEG) sensors, processed and decoded to identify the category to which the signal belongs. Once the signal category is determined, it can be used to control external devices. However, the success of such a system essentially relies on significant feature extraction and classification algorithms. One of the commonly used feature extraction technique for BCI systems is common spatial pattern (CSP). Results The performance of the proposed spatial-frequency-temporal feature extraction (SPECTRA) predictor is analysed using three public benchmark datasets. Our proposed predictor outperformed other competing methods achieving lowest average error rates of 8.55%, 17.90% and 20.26%, and highest average kappa coefficient values of 0.829, 0.643 and 0.595 for BCI Competition III dataset IVa, BCI Competition IV dataset I and BCI Competition IV dataset IIb, respectively.
Conclusions Our proposed SPECTRA predictor effectively finds features that are more separable and shows improvement in brain wave signal recognition that can be instrumental in developing improved real-time BCI systems that are computationally efficient.
Collapse
Affiliation(s)
- Shiu Kumar
- School of Electrical and Electronics Engineering, Fiji National University, Suva, Fiji.
| | - Tatsuhiko Tsunoda
- Laboratory for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, 230-0045, Japan.,Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical and Dental University (TMDU), Tokyo, 113-8510, Japan.,Laboratory for Medical Science Mathematics, Department of Biological Sciences, Graduate School of Science, University of Tokyo, Tokyo, 113-0033, Japan
| | - Alok Sharma
- Laboratory for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, 230-0045, Japan.,Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical and Dental University (TMDU), Tokyo, 113-8510, Japan.,School of Engineering and Physics, The University of the South Pacific, Suva, Fiji.,Institute for Integrated and Intelligent Systems, Griffith University, Nathan, Brisbane, QLD, Australia
| |
Collapse
|
11
|
Dagdevir E, Tokmakci M. Optimization of preprocessing stage in EEG based BCI systems in terms of accuracy and timing cost. Biomed Signal Process Control 2021. [DOI: 10.1016/j.bspc.2021.102548] [Citation(s) in RCA: 15] [Impact Index Per Article: 5.0] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/25/2022]
|
12
|
Li R, Wang L, Sourina O. Subject matching for cross-subject EEG-based recognition of driver states related to situation awareness. Methods 2021; 202:136-143. [PMID: 33845126 DOI: 10.1016/j.ymeth.2021.04.009] [Citation(s) in RCA: 4] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Key Words] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/02/2021] [Revised: 04/05/2021] [Accepted: 04/07/2021] [Indexed: 10/21/2022] Open
Abstract
Situation awareness (SA) has received much attention in recent years because of its importance for operators of dynamic systems. Electroencephalography (EEG) can be used to measure mental states of operators related to SA. However, cross-subject EEG-based SA recognition is a critical challenge, as data distributions of different subjects vary significantly. Subject variability is considered as a domain shift problem. Several attempts have been made to find domain-invariant features among subjects, where subject-specific information is neglected. In this work, we propose a simple but efficient subject matching framework by finding a connection between a target (test) subject and source (training) subjects. Specifically, the framework includes two stages: (1) we train the model with multi-source domain alignment layers to collect source domain statistics. (2) During testing, a distance is computed to perform subject matching in the latent representation space. We use a reciprocal exponential function as a similarity measure to dynamically select similar source subjects. Experiment results show that our framework achieves a state-of-the-art accuracy 74.32% for the Taiwan driving dataset.
Collapse
Affiliation(s)
- Ruilin Li
- School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore; Fraunhofer Singapore, Singapore 639798, Singapore
| | - Lipo Wang
- School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore 639798, Singapore.
| | - Olga Sourina
- Fraunhofer Singapore, Singapore 639798, Singapore.
| |
Collapse
|
13
|
Xiao X, Fang Y. Motor Imagery EEG Signal Recognition Using Deep Convolution Neural Network. Front Neurosci 2021; 15:655599. [PMID: 33841094 PMCID: PMC8027090 DOI: 10.3389/fnins.2021.655599] [Citation(s) in RCA: 8] [Impact Index Per Article: 2.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/19/2021] [Accepted: 03/03/2021] [Indexed: 11/21/2022] Open
Abstract
Brain computer interaction (BCI) based on EEG can help patients with limb dyskinesia to carry out daily life and rehabilitation training. However, due to the low signal-to-noise ratio and large individual differences, EEG feature extraction and classification have the problems of low accuracy and efficiency. To solve this problem, this paper proposes a recognition method of motor imagery EEG signal based on deep convolution network. This method firstly aims at the problem of low quality of EEG signal characteristic data, and uses short-time Fourier transform (STFT) and continuous Morlet wavelet transform (CMWT) to preprocess the collected experimental data sets based on time series characteristics. So as to obtain EEG signals that are distinct and have time-frequency characteristics. And based on the improved CNN network model to efficiently recognize EEG signals, to achieve high-quality EEG feature extraction and classification. Further improve the quality of EEG signal feature acquisition, and ensure the high accuracy and precision of EEG signal recognition. Finally, the proposed method is validated based on the BCI competiton dataset and laboratory measured data. Experimental results show that the accuracy of this method for EEG signal recognition is 0.9324, the precision is 0.9653, and the AUC is 0.9464. It shows good practicality and applicability.
Collapse
Affiliation(s)
- Xiongliang Xiao
- College of Electronic Science and Engineering, Hunan University of Information Technology, Changsha, China
| | - Yuee Fang
- College of Electric Power Engineering, Hunan Polytechnic of Water Resources and Electric Power, Changsha, China
| |
Collapse
|
14
|
Zhang C, Kim YK, Eskandarian A. EEG-inception: an accurate and robust end-to-end neural network for EEG-based motor imagery classification. J Neural Eng 2021; 18. [PMID: 33691299 DOI: 10.1088/1741-2552/abed81] [Citation(s) in RCA: 38] [Impact Index Per Article: 12.7] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/26/2020] [Accepted: 03/10/2021] [Indexed: 12/14/2022]
Abstract
Classification of EEG-based motor imagery (MI) is a crucial non-invasive application in brain-computer interface (BCI) research. This paper proposes a novel convolutional neural network (CNN) architecture for accurate and robust EEG-based MI classification that outperforms the state-of-the-art methods. The proposed CNN model, namely EEG-Inception, is built on the backbone of the Inception-Time network, which has showed to be highly efficient and accurate for time-series classification. Also, the proposed network is an end-to-end classification, as it takes the raw EEG signals as the input and does not require complex EEG signal-preprocessing. Furthermore, this paper proposes a novel data augmentation method for EEG signals to enhance the accuracy, at least by 3%, and reduce overfitting with limited BCI datasets. The proposed model outperforms all state-of-the-art methods by achieving the average accuracy of 88.4% and 88.6% on the 2008 BCI Competition IV 2a (four-classes) and 2b datasets (binary-classes), respectively. Furthermore, it takes less than 0.025 seconds to test a sample suitable for real-time processing. Moreover, the classification standard deviation for nine different subjects achieves the lowest value of 5.5 for the 2b dataset and 7.1 for the 2a dataset, which validates that the proposed method is highly robust. From the experiment results, it can be inferred that the EEG-Inception network exhibits a strong potential as a subject-independent classifier for EEG-based MI tasks.
