Pub Date : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737307
P. Douglas, D. Douglas
Electroencephalogram (EEG) has been a core tool used in functional neuroimaging in humans for nearly a hundred years. Because it is inexpensive, easy to implement, and noninvasive, it also represents an excellent candidate modality for use in the BCI setting. Nonetheless, a complete understanding of how EEG measurements (voltage fluctuations) relate to information processing in the brain remains somewhat elusive. A deeper understanding of the neuroanatomical underpinnings of the EEG signal may help explain inter-individual variability in evoked and induced potentials, which may improve BCI therapies targeted to the individual. According to classic biophysical models, EEG fluctuations are primarily a reflection of locally synchronized neuronal oscillations within the gray matter oriented approximately orthogonal to the scalp. In contrast, global models ignore local signals due to dendritic processing, and suggest that propagation delays due to white matter architecture are responsible for the EEG signal, and are capable of explaining the coherence between numerous rhythms (e.g., alpha) at spatially distinct areas of the scalp. Recently, combined local-global models suggest that the EEG signal may reflect a superposition of local processing along with global contributors including transduction along white matter tracts in the brain. Incorporating both local and global (e.g., white matter) priors into EEG source models may therefore improve source estimates. These models may also help disentangle which aspects of the EEG signal are predicted to colocalize spatially with measurements from functional MRI (fMRI). Here, we explore the possibility that white matter conductivity contributes to EEG measurements via a generative model based on classic axonal transduction models, and discuss its potential implications for source estimation.
{"title":"Reconsidering Spatial Priors In EEG Source Estimation : Does White Matter Contribute to EEG Rhythms?","authors":"P. Douglas, D. Douglas","doi":"10.1109/IWW-BCI.2019.8737307","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737307","url":null,"abstract":"Electroencephalogram (EEG) has been a core tool used in functional neuroimaging in humans for nearly a hundred years. Because it is inexpensive, easy to implement, and noninvasive, it also represents an excellent candidate modality for use in the BCI setting. Nonetheless, a complete understanding of how EEG measurements (voltage fluctuations) relate to information processing in the brain remains somewhat elusive. A deeper understanding of the neuroanatomical underpinnings of the EEG signal may help explain inter-individual variability in evoked and induced potentials, which may improve BCI therapies targeted to the individual. According to classic biophysical models, EEG fluctuations are primarily a reflection of locally synchronized neuronal oscillations within the gray matter oriented approximately orthogonal to the scalp. In contrast, global models ignore local signals due to dendritic processing, and suggest that propagation delays due to white matter architecture are responsible for the EEG signal, and are capable of explaining the coherence between numerous rhythms (e.g., alpha) at spatially distinct areas of the scalp. Recently, combined local-global models suggest that the EEG signal may reflect a superposition of local processing along with global contributors including transduction along white matter tracts in the brain. Incorporating both local and global (e.g., white matter) priors into EEG source models may therefore improve source estimates. These models may also help disentangle which aspects of the EEG signal are predicted to colocalize spatially with measurements from functional MRI (fMRI). Here, we explore the possibility that white matter conductivity contributes to EEG measurements via a generative model based on classic axonal transduction models, and discuss its potential implications for source estimation.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132539784","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737323
Dong-Kyun Han, Min-Ho Lee, J. Williamson, Seong-Whan Lee
In this study, we investigated the effect of real-time neurofeedback systems by adjusting the speed of a racing car and report the difference in effect between virtual and real environments. Thirty participants were divided into two conditions of the neurofeedback system (i.e., racing in real track and virtual game). For the performance evaluation, the band power of resting state EEG data and cognitive tests (Stroop and Digit span) were evaluated before and after the neurofeedback training. In the result, a significant increase of band power in the alpha frequency range (8–13Hz) as well as the test score were observed in both the virtual and real environments. Furthermore, neurofeedback in the virtual environment showed enhanced training effects compared to the real environment. We conclude that the performance of the neurofeedback training can be profoundly effected by the system environment as various factors (e.g., motivation, reward) are involved in the performance.
