Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311525
Soo-In Choi, G. Choi, Hyung-Tak Lee, Han-Jeong Hwang, Jaeyoung Shin
Electroencephalography (EEG) has been mainly utilized for developing brain-computer interface (BCI) systems. In recent, use of Ear-EEG measured around the ears has been proposed to enhance the practicality of conventional EEG-based BCI systems. Most of BCI systems based on Ear-EEG have used exogenous BCI paradigms employing external stimuli. In this study, we investigated the feasibility of using Ear-EEG in developing an endogenous BCI system that uses self-modulated brain signals. EEG data was measured while subjects performed mental arithmetic (MA) and baseline (BL) task. EEG data analysis was performed after dividing the whole brain area into four regions of interest (frontal, central, occipital, and ear area) to compare their EEG characteristics and classification performance. Similar event-related (de)synchronization (ERD/ERS) patterns were observed between the four ROIs, and classification performance was insignificant between them, except occipital area (frontal: 72.6 %, central: 76.7 %, occipital: 82.6 % and ear: 75.6 %). From the results, we could confirm the possibility of using Ear-EEG for developing an endogenous BCI system.
{"title":"Classification of mental arithmetic and resting-state based on Ear-EEG","authors":"Soo-In Choi, G. Choi, Hyung-Tak Lee, Han-Jeong Hwang, Jaeyoung Shin","doi":"10.1109/IWW-BCI.2018.8311525","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311525","url":null,"abstract":"Electroencephalography (EEG) has been mainly utilized for developing brain-computer interface (BCI) systems. In recent, use of Ear-EEG measured around the ears has been proposed to enhance the practicality of conventional EEG-based BCI systems. Most of BCI systems based on Ear-EEG have used exogenous BCI paradigms employing external stimuli. In this study, we investigated the feasibility of using Ear-EEG in developing an endogenous BCI system that uses self-modulated brain signals. EEG data was measured while subjects performed mental arithmetic (MA) and baseline (BL) task. EEG data analysis was performed after dividing the whole brain area into four regions of interest (frontal, central, occipital, and ear area) to compare their EEG characteristics and classification performance. Similar event-related (de)synchronization (ERD/ERS) patterns were observed between the four ROIs, and classification performance was insignificant between them, except occipital area (frontal: 72.6 %, central: 76.7 %, occipital: 82.6 % and ear: 75.6 %). From the results, we could confirm the possibility of using Ear-EEG for developing an endogenous BCI system.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"171 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75709496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311488
Minryung R. Song, Sang Wan Lee
Classifying neural signals is a crucial step in the brain-computer interface (BCI). Although Deep Neural Network (DNN) has been shown to be surprisingly good at classification, DNN suffers from long training time and catastrophic forgetting. Catastrophic forgetting refers to a phenomenon in which a DNN tends to forget previously learned task when it learns a new task. Here we argue that the solution to this problem may be found in the human brain, specifically, by combining functions of the two regions: the striatum and the hippocampus, which is pivotal for reinforcement learning and memory recall relevant to the current context, respectively. The mechanism of these brain regions provides insights into resolving catastrophic forgetting and long training time of DNNs. Referring to the hippocampus-striatum network we discuss design principles of combining different types of DNNs for building a new BCI architecture, called “Meta BCI”.
{"title":"Meta BCI : Hippocampus-striatum network inspired architecture towards flexible BCI","authors":"Minryung R. Song, Sang Wan Lee","doi":"10.1109/IWW-BCI.2018.8311488","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311488","url":null,"abstract":"Classifying neural signals is a crucial step in the brain-computer interface (BCI). Although Deep Neural Network (DNN) has been shown to be surprisingly good at classification, DNN suffers from long training time and catastrophic forgetting. Catastrophic forgetting refers to a phenomenon in which a DNN tends to forget previously learned task when it learns a new task. Here we argue that the solution to this problem may be found in the human brain, specifically, by combining functions of the two regions: the striatum and the hippocampus, which is pivotal for reinforcement learning and memory recall relevant to the current context, respectively. The mechanism of these brain regions provides insights into resolving catastrophic forgetting and long training time of DNNs. Referring to the hippocampus-striatum network we discuss design principles of combining different types of DNNs for building a new BCI architecture, called “Meta BCI”.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"103 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"78256913","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311512
Dongjae Kim, Sang Wan Lee
Arbitration between model-based (MB) and model-free (MF) reinforcement learning (RL) is key feature of human reinforcement learning. The computational model of arbitration control has been demonstrated to outperform conventional reinforcement learning algorithm, in terms of not only behavioral data but also neural signals. However, this arbitration process does not take full account of contextual changes in environment during learning. By incorporating a Dirichlet process Gaussian mixture model into the arbitration process, we propose a meta-controller for RL that quickly adapts to contextual changes of environment. The proposed model performs better than a conventional model-free RL, model-based RL, and arbitration model.