Collapse
Affiliation(s)
- Ce Zhang
- Mechanical Engineering, Virginia Polytechnic Institute and State University, 635 Prices Fork Rd, Blacksburg, Virginia, 24061-0131, UNITED STATES
| | - Young-Keun Kim
- Mechanical and Control Eng. Dept, Handong Global University, 558 Handong-ro Buk-gu, Pohang Gyeongbuk, Pohang, 37554, Korea (the Republic of)
| | - Azim Eskandarian
- Mechanical Engineering, Virginia Polytechnic Institute and State University, 635 Prices Fork Rd, Blacksburg, Virginia, 24060, UNITED STATES
| |
Collapse
|
15
|
Yang L, Song Y, Ma K, Xie L. Motor Imagery EEG Decoding Method Based on a Discriminative Feature Learning Strategy. IEEE Trans Neural Syst Rehabil Eng 2021; 29:368-379. [PMID: 33460382 DOI: 10.1109/tnsre.2021.3051958] [Citation(s) in RCA: 28] [Impact Index Per Article: 9.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/11/2022]
Abstract
With the rapid development of deep learning, more and more deep learning-based motor imagery electroencephalograph (EEG) decoding methods have emerged in recent years. However, the existing deep learning-based methods usually only adopt the constraint of classification loss, which hardly obtains the features with high discrimination and limits the improvement of EEG decoding accuracy. In this paper, a discriminative feature learning strategy is proposed to improve the discrimination of features, which includes the central distance loss (CD-loss), the central vector shift strategy, and the central vector update process. First, the CD-loss is proposed to make the same class of samples converge to the corresponding central vector. Then, the central vector shift strategy extends the distance between different classes of samples in the feature space. Finally, the central vector update process is adopted to avoid the non-convergence of CD-loss and weaken the influence of the initial value of central vectors on the final results. In addition, overfitting is another severe challenge for deep learning-based EEG decoding methods. To deal with this problem, a data augmentation method based on circular translation strategy is proposed to expand the experimental datasets without introducing any extra noise or losing any information of the original data. To validate the effectiveness of the proposed method, we conduct some experiments on two public motor imagery EEG datasets (BCI competition IV 2a and 2b dataset), respectively. The comparison with current state-of-the-art methods indicates that our method achieves the highest average accuracy and good stability on the two experimental datasets.
Collapse
|
16
|
Yang L, Song Y, Ma K, Su E, Xie L. A novel motor imagery EEG decoding method based on feature separation. J Neural Eng 2021; 18. [PMID: 33545691 DOI: 10.1088/1741-2552/abe39b] [Citation(s) in RCA: 6] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/31/2020] [Accepted: 02/05/2021] [Indexed: 11/11/2022]
Abstract
OBJECTIVE Motor imagery electroencephalography (EEG) decoding is a vital technology for the brain-computer interface (BCI) systems and has been widely studied in recent years. However, the original EEG signals usually contain a lot of class-independent information, and the existing motor imagery EEG decoding methods are easily interfered by this irrelevant information, which greatly limits the decoding accuracy of these methods. APPROACH To overcome the interference of the class-independent information, a motor imagery EEG decoding method based on feature separation is proposed in this paper. Furthermore, a feature separation network based on adversarial learning (FSNAL) is designed for the feature separation of the original EEG samples. First, the class-related features and class-independent features are separated by the proposed FSNAL framework, and then motor imagery EEG decoding is performed only according to the class-related features to avoid the adverse effects of class-independent features. MAIN RESULTS To validate the effectiveness of the proposed motor imagery EEG decoding method, we conduct some experiments on two public EEG datasets (the BCI competition IV 2a and 2b datasets). The experimental results comparison between our method and some state-of-the-art methods demonstrates that our motor imagery EEG decoding method outperforms all the compared methods on the two experimental datasets. SIGNIFICANCE Our motor imagery EEG decoding method can alleviate the interference of class-independent features, and it has great application potential for improving the performance of motor imagery BCI systems in the near future.
Collapse
Affiliation(s)
- Lie Yang
- Shien-Ming Wu School of Intelligent Engineering, South China University of Technology, No.777 East Xingye Avenue, Panyu District, Guangzhou, Guangzhou, 510460, CHINA
| | - Yonghao Song
- Shien-Ming Wu School of Intelligent Engineering, South China University of Technology, No.777 East Xingye Avenue, Panyu District, Guangzhou, Guangzhou, Guangdong, 510460, CHINA
| | - Ke Ma
- Sun Yat-Sen University Zhongshan Ophthalmic Center State Key Laboratory of Ophthalmology, No.54 Xianlie South Road, Yuexiu District, Guangzhou, Guangzhou, Guangdong, 510000, CHINA
| | - Enze Su
- Shien-Ming Wu School of Intelligent Engineering, South China University of Technology, No.777 East Xingye Avenue, Panyu District, Guangzhou, Guangzhou, Guangdong, 510460, CHINA
| | - Longhan Xie
- Shien-Ming Wu School of Intelligent Engineering, South China University of Technology, No.777 East Xingye Avenue, Panyu District, Guangzhou, Guangzhou, Guangdong, 510460, CHINA
| |
Collapse
|
17
|
Kumar S, Sharma R, Sharma A. OPTICAL+: a frequency-based deep learning scheme for recognizing brain wave signals. PeerJ Comput Sci 2021; 7:e375. [PMID: 33817023 PMCID: PMC7959638 DOI: 10.7717/peerj-cs.375] [Citation(s) in RCA: 12] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/10/2020] [Accepted: 01/06/2021] [Indexed: 06/12/2023]
Abstract
A human-computer interaction (HCI) system can be used to detect different categories of the brain wave signals that can be beneficial for neurorehabilitation, seizure detection and sleep stage classification. Research on developing HCI systems using brain wave signals has progressed a lot over the years. However, real-time implementation, computational complexity and accuracy are still a concern. In this work, we address the problem of selecting the appropriate filtering frequency band while also achieving a good system performance by proposing a frequency-based approach using long short-term memory network (LSTM) for recognizing different brain wave signals. Adaptive filtering using genetic algorithm is incorporated for a hybrid system utilizing common spatial pattern and LSTM network. The proposed method (OPTICAL+) achieved an overall average classification error rate of 30.41% and a kappa coefficient value of 0.398, outperforming the state-of-the-art methods. The proposed OPTICAL+ predictor can be used to develop improved HCI systems that will aid in neurorehabilitation and may also be beneficial for sleep stage classification and seizure detection.