{"title":"The Effect of Neurofeedback Training in Virtual and Real Environments based on BCI","authors":"Dong-Kyun Han, Min-Ho Lee, J. Williamson, Seong-Whan Lee","doi":"10.1109/IWW-BCI.2019.8737323","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737323","url":null,"abstract":"In this study, we investigated the effect of real-time neurofeedback systems by adjusting the speed of a racing car and report the difference in effect between virtual and real environments. Thirty participants were divided into two conditions of the neurofeedback system (i.e., racing in real track and virtual game). For the performance evaluation, the band power of resting state EEG data and cognitive tests (Stroop and Digit span) were evaluated before and after the neurofeedback training. In the result, a significant increase of band power in the alpha frequency range (8–13Hz) as well as the test score were observed in both the virtual and real environments. Furthermore, neurofeedback in the virtual environment showed enhanced training effects compared to the real environment. We conclude that the performance of the neurofeedback training can be profoundly effected by the system environment as various factors (e.g., motivation, reward) are involved in the performance.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134465630","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 : 2019-02-01DOI: 10.1109/iww-bci.2019.8737250
{"title":"BCI 2019 Welcome Message from the General Chairs","authors":"","doi":"10.1109/iww-bci.2019.8737250","DOIUrl":"https://doi.org/10.1109/iww-bci.2019.8737250","url":null,"abstract":"","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116656483","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737256
Amardeep Singh, Sunil Lal, H. Guesgen
Motor imagery based brain computer interface requires large number of labeled subject specific training trials to calibrate system for new subjects. This is due to huge variations in individual characteristics. Major challenge in development of brain computer interface is to reduce calibration time or completely eliminate. Existing approaches rise up to this challenge by incorporating Euclidean representation of the individual variations from other subjects’ trials. They use covariance matrices from other subjects but do not consider the geometry of the covariance matrices, which lies in space of Symmetric Positive Definite (SPD) matrices. This inevitably limits their performance. We focus on reducing calibration time by introducing Riemannian approach by incorporating geometrical properties of covariance matrices in the subject to subject transfer. Our method outperforms the state of the art methods on the BCI competition dataset IVa. Our proposed method yielded accuracy of 77.67%, 100%, 75%, 87.05% and 91.67% for five subjects (aa, al, av, aw and ay respectively) in the dataset resulting in an average accuracy of 86.27%.
{"title":"Motor Imagery Classification Based on Subject to Subject Transfer in Riemannian Manifold","authors":"Amardeep Singh, Sunil Lal, H. Guesgen","doi":"10.1109/IWW-BCI.2019.8737256","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737256","url":null,"abstract":"Motor imagery based brain computer interface requires large number of labeled subject specific training trials to calibrate system for new subjects. This is due to huge variations in individual characteristics. Major challenge in development of brain computer interface is to reduce calibration time or completely eliminate. Existing approaches rise up to this challenge by incorporating Euclidean representation of the individual variations from other subjects’ trials. They use covariance matrices from other subjects but do not consider the geometry of the covariance matrices, which lies in space of Symmetric Positive Definite (SPD) matrices. This inevitably limits their performance. We focus on reducing calibration time by introducing Riemannian approach by incorporating geometrical properties of covariance matrices in the subject to subject transfer. Our method outperforms the state of the art methods on the BCI competition dataset IVa. Our proposed method yielded accuracy of 77.67%, 100%, 75%, 87.05% and 91.67% for five subjects (aa, al, av, aw and ay respectively) in the dataset resulting in an average accuracy of 86.27%.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123837237","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737263
Soren Moller Christensen, Nicklas Stubkjær Holm, S. Puthusserypady
Worldwide, millions of people are locked in or in a wheelchair, due to several neuromuscular disorders or spinal cord injuries. These individuals are deprived of trivial social activities, like interacting or playing games with other people. Such activities are crucial for personal development, and can have a great impact on the quality of their lives. This work aims at the design and implementation of an electroencephalography (EEG) based motor imagery (MI) brain computer interface (BCI) system that would allow disabled, and able-bodied, individuals alike to control a drone in a 3D physical environment by only using their thoughts. An improved version of the filter bank common spatial pattern (FBCSP) algorithm was developed, and it has shown to perform superior (68.5% accuracy) to the winning FBCSP algorithm (67.8% accuracy), when tested on dataset 2a (4 class MI) of the BCI competition IV. A deep convolutional neural network (CNN) based algorithm was also implemented and tested on the same dataset, which however performed inferior (62.9% accuracy) to the winner, as well as our proposed FBCSP algorithms. The improved FBCSP was then tested on our in-house 5-class (left hand, right hand, tongue, both feet and rest) MI dataset (collected from 10 able-bodied subjects) and obtained a mean accuracy of 41.8±11.74%. This is considered a significant result though it is not good enough to attempt the control of a real drone.