{"title":"Context-dependent meta-control for reinforcement learning using a Dirichlet process Gaussian mixture model","authors":"Dongjae Kim, Sang Wan Lee","doi":"10.1109/IWW-BCI.2018.8311512","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311512","url":null,"abstract":"Arbitration between model-based (MB) and model-free (MF) reinforcement learning (RL) is key feature of human reinforcement learning. The computational model of arbitration control has been demonstrated to outperform conventional reinforcement learning algorithm, in terms of not only behavioral data but also neural signals. However, this arbitration process does not take full account of contextual changes in environment during learning. By incorporating a Dirichlet process Gaussian mixture model into the arbitration process, we propose a meta-controller for RL that quickly adapts to contextual changes of environment. The proposed model performs better than a conventional model-free RL, model-based RL, and arbitration model.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"27 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81631166","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311509
Changhee Han, C. Im
In this study, we developed a practical EEG-based BCI paradigm for online binary communication of patients in completely locked-in state (CLIS). The performance of our BCI paradigm was evaluated with a female patient in CLIS, who had never communicated even with her family for more than a year. An average online classification accuracy of 87.5 % was achieved using EEG data recorded just for 5 seconds. This is the first report of successful application of EEG-based BCI to the online yes/no communication of patients in CLIS.
{"title":"EEG-based brain-computer interface for real-time communication of patients in completely locked-in state","authors":"Changhee Han, C. Im","doi":"10.1109/IWW-BCI.2018.8311509","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311509","url":null,"abstract":"In this study, we developed a practical EEG-based BCI paradigm for online binary communication of patients in completely locked-in state (CLIS). The performance of our BCI paradigm was evaluated with a female patient in CLIS, who had never communicated even with her family for more than a year. An average online classification accuracy of 87.5 % was achieved using EEG data recorded just for 5 seconds. This is the first report of successful application of EEG-based BCI to the online yes/no communication of patients in CLIS.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"38 1","pages":"1-2"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"85308350","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311504
Min-Ki Kim, Sung-Phil Kim
Most invasive brain-machine interfaces (BMIs) have relied on the movement-related information in the firing activities of a number of cortical neurons. Recently, many efforts have been made to represent high-dimensional firing activities of a neuronal ensemble in a low-dimensional features space, visualizing the trajectory of temporal evolution of neural activities. The resulting neural trajectory often provides a sound means to visualize encoding of movement information in neuronal ensembles as well as to improve decoding performance by eliminating noise from irrelevant neurons. The present study aims to build the neural trajectory from motor cortical neurons in a primate performing a center-out task. The neural trajectory built by the standard principal component analysis method well represented hand speed profiles and provided proper feature vectors to a subsequent decoding algorithm. The results suggest an effective way of single-trial speed decoding for invasive BMIs.
{"title":"Decoding movement information from cortical activity for invasive BMIs","authors":"Min-Ki Kim, Sung-Phil Kim","doi":"10.1109/IWW-BCI.2018.8311504","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311504","url":null,"abstract":"Most invasive brain-machine interfaces (BMIs) have relied on the movement-related information in the firing activities of a number of cortical neurons. Recently, many efforts have been made to represent high-dimensional firing activities of a neuronal ensemble in a low-dimensional features space, visualizing the trajectory of temporal evolution of neural activities. The resulting neural trajectory often provides a sound means to visualize encoding of movement information in neuronal ensembles as well as to improve decoding performance by eliminating noise from irrelevant neurons. The present study aims to build the neural trajectory from motor cortical neurons in a primate performing a center-out task. The neural trajectory built by the standard principal component analysis method well represented hand speed profiles and provided proper feature vectors to a subsequent decoding algorithm. The results suggest an effective way of single-trial speed decoding for invasive BMIs.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"156 ","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91554975","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311516
K. Shiba, T. Kaburagi, Y. Kurihara
We aim to characterize functional connectivity during a listening comprehension task in terms of fit to common network topology models. The functional connectivity is expressed as a network structure which is reconstructed from cerebral blood volume measurements. The cerebral blood volume in the frontal lobe is measured using functional near-infrared spectroscopy (NIRS). Based on the reconstructed functional network structure, we discuss whether the functional connectivity has a scale-free or random graph structure. The feasibility of the reconstructed network is evaluated based on the distribution of the number of edges at nodes. In order to validate our proposed model, two language listening comprehension tasks were presented to subjects and the feasibility of the model structure is discussed. The experimental results suggest that the reconstructed functional connectivity network is more likely to be a scale-free network with an “ultra-small” world than a random network.