Collapse
Affiliation(s)
- Shiu Kumar
- School of Electrical and Electronic Engineering, Fiji National University, Suva, Fiji
| | - Ronesh Sharma
- School of Electrical and Electronic Engineering, Fiji National University, Suva, Fiji
| | - Alok Sharma
- STEMP, University of the South Pacific, Suva, Fiji
- Institute for Integrated and Intelligent Systems, Griffith University, Brisbane, Australia
- Laboratory for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa, Japan
| |
Collapse
|
18
|
Chen X, Tao X, Wang FL, Xie H. Global research on artificial intelligence-enhanced human electroencephalogram analysis. Neural Comput Appl 2021. [DOI: 10.1007/s00521-020-05588-x] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.3] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/16/2022]
|
19
|
Shamsi F, Haddad A, Najafizadeh L. Early classification of motor tasks using dynamic functional connectivity graphs from EEG. J Neural Eng 2020; 18. [PMID: 33246319 DOI: 10.1088/1741-2552/abce70] [Citation(s) in RCA: 10] [Impact Index Per Article: 2.5] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/06/2020] [Accepted: 11/27/2020] [Indexed: 12/13/2022]
Abstract
OBJECTIVE Classification of electroencephalography (EEG) signals with high accuracy using short recording intervals has been a challenging problem in developing brain computer interfaces (BCIs). This paper presents a novel feature extraction method for EEG recordings to tackle this problem. APPROACH The proposed approach is based on the concept that the brain functions in a dynamic manner, and utilizes dynamic functional connectivity graphs. The EEG data is first segmented into intervals during which functional networks sustain their connectivity. Functional connectivity networks for each identified segment are then localized, and graphs are constructed, which will be used as features. To take advantage of the dynamic nature of the generated graphs, a Long Short Term Memory (LSTM) classifier is employed for classification. MAIN RESULTS Features extracted from various durations of post-stimulus EEG data associated with motor execution and imagery tasks are used to test the performance of the classifier. Results show an average accuracy of 85.32% using features extracted from only 500 ms of the post-stimulus data. SIGNIFICANCE Our results demonstrate, for the first time, that using the proposed feature extraction method, it is possible to classify motor tasks from EEG recordings using a short interval of the data in the order of hundreds of milliseconds (e.g. 500 ms). This duration is considerably shorter than what has been reported before. These results will have significant implications for improving the effectiveness and the speed of BCIs, particularly for those used in assistive technologies.
Collapse
Affiliation(s)
- Foroogh Shamsi
- Electrical and Computer Engineering, Rutgers University, 94 Brett Rd, New Brunswick, New Jersey, NJ 08854, UNITED STATES
| | - Ali Haddad
- Electrical and Computer Engineering, Rutgers University, 94 Brett Rd, New Brunswick, New Jersey, NJ 08854, UNITED STATES
| | - Laleh Najafizadeh
- Electrical and Computer Engineering, Rutgers University, 94 Brett Rd, New Brunswick, New Jersey, 08901-8554, UNITED STATES
| |
Collapse
|
20
|
Luo J, Gao X, Zhu X, Wang B, Lu N, Wang J. Motor imagery EEG classification based on ensemble support vector learning. COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE 2020; 193:105464. [PMID: 32283387 DOI: 10.1016/j.cmpb.2020.105464] [Citation(s) in RCA: 19] [Impact Index Per Article: 4.8] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Subscribe] [Scholar Register] [Received: 10/09/2019] [Revised: 02/27/2020] [Accepted: 03/18/2020] [Indexed: 06/11/2023]
Abstract
BACKGROUND AND OBJECTIVE Brain-computer interfaces build a communication pathway from the human brain to a computer. Motor imagery-based electroencephalogram (EEG) classification is a widely applied paradigm in brain-computer interfaces. The common spatial pattern, based on the event-related desynchronization (ERD)/event-related synchronization (ERS) phenomenon, is one of the most popular algorithms for motor imagery-based EEG classification. Moreover, the spatiotemporal discrepancy feature based on the event-related potential phenomenon has been demonstrated to provide complementary information to ERD/ERS-based features. In this paper, aiming to improve the performance of motor imagery-based EEG classification in a few-channel situation, an ensemble support vector learning (ESVL)-based approach is proposed to combine the advantages of the ERD/ERS-based features and the event-related potential-based features in motor imagery-based EEG classification. METHODS ESVL is an ensemble learning algorithm based on support vector machine classifier. Specifically, the decision boundary with the largest interclass margin is obtained using the support vector machine algorithm, and the distances between sample points and the decision boundary are mapped to posterior probabilities. The probabilities obtained from different support vector machine classifiers are combined to make prediction. Thus, ESVL leverages the advantages of multiple trained support vector machine classifiers and makes a better prediction based on the posterior probabilities. The class discrepancy-guided sub-band-based common spatial pattern and the spatiotemporal discrepancy feature are applied to extract discriminative features, and then, the extracted features are used to train the ESVL classifier and make predictions. RESULTS The BCI Competition IV datasets 2a and 2b are employed to evaluate the performance of the proposed ESVL algorithm. Experimental comparisons with the state-of-the-art methods are performed, and the proposed ESVL-based approach achieves an average max kappa value of 0.60 and 0.71 on BCI Competition IV datasets 2a and 2b respectively. The results show that the proposed ESVL-based approach improves the performance of motor imagery-based brain-computer interfaces. CONCLUSION The proposed ESVL classifier could use the posterior probabilities to realize ensemble learning and the ESVL-based motor imagery classification approach takes advantage of the merits of ERD/ERS based feature and event-related potential based feature to improve the experimental performance.
