Multiple kernel learning based on three discriminant features for a P300 speller BCI

Yoon, Kyungae; Kim, Kiseon

doi:10.1016/j.neucom.2016.09.053

Cited by 20 publications

(7 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Bostanov has shown that decomposing P300 waves via t-statistics based continuous-wavelet transforms extracts reliable spectral and temporal ERP features [39]. In another study, Yoon et al proposed three features that composed of the raw samples and the amplitude and negative area of P300, and achieved an improved performance using a multiple kernel classifier [35]. The design of spatial feature extractors has been widely studied as well.…”

mentioning

confidence: 99%

“…All these ERP based BCI studies generally follow a common methodology: 1) conduct an ERP experiment to acquire EEG data by placing an electrode cap on the scalp of the subject; 2) pre-process the collected data to remove the artifacts and enhance the signal-to-noise ratio-for instance, restricting frequency contents to a range of (0.1 Hz-30 Hz) as suggested by [30], [46]; 3) extract raw temporal features with a predefined window size of 600 ms or 1000 ms followed by a customized spatiotemporal feature extraction method such as wavelets, spatial filters, beamformers, etc. ; and 4) use extracted features in a classification rule [30], [31], [35], [47]- [50].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Learning Discriminative Spatiospectral Features of ERPs for Accurate Brain–Computer Interfaces

Abibullaev

Zollanvari

2019

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

Constructing accurate predictive models is at the heart of brain-computer interfaces (BCIs) because these models can ultimately translate brain activities into communication and control commands. The majority of the previous work in BCI use spatial, temporal, or spatiotemporal features of event-related potentials (ERPs). In this study, we examined the discriminatory effect of their spatiospectral features to capture the most relevant set of neural activities from electroencephalographic recordings that represent users' mental intent. In this regard, we model ERP waveforms using a sum of sinusoids with unknown amplitudes, frequencies, and phases. The effect of this signal modeling step is to represent high-dimensional ERP waveforms in a substantially lower dimensionality space, which includes their dominant power spectral contents. We found that the most discriminative frequencies for accurate decoding of visual attention modulated ERPs lie in a spectral range less than 6.4 Hz. This was empirically verified by treating dominant frequency contents of ERP waveforms as feature vectors in the state-of-the-art machine learning techniques used herein. The constructed predictive models achieved remarkable performance, which for some subjects was as high as 94% as measured by the area under curve. Using these spectral contents, we further studied the discriminatory effect of each channel and proposed an efficient strategy to choose subject-specific subsets of channels that generally led to classifiers with comparable performance.

show abstract

mentioning

confidence: 99%

mentioning

confidence: 99%

Learning Discriminative Spatiospectral Features of ERPs for Accurate Brain–Computer Interfaces

Abibullaev

Zollanvari

2019

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

show abstract

“…Para tener una mejor perspectiva del alcance de estos resultados, se compararon con los obtenidos por Yoon y cols. en 2017 [17], donde utilizaron el algoritmo MKL con tres tipos de rasgos discriminantes, 5 caracteres, dos sujetos de prueba y número de repeticiones r = {1, 2, 3, … , 15}, utilizando los registros del Dataset II of BCI competition III; en este estudio se promediaron los resultados de los dos sujetos con el conjunto de prueba completo, por lo que se obtuvieron los porcentajes de exactitud de deletreos (reconocimiento de carácter) de 85.1% para r = 8 y de 95.7 para r = 12. Tabla 3.…”

Section: Resultsunclassified

“…En 2017, Yoon y cols. [17] utilizaron, en un deletreador BCI P300, un nuevo algoritmo de clasificación, el Multiple Kernel Learning (MKL) con tres tipos de características discriminantes (señal en crudo raw, amplitud y parte negativa de la señal), Los registros de señal utilizados en esta BCI fueron tomados del Dataset II of BCI competition III, que consiste en registros EEG de 64 canales. El estudio demostró que la característica discriminante raw tiene un mayor peso que las características amplitud y parte negativa.…”

Section: Figunclassified

Desarrollo de una BCI utilizando el potencial P300 y la diadema Mindwave®

Vera¹,

Figueroa²,

Vásquez³

et al. 2017

RCS

View full text Add to dashboard Cite

In this paper we propose a recording, processing and classification scheme for a spelling Brain-Computer Interface application based on the P300 potential and the electroencephalography (EEG) device Mindwave-NeuroSky®; this device has only one EEG recording channel and is currently the most affordable in the market for this purpose. Open access software Openvibe was chosen for signal recording, and Matlab scripts were written for signal analysis and classification, the latter being based on Linear Discrimant Analysis (LDA). Five test subjects were exposed to several spelling sessions each, in which different acquisition and stimulation parameters were varied to evaluate the effects in character identification. Those parameters were the number of repetitions of the stimulation sequence of the speller character matrix (r=8, 12), the type of character sequence (chosen randomly of beforehand), and the EEG recording site (Pz or Oz). The subjects achieved, an average performance of 45% correct row and column identifications. In comparison, a very recent work in which 64 EEG channels were acquired, and a complex classification and feature extraction scheme was used, they achieved an average of 90.4% correct character identifications. The MindWave-Neurosky® headset was not designed for application in BCI systems. However, by looking at the results in this work, its application for BCI it seems feasible when thinking of combining the proposed approach with some word prediction scheme based on the context of the spelled phrase.

show abstract

“…One of the most prominent classification methods for ERP system is support vector machine (SVM) [ 15 – 18 ]. SVM is mathematically simple and, with sufficient knowledge of feature matrix, the experimenter can modulate the kernel for the target problem.…”

Section: Introductionmentioning

confidence: 99%

Spatial and Time Domain Feature of ERP Speller System Extracted via Convolutional Neural Network

Yoon

Lee

Whang

2018

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Feature of event-related potential (ERP) has not been completely understood and illiteracy problem remains unsolved. To this end, P300 peak has been used as the feature of ERP in most brain–computer interface applications, but subjects who do not show such peak are common. Recent development of convolutional neural network provides a way to analyze spatial and temporal features of ERP. Here, we train the convolutional neural network with 2 convolutional layers whose feature maps represented spatial and temporal features of event-related potential. We have found that nonilliterate subjects' ERP show high correlation between occipital lobe and parietal lobe, whereas illiterate subjects only show correlation between neural activities from frontal lobe and central lobe. The nonilliterates showed peaks in P300, P500, and P700, whereas illiterates mostly showed peaks in around P700. P700 was strong in both subjects. We found that P700 peak may be the key feature of ERP as it appears in both illiterate and nonilliterate subjects.

show abstract

Multiple kernel learning based on three discriminant features for a P300 speller BCI

Cited by 20 publications

References 21 publications

Learning Discriminative Spatiospectral Features of ERPs for Accurate Brain–Computer Interfaces

Learning Discriminative Spatiospectral Features of ERPs for Accurate Brain–Computer Interfaces

Desarrollo de una BCI utilizando el potencial P300 y la diadema Mindwave®

Spatial and Time Domain Feature of ERP Speller System Extracted via Convolutional Neural Network

Contact Info

Product

Resources

About