{"title":"An Improved Five Class MI Based BCI Scheme for Drone Control Using Filter Bank CSP","authors":"Soren Moller Christensen, Nicklas Stubkjær Holm, S. Puthusserypady","doi":"10.1109/IWW-BCI.2019.8737263","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737263","url":null,"abstract":"Worldwide, millions of people are locked in or in a wheelchair, due to several neuromuscular disorders or spinal cord injuries. These individuals are deprived of trivial social activities, like interacting or playing games with other people. Such activities are crucial for personal development, and can have a great impact on the quality of their lives. This work aims at the design and implementation of an electroencephalography (EEG) based motor imagery (MI) brain computer interface (BCI) system that would allow disabled, and able-bodied, individuals alike to control a drone in a 3D physical environment by only using their thoughts. An improved version of the filter bank common spatial pattern (FBCSP) algorithm was developed, and it has shown to perform superior (68.5% accuracy) to the winning FBCSP algorithm (67.8% accuracy), when tested on dataset 2a (4 class MI) of the BCI competition IV. A deep convolutional neural network (CNN) based algorithm was also implemented and tested on the same dataset, which however performed inferior (62.9% accuracy) to the winner, as well as our proposed FBCSP algorithms. The improved FBCSP was then tested on our in-house 5-class (left hand, right hand, tongue, both feet and rest) MI dataset (collected from 10 able-bodied subjects) and obtained a mean accuracy of 41.8±11.74%. This is considered a significant result though it is not good enough to attempt the control of a real drone.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128549729","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737319
Jaehyung Lee, Kabmun Cha, Hyungmin Kim, Junhyuk Choi, Choong Hyun Kim, S. Lee
The goal of this study was to compare decoding accuracy of left and right movement intention from electroencephalography (EEG) using three different types of paradigms: Motor Imagery (MI), Selective Attention (SA), and Hybrid task (HY)). Specifically, SA and HY are the Steady-State Somatosensory Evoked potential (SSSEP) paradigms which elicit brain responses to tactile stimulation. One subject participated in two sessions (Screening and Study session). In the screening session, resonance-like frequency of the subject was found at each hand while sitting on a chair. In the study session, the subject was asked to imagine either left of right hand open-close movement (MI task), to give selective attention to the vibrotactile stimulation (SA task), and to perform combined MI and SA task (HY) according to a randomly assigned directional cue. The accuracies of 3 paradigms were MI-left 65.8%, MI-right 69.2% (mean: 67.5%), SA-left 76.6%, SA-right 84.0% (mean: 80.3%) and HY-left 93.8%, HY-right 95.9% (mean: 94.9%). The method and results of the current study could be a basis for controlling the left and right movement direction of an exoskeleton robot using EEG.