Collapse
Affiliation(s)
- Jing Luo
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi'an University of Technology, Xi'an Shaanxi, China.
| | - Xing Gao
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi'an University of Technology, Xi'an Shaanxi, China
| | - Xiaobei Zhu
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi'an University of Technology, Xi'an Shaanxi, China
| | - Bin Wang
- Shaanxi Key Laboratory for Network Computing and Security Technology, School of Computer Science and Engineering, Xi'an University of Technology, Xi'an Shaanxi, China
| | - Na Lu
- State Key Laboratory for Manufacturing Systems Engineering, Systems Engineering Institute, Xi'an Jiaotong University, Xi'an, Shaanxi, China
| | - Jie Wang
- State Key Laboratory for Manufacturing Systems Engineering, Systems Engineering Institute, Xi'an Jiaotong University, Xi'an, Shaanxi, China
| |
Collapse
|
21
|
Dai G, Zhou J, Huang J, Wang N. HS-CNN: a CNN with hybrid convolution scale for EEG motor imagery classification. J Neural Eng 2020; 17:016025. [PMID: 31476743 DOI: 10.1088/1741-2552/ab405f] [Citation(s) in RCA: 86] [Impact Index Per Article: 21.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/03/2023]
Abstract
OBJECTIVE Electroencephalography (EEG) motor imagery classification has been widely used in healthcare applications such as mobile assistive robots and post-stroke rehabilitation. Recently, EEG motor imagery classification methods based on convolutional neural networks (CNNs) have been proposed and have achieved relatively high classification accuracy. However, these methods use single convolution scale in the CNN, while the best convolution scale differs from subject to subject. This limits the classification accuracy. Another issue is that the classification accuracy degrades when training data is limited. APPROACH To address these issues, we have proposed a hybrid-scale CNN architecture with a data augmentation method for EEG motor imagery classification. MAIN RESULTS Compared with several state-of-the-art methods, the proposed method achieves an average classification accuracy of 91.57% and 87.6% on two commonly used datasets, which outperforms several state-of-the-art EEG motor imagery classification methods. SIGNIFICANCE The proposed method effectively addresses the issues of existing CNN-based EEG motor imagery classification methods and improves the classification accuracy.
Collapse
Affiliation(s)
- Guanghai Dai
- School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, Sichuan, People's Republic of China
| | | | | | | |
Collapse
|
22
|
Kumar S, Sharma A, Tsunoda T. Brain wave classification using long short-term memory network based OPTICAL predictor. Sci Rep 2019; 9:9153. [PMID: 31235800 PMCID: PMC6591300 DOI: 10.1038/s41598-019-45605-1] [Citation(s) in RCA: 44] [Impact Index Per Article: 8.8] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/11/2018] [Accepted: 06/07/2019] [Indexed: 11/09/2022] Open
Abstract
Brain-computer interface (BCI) systems having the ability to classify brain waves with greater accuracy are highly desirable. To this end, a number of techniques have been proposed aiming to be able to classify brain waves with high accuracy. However, the ability to classify brain waves and its implementation in real-time is still limited. In this study, we introduce a novel scheme for classifying motor imagery (MI) tasks using electroencephalography (EEG) signal that can be implemented in real-time having high classification accuracy between different MI tasks. We propose a new predictor, OPTICAL, that uses a combination of common spatial pattern (CSP) and long short-term memory (LSTM) network for obtaining improved MI EEG signal classification. A sliding window approach is proposed to obtain the time-series input from the spatially filtered data, which becomes input to the LSTM network. Moreover, instead of using LSTM directly for classification, we use regression based output of the LSTM network as one of the features for classification. On the other hand, linear discriminant analysis (LDA) is used to reduce the dimensionality of the CSP variance based features. The features in the reduced dimensional plane after performing LDA are used as input to the support vector machine (SVM) classifier together with the regression based feature obtained from the LSTM network. The regression based feature further boosts the performance of the proposed OPTICAL predictor. OPTICAL showed significant improvement in the ability to accurately classify left and right-hand MI tasks on two publically available datasets. The improvements in the average misclassification rates are 3.09% and 2.07% for BCI Competition IV Dataset I and GigaDB dataset, respectively. The Matlab code is available at https://github.com/ShiuKumar/OPTICAL .
Collapse
Affiliation(s)
- Shiu Kumar
- The University of the South Pacific, Suva, Fiji. .,Fiji National University, Suva, Fiji.
| | - Alok Sharma
- Institute for Integrated and Intelligent Systems, Griffith University, Brisbane, QLD-4111, Australia. .,Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan. .,Laboratory for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa, Japan. .,The University of the South Pacific, Suva, Fiji.
| | - Tatsuhiko Tsunoda
- Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan.,Laboratory for Medical Science Mathematics, RIKEN Center for Integrative Medical Sciences, Yokohama, Kanagawa, Japan.,CREST, JST, Tokyo, 102-8666, Japan
| |
Collapse
|
23
|
Luo J, Wang J, Xu R, Xu K. Class discrepancy-guided sub-band filter-based common spatial pattern for motor imagery classification. J Neurosci Methods 2019; 323:98-107. [PMID: 31141703 DOI: 10.1016/j.jneumeth.2019.05.011] [Citation(s) in RCA: 13] [Impact Index Per Article: 2.6] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/02/2019] [Revised: 05/21/2019] [Accepted: 05/22/2019] [Indexed: 10/26/2022]
Abstract
BACKGROUND Motor imagery classification, an important branch of brain-computer interface (BCI), recognizes the intention of subjects to control external auxiliary equipment. Therefore, EEG-based motor imagery classification has received increasing attention in the fields of neuroscience. The common spatial pattern (CSP) algorithm has recently achieved great success in motor imagery classification. However, varying discriminative frequency bands and few-channel EEG limit the performance of CSP. NEW METHOD A class discrepancy-guided sub-band filter-based CSP (CDFCSP) algorithm is proposed to automatically recognize and augment the discriminative frequency bands for CSP algorithms. Specifically, a priori knowledge and templates obtained from the training set were applied as the design guidelines of the class discrepancy-guided sub-band filter (CDF). Second, a filter bank CSP was used to extract features from EEG traces filtered by the CDF. Finally, the CSP features of multiple frequency bands were leveraged to train linear support vector machine classifier and generate prediction. RESULTS BCI competition IV datasets 2a and 2b, which include EEGs from 18 subjects, were used to validate the performance improvement provided by the CDF. Student's t-tests of the CDFCSP versus the filter bank CSP without the CDF showed that the performance improvement was significant (i.e., p-values of 0.040 and 0.032 for the ratio and normalization mode CDFCSP, respectively). COMPARISON WITH EXISTING METHOD(S) The experiments show that the proposed CDFCSP improves the CSP algorithm and outperforms the other state-of-the-art algorithms evaluated in this paper. CONCLUSIONS The increased performance of the proposed CDFCSP algorithm can promote the application of BCI systems.