{"title":"A dynamic Bayesian network analysis of functional connectivity during a language listening comprehension task","authors":"K. Shiba, T. Kaburagi, Y. Kurihara","doi":"10.1109/IWW-BCI.2018.8311516","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311516","url":null,"abstract":"We aim to characterize functional connectivity during a listening comprehension task in terms of fit to common network topology models. The functional connectivity is expressed as a network structure which is reconstructed from cerebral blood volume measurements. The cerebral blood volume in the frontal lobe is measured using functional near-infrared spectroscopy (NIRS). Based on the reconstructed functional network structure, we discuss whether the functional connectivity has a scale-free or random graph structure. The feasibility of the reconstructed network is evaluated based on the distribution of the number of edges at nodes. In order to validate our proposed model, two language listening comprehension tasks were presented to subjects and the feasibility of the model structure is discussed. The experimental results suggest that the reconstructed functional connectivity network is more likely to be a scale-free network with an “ultra-small” world than a random network.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"1 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"88294433","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311510
Jihyeon Ha, Da-hye Kim, Laehyun Kim
The electroencephalography (EEG) based brain-computer interface (BCI) presented a new paradigm of rehabilitation. Especially, rehabilitation incorporating EEG based BCI for stroke with motor impairment makes rehabilitation more effective than previously; for example, it provides neurofeedback to improve engagement of the brain. In this study, we measured EEG data of nine patients with chronic stroke accompanied with a unilateral motor problem while all patients performed upper limb rehabilitation (performing a grasping task with the affected hand). As a result, we found that the EEG feature showed similar EEG power spectral densities between the ipsilesional area and contralesional area. Additionally, this feature was significantly correlated (Spearman correlation coefficient p = −0.7280, p < 0.05) with the Fugl-Meyer Assessment score of the affected hand, indicating a degree of motor function. These results showed that brain activity of patients who had low motor function bilaterally appeared in ipsilesional and contralesional areas, whereas brain activity of patients who had high motor function specifically appeared in the ipsilesional area only.
基于脑电图(EEG)的脑机接口(BCI)为康复治疗提供了一种新的范式。特别是，结合脑电脑机接口的康复治疗卒中合并运动障碍，使康复比以前更有效;例如，它提供神经反馈来提高大脑的参与度。在这项研究中，我们测量了9例伴有单侧运动问题的慢性中风患者的脑电图数据，同时所有患者都进行了上肢康复(用患手执行抓握任务)。结果表明，同侧和对侧的脑电特征具有相似的功率谱密度。此外，该特征与患手的Fugl-Meyer评估评分显著相关(Spearman相关系数p = - 0.7280, p < 0.05)，表明运动功能的程度。这些结果表明，双侧低运动功能患者的脑活动出现在同侧和对侧区域，而高运动功能患者的脑活动只出现在同侧区域。
{"title":"An approach for assessing stroke motor function ability using the similarity between electroencephalographic power spectral densities on both motor cortices","authors":"Jihyeon Ha, Da-hye Kim, Laehyun Kim","doi":"10.1109/IWW-BCI.2018.8311510","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311510","url":null,"abstract":"The electroencephalography (EEG) based brain-computer interface (BCI) presented a new paradigm of rehabilitation. Especially, rehabilitation incorporating EEG based BCI for stroke with motor impairment makes rehabilitation more effective than previously; for example, it provides neurofeedback to improve engagement of the brain. In this study, we measured EEG data of nine patients with chronic stroke accompanied with a unilateral motor problem while all patients performed upper limb rehabilitation (performing a grasping task with the affected hand). As a result, we found that the EEG feature showed similar EEG power spectral densities between the ipsilesional area and contralesional area. Additionally, this feature was significantly correlated (Spearman correlation coefficient p = −0.7280, p < 0.05) with the Fugl-Meyer Assessment score of the affected hand, indicating a degree of motor function. These results showed that brain activity of patients who had low motor function bilaterally appeared in ipsilesional and contralesional areas, whereas brain activity of patients who had high motor function specifically appeared in the ipsilesional area only.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"98 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75441817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311490
L. Botrel, A. Kübler
Reliable predictors of BCI performance would be desirable for basic research and application of BCI in a clinical context alike. In basic research, predictors help to elucidate how the brain instantiates BCI control. With respect to BCI controlled applications to be used by patient end-users with disease, predictors could support the choice of the optimal brain signal. Training of the predicting variable may support later BCI control. Among others, physiologic and psychologic variables have been suggested as such predictors. For example, the resting state μ-rhythm peak, the activation of dorsolateral prefrontal cortex during motor imagery, and the ability to coordinate visual and motor information were related to performance in different motor imagery BCI paradigms. The predictive power was low to medium, few even high, where the physiologic predictor was most powerful. To identify predictors, those and the related criterion variable have to be unambiguously defined. Likewise, reliability and validity have to be specified in the realm of BCI.