{"title":"Hybrid MI-SSSEP Paradigm for classifying left and right movement toward BCI for exoskeleton control","authors":"Jaehyung Lee, Kabmun Cha, Hyungmin Kim, Junhyuk Choi, Choong Hyun Kim, S. Lee","doi":"10.1109/IWW-BCI.2019.8737319","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737319","url":null,"abstract":"The goal of this study was to compare decoding accuracy of left and right movement intention from electroencephalography (EEG) using three different types of paradigms: Motor Imagery (MI), Selective Attention (SA), and Hybrid task (HY)). Specifically, SA and HY are the Steady-State Somatosensory Evoked potential (SSSEP) paradigms which elicit brain responses to tactile stimulation. One subject participated in two sessions (Screening and Study session). In the screening session, resonance-like frequency of the subject was found at each hand while sitting on a chair. In the study session, the subject was asked to imagine either left of right hand open-close movement (MI task), to give selective attention to the vibrotactile stimulation (SA task), and to perform combined MI and SA task (HY) according to a randomly assigned directional cue. The accuracies of 3 paradigms were MI-left 65.8%, MI-right 69.2% (mean: 67.5%), SA-left 76.6%, SA-right 84.0% (mean: 80.3%) and HY-left 93.8%, HY-right 95.9% (mean: 94.9%). The method and results of the current study could be a basis for controlling the left and right movement direction of an exoskeleton robot using EEG.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124339867","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737305
Jisung Park, Sung-Phil Kim
The current neural decoding algorithms for brain-machine interfaces (BMIs) have largely focused on predicting the velocity of arm movements from neuronal ensemble activity. Yet, mounting evidence indicates that velocity is encoded separately in motor cortical activity. In this regard, we aimed to decode separate speed and direction information independently using a machine learning algorithm based on long short-term memory (LSTM). The performance of the proposed decoder was compared with the traditional decodres using velocity Kalman filter and the velocity LSTM. The proposed decoder showed better angular prediction than the other decoders. Also, the reconstruction hand trajectories with the proposed decoder acquired the targets more often. Movement time of the reconstructed trajectories by the proposed decoder was shorter than the others. Our results suggest advantages of decoding speed and direction independently using a nonlinear model such as LSTM for intracortical BMIs.
{"title":"Estimation of speed and direction of arm movements from M1 activity using a nonlinear neural decoder","authors":"Jisung Park, Sung-Phil Kim","doi":"10.1109/IWW-BCI.2019.8737305","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737305","url":null,"abstract":"The current neural decoding algorithms for brain-machine interfaces (BMIs) have largely focused on predicting the velocity of arm movements from neuronal ensemble activity. Yet, mounting evidence indicates that velocity is encoded separately in motor cortical activity. In this regard, we aimed to decode separate speed and direction information independently using a machine learning algorithm based on long short-term memory (LSTM). The performance of the proposed decoder was compared with the traditional decodres using velocity Kalman filter and the velocity LSTM. The proposed decoder showed better angular prediction than the other decoders. Also, the reconstruction hand trajectories with the proposed decoder acquired the targets more often. Movement time of the reconstructed trajectories by the proposed decoder was shorter than the others. Our results suggest advantages of decoding speed and direction independently using a nonlinear model such as LSTM for intracortical BMIs.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115312720","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737347
Gaeun Jeong, J. Kim, Seokyun Ryun, C. Chung
Perceiving and processing sensory stimuli are essential to generate motor action. Previous studies suggested features of vibrotactile stimulus such as amplitude and frequency are differently represented onto somatosensory cortices so that the stimulus can be discriminated. In the present study, we aimed to demonstrate the effect of transcranial magnetic stimulation (TMS) triplet pulses over primary somatosensory cortex (S1) or secondary somatosensory cortex (S2) on a tactile discrimination task. In two alternative forced choice task, TMS over S1 or S2 significantly interfered with the discrimination performance. This disruptive influence was mostly shown when the vibrotactile stimulus was close to high frequency (320Hz). Therefore we concluded the inhibitory effect of TMS is selective with tactile frequency.
{"title":"Interference in tactile discrmination performance by neuronal modulation","authors":"Gaeun Jeong, J. Kim, Seokyun Ryun, C. Chung","doi":"10.1109/IWW-BCI.2019.8737347","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737347","url":null,"abstract":"Perceiving and processing sensory stimuli are essential to generate motor action. Previous studies suggested features of vibrotactile stimulus such as amplitude and frequency are differently represented onto somatosensory cortices so that the stimulus can be discriminated. In the present study, we aimed to demonstrate the effect of transcranial magnetic stimulation (TMS) triplet pulses over primary somatosensory cortex (S1) or secondary somatosensory cortex (S2) on a tactile discrimination task. In two alternative forced choice task, TMS over S1 or S2 significantly interfered with the discrimination performance. This disruptive influence was mostly shown when the vibrotactile stimulus was close to high frequency (320Hz). Therefore we concluded the inhibitory effect of TMS is selective with tactile frequency.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124369943","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737260
Sujin Bak, Yeon Pyo, Jichai Jeong
Brain-computer interface can be currently accelerated to develop the exoskeleton for healthy people as well as patients who are unable to move muscles around the world. In this situation, the communication between the electroencephalogram (EEG) and prosthesis discovers the vulnerabilities to taking personal information. However, previous researches only focus on the analysis of attack pattern rather than fixing the vulnerability. In order to complement the vulnerability, we propose a blockchain platform in which try to identify the modulated data when server is attacked. Also, we find out potential risks in EEG data with non-blockchain environments after attack in our study. As a result, the proposed system can guarantee the integrity of EEG data by knowing the change of hash, and can prevent attacks such as hijacking, sniffing, and eavesdropping.