Collapse
Affiliation(s)
- Jing Luo
- School of Computer Science and Engineering, Xi'an University of Technology, Xi'an, Shaanxi, China; Shaanxi Key Laboratory for Network Computing and Security Technology, Xi'an University of Technology, Xi'an, Shaanxi, China.
| | - Jie Wang
- State Key Laboratory for Manufacturing Systems Engineering, Systems Engineering Institute, Xi'an Jiaotong University, Xi'an, Shaanxi, China
| | - Rong Xu
- Shaanxi Province Institute of Water Resources and Electric Power Investigation and Design, Xi'an, Shaanxi, China
| | - Kailiang Xu
- School of Automation and Information Engineering, Xi'an University of Technology, Xi'an, Shaanxi, China
| |
Collapse
|
24
|
An Incremental Version of L-MVU for the Feature Extraction of MI-EEG. COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2019; 2019:4317078. [PMID: 31191631 PMCID: PMC6525943 DOI: 10.1155/2019/4317078] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.2] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 02/27/2019] [Revised: 04/04/2019] [Accepted: 04/07/2019] [Indexed: 11/17/2022]
Abstract
Due to the nonlinear and high-dimensional characteristics of motor imagery electroencephalography (MI-EEG), it can be challenging to get high online accuracy. As a nonlinear dimension reduction method, landmark maximum variance unfolding (L-MVU) can completely retain the nonlinear features of MI-EEG. However, L-MVU still requires considerable computation costs for out-of-sample data. An incremental version of L-MVU (denoted as IL-MVU) is proposed in this paper. The low-dimensional representation of the training data is generated by L-MVU. For each out-of-sample data, its nearest neighbors will be found in the high-dimensional training samples and the corresponding reconstruction weight matrix be calculated to generate its low-dimensional representation as well. IL-MVU is further combined with the dual-tree complex wavelet transform (DTCWT), which develops a hybrid feature extraction method (named as IL-MD). IL-MVU is applied to extract the nonlinear features of the specific subband signals, which are reconstructed by DTCWT and have the obvious event-related synchronization/event-related desynchronization phenomenon. The average energy features of α and β waves are calculated simultaneously. The two types of features are fused and are evaluated by a linear discriminant analysis classifier. Based on the two public datasets with 12 subjects, extensive experiments were conducted. The average recognition accuracies of 10-fold cross-validation are 92.50% on Dataset 3b and 88.13% on Dataset 2b, and they gain at least 1.43% and 3.45% improvement, respectively, compared to existing methods. The experimental results show that IL-MD can extract more accurate features with relatively lower consumption cost, and it also has better feature visualization and self-adaptive characteristics to subjects. The t-test results and Kappa values suggest the proposed feature extraction method reaches statistical significance and has high consistency in classification.
Collapse
|
25
|
Yang S, Yin Z, Wang Y, Zhang W, Wang Y, Zhang J. Assessing cognitive mental workload via EEG signals and an ensemble deep learning classifier based on denoising autoencoders. Comput Biol Med 2019; 109:159-170. [PMID: 31059900 DOI: 10.1016/j.compbiomed.2019.04.034] [Citation(s) in RCA: 21] [Impact Index Per Article: 4.2] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/15/2019] [Revised: 04/26/2019] [Accepted: 04/26/2019] [Indexed: 10/26/2022]
Abstract
To estimate the reliability and cognitive states of operator performance in a human-machine collaborative environment, we propose a novel human mental workload (MW) recognizer based on deep learning principles and utilizing the features of the electroencephalogram (EEG). To determine personalized properties in high dimensional EEG indicators, we introduce a feature mapping layer in stacked denoising autoencoder (SDAE) that is capable of preserving the local information in EEG dynamics. The ensemble classifier is then built via the subject-specific integrated deep learning committee, and adapts to the cognitive properties of a specific human operator and alleviates inter-subject feature variations. We validate our algorithms and the ensemble SDAE classifier with local information preservation (denoted by EL-SDAE) on an EEG database collected during the execution of complex human-machine tasks. The classification performance indicates that the EL-SDAE outperforms several classical MW estimators when its optimal network architecture has been identified.
Collapse
Affiliation(s)
- Shuo Yang
- School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, PR China; Engineering Research Center of Optical Instrument and System, Ministry of Education, Shanghai Key Lab of Modern Optical System, University of Shanghai for Science and Technology, Shanghai, 200093, PR China
| | - Zhong Yin
- School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, PR China; Engineering Research Center of Optical Instrument and System, Ministry of Education, Shanghai Key Lab of Modern Optical System, University of Shanghai for Science and Technology, Shanghai, 200093, PR China.