{"title":"Reliable predictors of SMR BCI performance — Do they exist?","authors":"L. Botrel, A. Kübler","doi":"10.1109/IWW-BCI.2018.8311490","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311490","url":null,"abstract":"Reliable predictors of BCI performance would be desirable for basic research and application of BCI in a clinical context alike. In basic research, predictors help to elucidate how the brain instantiates BCI control. With respect to BCI controlled applications to be used by patient end-users with disease, predictors could support the choice of the optimal brain signal. Training of the predicting variable may support later BCI control. Among others, physiologic and psychologic variables have been suggested as such predictors. For example, the resting state μ-rhythm peak, the activation of dorsolateral prefrontal cortex during motor imagery, and the ability to coordinate visual and motor information were related to performance in different motor imagery BCI paradigms. The predictive power was low to medium, few even high, where the physiologic predictor was most powerful. To identify predictors, those and the related criterion variable have to be unambiguously defined. Likewise, reliability and validity have to be specified in the realm of BCI.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"32 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81645557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we propose a novel architecture of a deep neural network for EEG-based motor imagery classification. Unlike the existing deep neural networks in the literature, the proposed network allows us to analyze the learned network weights from a neurophysiological perspective, thus providing an insight into the underlying patterns inherent in motor imagery induced EEG signals. In order to validate the effectiveness of the proposed method, we conducted experiments on the BCI Competition IV-IIa dataset by comparing with the competing methods in terms of the Cohen's k value. For qualitative analysis, we also performed visual inspection of the activation patterns estimated from the learned network weights.
{"title":"Deep recurrent spatio-temporal neural network for motor imagery based BCI","authors":"Wonjun Ko, Jee Seok Yoon, Eunsong Kang, E. Jun, Jun-Sik Choi, Heung-Il Suk","doi":"10.1109/IWW-BCI.2018.8311535","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311535","url":null,"abstract":"In this paper, we propose a novel architecture of a deep neural network for EEG-based motor imagery classification. Unlike the existing deep neural networks in the literature, the proposed network allows us to analyze the learned network weights from a neurophysiological perspective, thus providing an insight into the underlying patterns inherent in motor imagery induced EEG signals. In order to validate the effectiveness of the proposed method, we conducted experiments on the BCI Competition IV-IIa dataset by comparing with the competing methods in terms of the Cohen's k value. For qualitative analysis, we also performed visual inspection of the activation patterns estimated from the learned network weights.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"25 1","pages":"1-3"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81813895","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2018-01-01DOI: 10.1109/IWW-BCI.2018.8311494
Yi-Ming Jin, Mahta Mousavi, V. D. Sa
In brain-computer interfaces, adapting a classifier from one user to another is challenging but essential to reduce training time for new users. Common Spatial Patterns (CSP) is a widely used method for learning spatial filters for user specific feature extraction but the performance is degraded when applied to a different user. This paper proposes a novel Adaptive Selective Common Spatial Pattern (ASCSP) method to update the covariance matrix using selected candidates. Subspace alignment is then applied to the extracted features before classification. The proposed method outperforms the standard CSP and adaptive CSP algorithms previously proposed. Visualization of extracted features is provided to demonstrate how subspace alignment contributes to reduce the domain variance between source and target domains.
{"title":"Adaptive CSP with subspace alignment for subject-to-subject transfer in motor imagery brain-computer interfaces","authors":"Yi-Ming Jin, Mahta Mousavi, V. D. Sa","doi":"10.1109/IWW-BCI.2018.8311494","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2018.8311494","url":null,"abstract":"In brain-computer interfaces, adapting a classifier from one user to another is challenging but essential to reduce training time for new users. Common Spatial Patterns (CSP) is a widely used method for learning spatial filters for user specific feature extraction but the performance is degraded when applied to a different user. This paper proposes a novel Adaptive Selective Common Spatial Pattern (ASCSP) method to update the covariance matrix using selected candidates. Subspace alignment is then applied to the extracted features before classification. The proposed method outperforms the standard CSP and adaptive CSP algorithms previously proposed. Visualization of extracted features is provided to demonstrate how subspace alignment contributes to reduce the domain variance between source and target domains.","PeriodicalId":6537,"journal":{"name":"2018 6th International Conference on Brain-Computer Interface (BCI)","volume":"4 1","pages":"1-4"},"PeriodicalIF":0.0,"publicationDate":"2018-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74059660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：