{"title":"Protection of EEG Data using Blockchain Platform","authors":"Sujin Bak, Yeon Pyo, Jichai Jeong","doi":"10.1109/IWW-BCI.2019.8737260","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737260","url":null,"abstract":"Brain-computer interface can be currently accelerated to develop the exoskeleton for healthy people as well as patients who are unable to move muscles around the world. In this situation, the communication between the electroencephalogram (EEG) and prosthesis discovers the vulnerabilities to taking personal information. However, previous researches only focus on the analysis of attack pattern rather than fixing the vulnerability. In order to complement the vulnerability, we propose a blockchain platform in which try to identify the modulated data when server is attacked. Also, we find out potential risks in EEG data with non-blockchain environments after attack in our study. As a result, the proposed system can guarantee the integrity of EEG data by knowing the change of hash, and can prevent attacks such as hijacking, sniffing, and eavesdropping.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"36 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123722438","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 : 2019-02-01DOI: 10.1109/IWW-BCI.2019.8737340
Eunjin Jeon, Wonjun Ko, Heung-Il Suk
Recent successes of deep learning methods in various applications have inspired BCI researchers for their use in EEG classification. However, data insufficiency and high intra- and inter-subject variabilities hinder from taking their advantage of discovering complex patterns inherent in data, which can be potentially useful to enhance EEG classification accuracy. In this paper, we devise a novel framework of training a deep network by adapting samples of other subjects as a means of domain adaptation. Assuming that there are EEG trials of motor-imagery tasks from multiple subjects available, we first select a subject whose EEG signal characteristics are similar to the target subject based on their power spectral density in resting-state EEG signals. We then use EEG signals of both the selected subject (called a source subject) and the target subject jointly in training a deep network. Rather than training a single path network, we adopt a multi-path network architecture, where the shared bottom layers are used to discover common features for both source and target subjects, while the upper layers branch out into (1) source-target subject identification, (2) label prediction optimized for a source subject, and (3) label prediction optimized for a target subject. Based on our experimental results over the BCI Competition IV-IIa dataset, we validated the effectiveness of the proposed framework in various aspects.
{"title":"Domain Adaptation with Source Selection for Motor-Imagery based BCI","authors":"Eunjin Jeon, Wonjun Ko, Heung-Il Suk","doi":"10.1109/IWW-BCI.2019.8737340","DOIUrl":"https://doi.org/10.1109/IWW-BCI.2019.8737340","url":null,"abstract":"Recent successes of deep learning methods in various applications have inspired BCI researchers for their use in EEG classification. However, data insufficiency and high intra- and inter-subject variabilities hinder from taking their advantage of discovering complex patterns inherent in data, which can be potentially useful to enhance EEG classification accuracy. In this paper, we devise a novel framework of training a deep network by adapting samples of other subjects as a means of domain adaptation. Assuming that there are EEG trials of motor-imagery tasks from multiple subjects available, we first select a subject whose EEG signal characteristics are similar to the target subject based on their power spectral density in resting-state EEG signals. We then use EEG signals of both the selected subject (called a source subject) and the target subject jointly in training a deep network. Rather than training a single path network, we adopt a multi-path network architecture, where the shared bottom layers are used to discover common features for both source and target subjects, while the upper layers branch out into (1) source-target subject identification, (2) label prediction optimized for a source subject, and (3) label prediction optimized for a target subject. Based on our experimental results over the BCI Competition IV-IIa dataset, we validated the effectiveness of the proposed framework in various aspects.","PeriodicalId":345970,"journal":{"name":"2019 7th International Winter Conference on Brain-Computer Interface (BCI)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122410188","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：