| | - Yagang Wang
- School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, PR China; Engineering Research Center of Optical Instrument and System, Ministry of Education, Shanghai Key Lab of Modern Optical System, University of Shanghai for Science and Technology, Shanghai, 200093, PR China
| | - Wei Zhang
- School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, PR China; Engineering Research Center of Optical Instrument and System, Ministry of Education, Shanghai Key Lab of Modern Optical System, University of Shanghai for Science and Technology, Shanghai, 200093, PR China
| | - Yongxiong Wang
- School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai, 200093, PR China; Engineering Research Center of Optical Instrument and System, Ministry of Education, Shanghai Key Lab of Modern Optical System, University of Shanghai for Science and Technology, Shanghai, 200093, PR China
| | - Jianhua Zhang
- OsloMet Artificial Intelligence Lab, Department of Computer Science, Oslo Metropolitan University, Oslo, N-0130, Norway
| |
Collapse
|
26
|
González J, Ortega J, Damas M, Martín-Smith P, Gan JQ. A new multi-objective wrapper method for feature selection – Accuracy and stability analysis for BCI. Neurocomputing 2019. [DOI: 10.1016/j.neucom.2019.01.017] [Citation(s) in RCA: 33] [Impact Index Per Article: 6.6] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/27/2022]
|
27
|
Xu X, Liang T, Zhu J, Zheng D, Sun T. Review of classical dimensionality reduction and sample selection methods for large-scale data processing. Neurocomputing 2019. [DOI: 10.1016/j.neucom.2018.02.100] [Citation(s) in RCA: 31] [Impact Index Per Article: 6.2] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/21/2023]
|
28
|
Luo J, Feng Z, Lu N. Spatio-temporal discrepancy feature for classification of motor imageries. Biomed Signal Process Control 2019. [DOI: 10.1016/j.bspc.2018.07.003] [Citation(s) in RCA: 11] [Impact Index Per Article: 2.2] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 10/28/2022]
|
29
|
Mahmoudi M, Shamsi M. Multi-class EEG classification of motor imagery signal by finding optimal time segments and features using SNR-based mutual information. AUSTRALASIAN PHYSICAL & ENGINEERING SCIENCES IN MEDICINE 2018; 41:957-972. [PMID: 30338495 DOI: 10.1007/s13246-018-0691-2] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Received: 04/11/2018] [Accepted: 10/01/2018] [Indexed: 10/28/2022]
Abstract
The electroencephalogram signals are used to distinguish different motor imagery tasks in brain-computer interfaces. In most studies, in order to classify the EEG signals recorded in a cue-guided BCI paradigm, time segments for feature extraction after the onset of the visual cue were selected manually. In addition, in these studies the authors have selected a single identical time segment for different subjects. The present study emphasized on the inter-individual variability and difference between different motor imagery tasks as the potential source of erroneous results and used mutual information and the subject specific time interval to overcome this problem. More specifically, a new method was proposed to automatically find the best subject specific time intervals for the classification of four-class motor imagery tasks by using MI between the BCI input and output. Moreover, the signal-to-noise ratio was used to calculate the MI values, while the MI values were used as feature selection criteria to select the discriminative features. The time segments and the best discriminative features were found by using training data and used to assess the evaluation data. Furthermore, the CSP algorithm was used to extract signal features. The dataset 2A of BCI competition IV used in this study consisted of four different motor imagery signals, which were obtained from nine different subjects. One Vs One decomposition scheme was used to deal with the multi-class nature of the problem. The MI values showed that the obtained time segments not only varied between different subjects but also varied between different classifiers of different pair of classes. Finally, the results suggested that the proposed method was efficient in classifying multi-class motor imagery signals as compared to other classification strategies proposed by the other studies.
Collapse
Affiliation(s)
- Mahmoud Mahmoudi
- Faculty of Biomedical Engineering, Sahand University of Technology, Sahand New Town, Tabriz, Iran.
| | - Mousa Shamsi
- Faculty of Biomedical Engineering, Sahand University of Technology, Sahand New Town, Tabriz, Iran
| |
Collapse
|
30
|
Appavu alias Balamurugan S, Nancy SG. An Efficient Feature Selection and Classification Using Optimal Radial Basis Function Neural Network. INT J UNCERTAIN FUZZ 2018. [DOI: 10.1142/s0218488518500320] [Citation(s) in RCA: 3] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/18/2022]
Abstract
Feature selection is the process of identifying and removing many irrelevant and redundant features. Irrelevant features, along with redundant features, severely affect the accuracy of the learning machines. In high dimensional space finding clusters of data objects is challenging due to the curse of dimensionality. When the dimensionality increases, data in the irrelevant dimensions may produce much noise. And also, time complexity is the major issues in existing approach. In order to rectify these issues our proposed method made use of efficient feature subset selection in high dimensional data. Here we are considering the input dataset is the high dimensional micro array dataset. Initially, we have to select the optimal features so that our proposed technique employed Modified Social Spider Optimization (MSSO) algorithm. Here the traditional Social Spider Optimization is modified with the help of fruit fly optimization algorithm. Next the selected features are the input for the classifier. Here the classification is performed using Optimized Radial basis Function based neural network (ORBFNN) technique to classify the micro array data as normal or abnormal data. The effectiveness of RBFNN is optimized by means of artificial bee colony algorithm (ABC). Experimental results indicate that the proposed classification framework have outperformed by having better accuracy for five benchmark dataset 93.66%, 97.09%, 98.66%, 98.28% and 98.93% which is minimum value when compared to the existing technique. The proposed method is executed in MATLAB platform.
Collapse
Affiliation(s)
| | - S. Gilbert Nancy
- Department of Computer Science and Engineering, Bannari Amman Institute of Technology, Sathyamangalam, Erode District, Tamil Nadu, India
| |
Collapse
|
31
|
Wang J, Feng Z, Lu N, Luo J. Toward optimal feature and time segment selection by divergence method for EEG signals classification. Comput Biol Med 2018; 97:161-170. [PMID: 29747059 DOI: 10.1016/j.compbiomed.2018.04.022] [Citation(s) in RCA: 23] [Impact Index Per Article: 3.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/04/2018] [Revised: 04/23/2018] [Accepted: 04/23/2018] [Indexed: 11/16/2022]
Abstract
Feature selection plays an important role in the field of EEG signals based motor imagery pattern classification. It is a process that aims to select an optimal feature subset from the original set. Two significant advantages involved are: lowering the computational burden so as to speed up the learning procedure and removing redundant and irrelevant features so as to improve the classification performance. Therefore, feature selection is widely employed in the classification of EEG signals in practical brain-computer interface systems. In this paper, we present a novel statistical model to select the optimal feature subset based on the Kullback-Leibler divergence measure, and automatically select the optimal subject-specific time segment. The proposed method comprises four successive stages: a broad frequency band filtering and common spatial pattern enhancement as preprocessing, features extraction by autoregressive model and log-variance, the Kullback-Leibler divergence based optimal feature and time segment selection and linear discriminate analysis classification. More importantly, this paper provides a potential framework for combining other feature extraction models and classification algorithms with the proposed method for EEG signals classification. Experiments on single-trial EEG signals from two public competition datasets not only demonstrate that the proposed method is effective in selecting discriminative features and time segment, but also show that the proposed method yields relatively better classification results in comparison with other competitive methods.
