A review of classification algorithms for EEG-based brain–computer interfaces

Lotte, Fabien; Congedo, Marco; Lécuyer, Anatole; Lamarche, Fabrice; Arnaldi, Bruno

doi:10.1088/1741-2560/4/2/r01

Cited by 2,291 publications

(1,568 citation statements)

References 76 publications

Supporting

Mentioning

1,488

Contrasting

Unclassified

Order By: Relevance

“…The trade-off parameter C was selected at 100 and kernel parameter σ was chosen 0.5 [16], [17]. 10-fold cross validation was applied also in order to find the optimized parameters where the results were close with the selected parameters [18].…”

Section: Svm Classifiermentioning

confidence: 99%

“…Hidden Markov Model (HMM) is a statistical model that has been used to different areas [17]- [20]. Lotte et al applied HMM on EEG based brain computer interfaces [17].…”

Section: E Hmmmentioning

confidence: 99%

“…Lotte et al applied HMM on EEG based brain computer interfaces [17]. Obermaier et al used mHMM for online classification of single trial EEG data and improved online experiment determination of minimal classification error [19].…”

Section: E Hmmmentioning

confidence: 99%

See 2 more Smart Citations

Pain Level Measurement Using Discrete Wavelet Transform

Vatankhah¹,

Toliyat²

2016

IJET

View full text Add to dashboard Cite

Abstract-Experimental pain has extensively been used as a tool for investigating neural mechanisms and the psychological factors involved in pain processing. The detection of existence and/or level of pain is vital when verbal information is not present e.g. for infants, disabled persons, anesthetized patients and animals also. This study shows that there is a firm relation between Electroencephalogram (EEG) and chronic pain levels and EEG can be used as a reliable tool for detecting, measuring and diagnosing pain levels in humans.This paper proposed a use of wavelet coherency in order to estimate the three pain levels and its usage as an index for pain measurement. Besides, wavelet coefficients are studied to show consistencies with EEG dynamic were extracted to provide the feature vector. A Hidden Markov Model (HMM) and a support vector machine (SVM) scheme was used for pain levels classification. This study confirms the hypothesis that brain pattern under the chronic pain mental task is mapped on EEG and the dependency of brain patterns to EEG is possible and detectable.Index Terms-Chronic pain index, electroencephalogram, SVM, HMM.

show abstract

Section: Svm Classifiermentioning

confidence: 99%

“…Hidden Markov Model (HMM) is a statistical model that has been used to different areas [17]- [20]. Lotte et al applied HMM on EEG based brain computer interfaces [17].…”

Section: E Hmmmentioning

confidence: 99%

See 1 more Smart Citation

Pain Level Measurement Using Discrete Wavelet Transform

Vatankhah¹,

Toliyat²

2016

IJET

View full text Add to dashboard Cite

show abstract

“…The use of a large number of time segments leads to high dimensionality of the feature vectors (which is [100 × 4097] for each class). Theoretically, if the number of training data is small compared to the size of the feature vectors, the classifier will most probably give poor results [20]. It is recommended to use at least five to ten times as many training samples per class as the dimensionality [21], [22].…”

Section: Feature Extractionmentioning

confidence: 99%

Analysis of Epileptic EEG Signals with Simple Random Sampling J48 Algorithm

Wang¹,

Zhu²,

Li³

et al. 2014

IJBBB

View full text Add to dashboard Cite

“…In order to assist in ambulation rehabilitation, the method needs to detect the feature signal related to the voluntary walking in the EEG. The techniques that target various EEG features exist for analyzing EEG signals [1], such as power spectrum, spectral centroid, event-related potential (ERP), and principal component analysis (PCA) as well as factor analysis, independent component analysis (ICA), k-nearest neighbor (kNN), linear discriminant analysis (LDA), neural network analysis (NN), and support vector machine (SVM) classification. We focused on a combination of feature extraction techniques for real-time processing because our final goal is to develop a system that can be used on a daily basis.…”

Section: Introductionmentioning

confidence: 99%

Electroencephalogram Analysis of Mechanisms Underlying Brain Activity during Voluntary Movement

Kuramoto¹,

Ito²,

Sato³

et al. 2014

IJBBB

View full text Add to dashboard Cite

Abstract-This paper proposes an electroencephalogram (EEG)-based method for identifying the mechanism underlying brain activity during a voluntary movement. Our final goal is to develop a rehabilitation assistance system that can be used on a daily basis by observing a patient's degree of functional recovery. In order to assist in ambulation rehabilitation, the mechanism related to the voluntary movement need to be identified. As a first step, we elucidated the mechanism underlying brain activity. The proposed method involves EEG recording, blind source separation, noise reduction, and feature enhancement. The EEG device has dry-type sensors. Independent component analysis is used to delink the signal. Median morphological filters are used to reduce EEG noise and emphasize the feature signal. To demonstrate the effectiveness of the proposed method, we conducted experiments using real EEG data.Index Terms-Electroencephalogram (EEG), voluntary movement, brain activity, independent component analysis, median morphological filters, decision making.

show abstract

A review of classification algorithms for EEG-based brain–computer interfaces

Cited by 2,291 publications

References 76 publications

Pain Level Measurement Using Discrete Wavelet Transform

Pain Level Measurement Using Discrete Wavelet Transform

Analysis of Epileptic EEG Signals with Simple Random Sampling J48 Algorithm

Electroencephalogram Analysis of Mechanisms Underlying Brain Activity during Voluntary Movement

Contact Info

Product

Resources

About