Information Theoretic Approaches for Motor-Imagery BCI Systems: Review and Experimental Comparison

Martı́n-Clemente, Rubén; Olias, Javier; Thiyam, Deepa Beeta; Cichocki, Andrzej; Cruces, Sergio

doi:10.3390/e20010007

Cited by 29 publications

(23 citation statements)

References 66 publications

(121 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In EEG-based BCIs, control signals are obtained by different strategies due to the presence or absence of external stimulus to evoke user intention [13]. Self-modulation of sensory-motor rhythms through motor imagination (MI) of different parts of the body seems to be the most popular paradigm since it can be operated only by the user's will, independently of any external stimulation [14]. The imagination of right or left limbs movement, changing the neuronal activity in contralateral hemisphere of the brain motor cortex leads to physiological phenomenon known as event-related DE synchronization (ERD) and event-related synchronization (ERS) which are reflected respectively as decreasing and increasing power within either alpha and beta bands in EEG signal [15].…”

Section: Introductionmentioning

confidence: 99%

“…The imagination of right or left limbs movement, changing the neuronal activity in contralateral hemisphere of the brain motor cortex leads to physiological phenomenon known as event-related DE synchronization (ERD) and event-related synchronization (ERS) which are reflected respectively as decreasing and increasing power within either alpha and beta bands in EEG signal [15]. The distinctive pattern of these variations in the signal can be recognized by advanced signal processing and machine learning algorithms to reveal which kind of motion the user has imagined [14].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Identifying motor imagery activities in brain computer interfaces based on the intelligent selection of most informative timeframe

2020

View full text Add to dashboard Cite

Removing the contribution of dispensable mental activities dispersed across the electroencephalogram (EEG) signal improves the interpretability and efficiency of the intended neuronal responses to control a brain computer interface (BCI). Performing motor imagery tasks causes proper formation of special patterns at a specific timeframe of the EEG signal. The accurate selection of this optimal informative timeframe has not yet been investigated. Previous studies have considered an identic portion of data for all individuals, while neglecting that the duration and delay takes for the motor imagery brain activities to be well reflected in EEG signals vary between individuals. We propose an intelligent hybrid genetic algorithm-support vector machine (SVM) method to select the most stimulated timeframe of interest. The method also selects the most distinctive subset of features (through a comprehensive fused set of temporal, spectral and wavelet inspected information) while simultaneously optimize the SVM classifier's parameters. Evaluation results show that not only the most stimulated timeframe has a short duration but also occurs after a specific delay: that they are different between individuals. Using this optimal timeframe, the classification accuracy increased up to 92.14% for Graz 2003 and 89.00%, 84.81% and 85.00% for O3, S4 and X11 subjects of Graz 2005 database respectively. These results that were obtained despite the use of a small set of features confirm that this intelligent method can be effective in increasing the computational speed while decreasing the computational complexity which provides the potential capability of including in real time BCI systems.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Identifying motor imagery activities in brain computer interfaces based on the intelligent selection of most informative timeframe

2020

View full text Add to dashboard Cite

show abstract

“…The conventional method of MI BCI with EEG signals consists of the extraction of hidden features and subsequent classification based on various machine learning methods. The common spatial pattern (CSP) algorithm is one of the most popular feature extraction methods [10,11]. CSP is commonly used to analyze spatial patterns of multichannel MI EEG signals.…”

Section: Introductionmentioning

confidence: 99%

Application of Continuous Wavelet Transform and Convolutional Neural Network in Decoding Motor Imagery Brain-Computer Interface

Lee

Choi

2019

Entropy

View full text Add to dashboard Cite

The motor imagery-based brain-computer interface (BCI) using electroencephalography (EEG) has been receiving attention from neural engineering researchers and is being applied to various rehabilitation applications. However, the performance degradation caused by motor imagery EEG with very low single-to-noise ratio faces several application issues with the use of a BCI system. In this paper, we propose a novel motor imagery classification scheme based on the continuous wavelet transform and the convolutional neural network. Continuous wavelet transform with three mother wavelets is used to capture a highly informative EEG image by combining time-frequency and electrode location. A convolutional neural network is then designed to both classify motor imagery tasks and reduce computation complexity. The proposed method was validated using two public BCI datasets, BCI competition IV dataset 2b and BCI competition II dataset III. The proposed methods were found to achieve improved classification performance compared with the existing methods, thus showcasing the feasibility of motor imagery BCI.

show abstract

“…Typically, electroencephalogram (EEG) samples acquired under two different experimental conditions provide a multivariate data set with two classes. CSP linearly projects the data onto directions where the variance of the projected data points is significantly higher for one class than for the other [5]- [7]. The projected data variances can then be used as features for classification.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Common Spatial Patterns

Martı́n-Clemente

Olias

Cruces

et al. 2019

IEEE Trans. Neural Syst. Rehabil. Eng.

Self Cite

View full text Add to dashboard Cite

The common spatial pattern (CSP) method is a dimensionality reduction technique widely used in brain-computer interface (BCI) systems. In the two-class CSP problem, training data are linearly projected onto directions maximizing or minimizing the variance ratio between the two classes. The present contribution proves that kurtosis maximization performs CSP in an unsupervised manner, i.e., with no need for labeled data, when the classes follow Gaussian or elliptically symmetric distributions. Numerical analyses on synthetic and real data validate these findings in various experimental conditions, and demonstrate the interest of the proposed unsupervised approach.

show abstract

Information Theoretic Approaches for Motor-Imagery BCI Systems: Review and Experimental Comparison

Cited by 29 publications

References 66 publications

Identifying motor imagery activities in brain computer interfaces based on the intelligent selection of most informative timeframe

Identifying motor imagery activities in brain computer interfaces based on the intelligent selection of most informative timeframe

Application of Continuous Wavelet Transform and Convolutional Neural Network in Decoding Motor Imagery Brain-Computer Interface

Unsupervised Common Spatial Patterns

Contact Info

Product

Resources

About