Collapse
Affiliation(s)
- Jie Wang
- State Key Laboratory for Manufacturing System Engineering, System Engineering Institute, Xi'an Jiaotong University, Xi'an, Shanxi, China.
| | - Zuren Feng
- State Key Laboratory for Manufacturing System Engineering, System Engineering Institute, Xi'an Jiaotong University, Xi'an, Shanxi, China
| | - Na Lu
- State Key Laboratory for Manufacturing System Engineering, System Engineering Institute, Xi'an Jiaotong University, Xi'an, Shanxi, China; Beijing Advanced Innovation Center for Intelligent Robots and Systems, Beijing, China.
| | - Jing Luo
- State Key Laboratory for Manufacturing System Engineering, System Engineering Institute, Xi'an Jiaotong University, Xi'an, Shanxi, China
| |
Collapse
|
32
|
Kumar S, Sharma A, Tsunoda T. An improved discriminative filter bank selection approach for motor imagery EEG signal classification using mutual information. BMC Bioinformatics 2017; 18:545. [PMID: 29297303 PMCID: PMC5751568 DOI: 10.1186/s12859-017-1964-6] [Citation(s) in RCA: 51] [Impact Index Per Article: 7.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 12/31/2022] Open
Abstract
BACKGROUND Common spatial pattern (CSP) has been an effective technique for feature extraction in electroencephalography (EEG) based brain computer interfaces (BCIs). However, motor imagery EEG signal feature extraction using CSP generally depends on the selection of the frequency bands to a great extent. METHODS In this study, we propose a mutual information based frequency band selection approach. The idea of the proposed method is to utilize the information from all the available channels for effectively selecting the most discriminative filter banks. CSP features are extracted from multiple overlapping sub-bands. An additional sub-band has been introduced that cover the wide frequency band (7-30 Hz) and two different types of features are extracted using CSP and common spatio-spectral pattern techniques, respectively. Mutual information is then computed from the extracted features of each of these bands and the top filter banks are selected for further processing. Linear discriminant analysis is applied to the features extracted from each of the filter banks. The scores are fused together, and classification is done using support vector machine. RESULTS The proposed method is evaluated using BCI Competition III dataset IVa, BCI Competition IV dataset I and BCI Competition IV dataset IIb, and it outperformed all other competing methods achieving the lowest misclassification rate and the highest kappa coefficient on all three datasets. CONCLUSIONS Introducing a wide sub-band and using mutual information for selecting the most discriminative sub-bands, the proposed method shows improvement in motor imagery EEG signal classification.
Collapse
Affiliation(s)
- Shiu Kumar
- Department of Electronics, Instrumentation and Control Engineering, School of Electrical & Electronics Engineering, Fiji National University, Suva, Fiji. .,School of Engineering and Physics, Faculty of Science, Technology and Environment, The University of the South Pacific, Suva, Fiji.
| | - Alok Sharma
- School of Engineering and Physics, Faculty of Science, Technology and Environment, The University of the South Pacific, Suva, Fiji.,Institute for Integrated and Intelligent Systems (IIIS), Griffith University, Brisbane, Australia.,RIKEN Center for Integrative Medical Sciences, Yokohama, 230-0045, Japan.,CREST, JST, Yokohama, 230-0045, Japan
| | - Tatsuhiko Tsunoda
- RIKEN Center for Integrative Medical Sciences, Yokohama, 230-0045, Japan.,CREST, JST, Yokohama, 230-0045, Japan.,Department of Medical Science Mathematics, Medical Research Institute, Tokyo Medical and Dental University (TMDU), Tokyo, 113-8510, Japan
| |
Collapse
|
33
|
Yang Y, Chevallier S, Wiart J, Bloch I. Subject-specific time-frequency selection for multi-class motor imagery-based BCIs using few Laplacian EEG channels. Biomed Signal Process Control 2017. [DOI: 10.1016/j.bspc.2017.06.016] [Citation(s) in RCA: 44] [Impact Index Per Article: 6.3] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/24/2022]
|
34
|
Li L, Xu G, Zhang F, Xie J, Li M. Relevant Feature Integration and Extraction for Single-Trial Motor Imagery Classification. Front Neurosci 2017; 11:371. [PMID: 28706472 PMCID: PMC5489604 DOI: 10.3389/fnins.2017.00371] [Citation(s) in RCA: 8] [Impact Index Per Article: 1.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/20/2017] [Accepted: 06/13/2017] [Indexed: 11/16/2022] Open
Abstract
Brain computer interfaces provide a novel channel for the communication between brain and output devices. The effectiveness of the brain computer interface is based on the classification accuracy of single trial brain signals. The common spatial pattern (CSP) algorithm is believed to be an effective algorithm for the classification of single trial brain signals. As the amplitude feature for spatial projection applied by this algorithm is based on a broad frequency bandpass filter (mainly 5–30 Hz) in which the frequency band is often selected by experience, the CSP is sensitive to noise and the influence of other irrelevant information in the selected broad frequency band. In this paper, to improve the CSP, a novel relevant feature integration and extraction algorithm is proposed. Before projecting, we integrated the motor relevant information to suppress the interference of noise and irrelevant information, as well as to improve the spatial difference for projection. The algorithm was evaluated with public datasets. It showed significantly better classification performance with single trial electroencephalography (EEG) data, increasing by 6.8% compared with the CSP.
Collapse
Affiliation(s)
- Lili Li
- School of Mechanical Engineering, Xi'an Jiaotong UniversityXi'an, China
| | - Guanghua Xu
- School of Mechanical Engineering, Xi'an Jiaotong UniversityXi'an, China.,State Key Laboratory for Manufacturing Systems Engineering, Xi'an Jiaotong UniversityXi'an, China
| | - Feng Zhang
- School of Mechanical Engineering, Xi'an Jiaotong UniversityXi'an, China
| | - Jun Xie
- School of Mechanical Engineering, Xi'an Jiaotong UniversityXi'an, China
| | - Min Li
- School of Mechanical Engineering, Xi'an Jiaotong UniversityXi'an, China
| |
Collapse
|
35
|
A Fast, Efficient Domain Adaptation Technique for Cross-Domain Electroencephalography(EEG)-Based Emotion Recognition. SENSORS 2017; 17:s17051014. [PMID: 28467371 PMCID: PMC5469537 DOI: 10.3390/s17051014] [Citation(s) in RCA: 64] [Impact Index Per Article: 9.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Subscribe] [Scholar Register] [Received: 03/20/2017] [Revised: 04/23/2017] [Accepted: 04/26/2017] [Indexed: 11/19/2022]
Abstract
Electroencephalography (EEG)-based emotion recognition is an important element in psychiatric health diagnosis for patients. However, the underlying EEG sensor signals are always non-stationary if they are sampled from different experimental sessions or subjects. This results in the deterioration of the classification performance. Domain adaptation methods offer an effective way to reduce the discrepancy of marginal distribution. However, for EEG sensor signals, both marginal and conditional distributions may be mismatched. In addition, the existing domain adaptation strategies always require a high level of additional computation. To address this problem, a novel strategy named adaptive subspace feature matching (ASFM) is proposed in this paper in order to integrate both the marginal and conditional distributions within a unified framework (without any labeled samples from target subjects). Specifically, we develop a linear transformation function which matches the marginal distributions of the source and target subspaces without a regularization term. This significantly decreases the time complexity of our domain adaptation procedure. As a result, both marginal and conditional distribution discrepancies between the source domain and unlabeled target domain can be reduced, and logistic regression (LR) can be applied to the new source domain in order to train a classifier for use in the target domain, since the aligned source domain follows a distribution which is similar to that of the target domain. We compare our ASFM method with six typical approaches using a public EEG dataset with three affective states: positive, neutral, and negative. Both offline and online evaluations were performed. The subject-to-subject offline experimental results demonstrate that our component achieves a mean accuracy and standard deviation of 80.46% and 6.84%, respectively, as compared with a state-of-the-art method, the subspace alignment auto-encoder (SAAE), which achieves values of 77.88% and 7.33% on average, respectively. For the online analysis, the average classification accuracy and standard deviation of ASFM in the subject-to-subject evaluation for all the 15 subjects in a dataset was 75.11% and 7.65%, respectively, gaining a significant performance improvement compared to the best baseline LR which achieves 56.38% and 7.48%, respectively. The experimental results confirm the effectiveness of the proposed method relative to state-of-the-art methods. Moreover, computational efficiency of the proposed ASFM method is much better than standard domain adaptation; if the numbers of training samples and test samples are controlled within certain range, it is suitable for real-time classification. It can be concluded that ASFM is a useful and effective tool for decreasing domain discrepancy and reducing performance degradation across subjects and sessions in the field of EEG-based emotion recognition.
Collapse
|
36
|
Adaptive Feature Extraction of Motor Imagery EEG with Optimal Wavelet Packets and SE-Isomap. APPLIED SCIENCES-BASEL 2017. [DOI: 10.3390/app7040390] [Citation(s) in RCA: 24] [Impact Index Per Article: 3.4] [Reference Citation Analysis] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 11/17/2022]
|
37
|
Ksieniewicz P, Graña M, Woźniak M. Paired feature multilayer ensemble – concept and evaluation of a classifier. JOURNAL OF INTELLIGENT & FUZZY SYSTEMS 2017. [DOI: 10.3233/jifs-169139] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.1] [Reference Citation Analysis] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/15/2022]
Affiliation(s)
- Paweł Ksieniewicz
- Department of Systems and Computer Networks, Faculty of Electronics Wrocław University of Science and Technology, Wrocław, Poland
| | - Manuel Graña
- University of the Basque Country, Leioa, Bizkaia, Spain
| | - Michał Woźniak
- Department of Systems and Computer Networks, Faculty of Electronics Wrocław University of Science and Technology, Wrocław, Poland
| |
Collapse
|
38
|
Chai X, Wang Q, Zhao Y, Liu X, Bai O, Li Y. Unsupervised domain adaptation techniques based on auto-encoder for non-stationary EEG-based emotion recognition. Comput Biol Med 2016; 79:205-214. [PMID: 27810626 DOI: 10.1016/j.compbiomed.2016.10.019] [Citation(s) in RCA: 70] [Impact Index Per Article: 8.8] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/17/2016] [Revised: 10/17/2016] [Accepted: 10/19/2016] [Indexed: 11/30/2022]
Abstract
In electroencephalography (EEG)-based emotion recognition systems, the distribution between the training samples and the testing samples may be mismatched if they are sampled from different experimental sessions or subjects because of user fatigue, different electrode placements, varying impedances, etc. Therefore, it is difficult to directly classify the EEG patterns with a conventional classifier. The domain adaptation method, which is aimed at obtaining a common representation across training and test domains, is an effective method for reducing the distribution discrepancy. However, the existing domain adaptation strategies either employ a linear transformation or learn the nonlinearity mapping without a consistency constraint; they are not sufficiently powerful to obtain a similar distribution from highly non-stationary EEG signals. To address this problem, in this paper, a novel component, called the subspace alignment auto-encoder (SAAE), is proposed. Taking advantage of both nonlinear transformation and a consistency constraint, we combine an auto-encoder network and a subspace alignment solution in a unified framework. As a result, the source domain can be aligned with the target domain together with its class label, and any supervised method can be applied to the new source domain to train a classifier for classification in the target domain, as the aligned source domain follows a distribution similar to that of the target domain. We compared our SAAE method with six typical approaches using a public EEG dataset containing three affective states: positive, neutral, and negative. Subject-to-subject and session-to-session evaluations were performed. The subject-to-subject experimental results demonstrate that our component achieves a mean accuracy of 77.88% in comparison with a state-of-the-art method, TCA, which achieves 73.82% on average. In addition, the average classification accuracy of SAAE in the session-to-session evaluation for all the 15 subjects in a dataset is 81.81%, an improvement of up to 1.62% on average as compared to the best baseline TCA. The experimental results show the effectiveness of the proposed method relative to state-of-the-art methods. It can be concluded that SAAE is a useful and effective tool for decreasing domain discrepancy and reducing performance degradation across subjects and sessions in the EEG-based emotion recognition field.
Collapse
Affiliation(s)
- Xin Chai
- School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, China
| | - Qisong Wang
- School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, China.
| | - Yongping Zhao
- School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, China
| | - Xin Liu
- Department of Traffic Information and Control Engineering, School of Transportation Science and Engineering, Harbin Institute of Technology, Harbin, China
| | - Ou Bai
- Department of Electrical and Computer Engineering, Florida International University, USA
| | - Yongqiang Li
- School of Electrical Engineering and Automation, Harbin Institute of Technology, Harbin, China
| |
Collapse
|