Automated identification of sleep states from EEG signals by means of ensemble empirical mode decomposition and random under sampling boosting

Hassan, Ahnaf Rashik; Bhuiyan, Mohammed Imamul Hassan

doi:10.1016/j.cmpb.2016.12.015

Cited by 222 publications

(106 citation statements)

References 31 publications

Supporting

Mentioning

100

Contrasting

Unclassified

Order By: Relevance

“…In Figure 2, the information presented in Figure 1 is now depicted in terms of the box plot specifically for the selected range of frequency (30-50 Hz), highlighting the main differences among the awake stage and all the remainders. As suggested in [26], the non-overlapping and non-collinear notches of the box plots suggest good inter-class variation of the selected features. The mean of the FFT coefficients from 30 to 50Hz allows the classifier to distinguish the awake stage from the others.…”

Section: Mean Of the Fft Coefficients From 30 Hz To 50 Hzmentioning

confidence: 93%

“…In opposition to them, the work of Ronzhina et al [11] rated statistical moments as important parameters in signal analysis at time domain. More recently, Hassan and Bhuiyan [26] and Silveira et al [12] extracted statistical information after decomposing the signals in time-frequency domain, and used them as classification attributes for sleep staging. Nevertheless in [26] the preprocessing stage applied an expensive datadriven Ensemble Empirical Mode Decomposition (EEMD) method for obtaining the frequency bands.…”

Section: Related Workmentioning

confidence: 99%

“…More recently, Hassan and Bhuiyan [26] and Silveira et al [12] extracted statistical information after decomposing the signals in time-frequency domain, and used them as classification attributes for sleep staging. Nevertheless in [26] the preprocessing stage applied an expensive datadriven Ensemble Empirical Mode Decomposition (EEMD) method for obtaining the frequency bands. The high classification performances obtained in both works highlight the potential in associating time-frequency information with statistical tools for defining features.…”

Section: Related Workmentioning

confidence: 99%

“…Sigma oscillations occur mainly during S2 stage [19]. Corsi-Cabrera et al [20] claim that the beta rhythm (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) is primarily associated to the awake stage. Finally, the gamma rhythm (30-80 Hz) is related to the working memory and attention [21] and, thus, arises mainly during the awake stage.…”

Section: Related Workmentioning

confidence: 99%

See 3 more Smart Citations

Sleep Stages Classification Using Spectral Based Statistical Moments as Features

Braun

Kozakevicius

Silveira

et al. 2018

RITA

View full text Add to dashboard Cite

Abstract:In the pursuit of portable, efficient and effective sleep staging systems, researchers have been testing a massive number of combinations of EEG features and classifiers. State of the art sleep classification ensembles achieve accuracy in the order of 90%. However, there is presently no consensus regarding the best set of features for identifying sleep stages with a single EEG channel, leading researchers to modify the feature selection according to the number of classification stages. This paper introduces a reduced set of frequency-domain features capable of yielding high classification accuracy (90.9%, 91.8%, 92.4%, 94.3% and 97.1%) for all 6-to 2-state sleep stages. The proposed system uses fast Fourier transform (FFT) to convert data from Pz-Oz EEG channel into the frequency domain. Afterwards, eight statistical features are extracted from specific frequency ranges associated to brain rhythms, feeding a random forest classifier. Keywords: sleep stage classification -random forest -FFT -statistical momentsResumo: Na busca por sistemas portáveis, eficientes e efetivos para classificação de sono, pesquisadores têm testado um grande número de combinações de atributos de sinais EEG e classificadores. O estado da arte dos aplicativos de classificação para estágios de sono atinge taxas de acerto na ordem de 90%. Contudo, não existe um consenso em relação ao melhor grupo de atributos para identificar estágios de sono em sinais EEG de umúnico canal, levando pesquisadores a modificar a seleção dos atributos de acordo com o número de estágios a serem classificados. Este trabalho apresenta um conjunto reduzido de atributos, capaz de produzir uma classificação com taxas elevadas de acerto (90.9%, 91.8%, 92.4%, 94.3% and 97.1%) para todos os agrupamentos de estágios de sono (de 6 a 2 estágios). O sistema proposto utiliza a transformada rápida de Fourier (FFT) para converter os dados do canal de EEG Pz-Oz para o domínio da frequência. Posteriormente, oito atributos estatísticos são extraídos de faixas de frequência específicas associadas a ritmos cerebrais, alimentando um classificador do tipo floresta aleatória. Palavras-Chave: classificação de estágios de sono -floresta aleatória -FFT -momentos estatísticos

show abstract

Section: Mean Of the Fft Coefficients From 30 Hz To 50 Hzmentioning

confidence: 93%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Sleep Stages Classification Using Spectral Based Statistical Moments as Features

Braun

Kozakevicius

Silveira

et al. 2018

RITA

View full text Add to dashboard Cite

show abstract

“…Previous studies have also demonstrated that electroencephalograph (EEG) signals vary with sleep progresses, thus EEG has been widely used for assessing sleep stages 5,6 . Hassan et al extracted many features such as tunable-Q factor wavelet transform 7 , spectral features, empirical mode decomposition 8 , Gaussian parameters and statistical features 9 from EEG signals to classify sleep states. Liang et al employed multi-scale entropy and auto-regressive model parameters to classify sleep stages 10 , considering that the characteristics of EEG signals are somewhat chaotic, not only the traditional features, the correlation dimension (CD) derived from nonlinear dynamical analysis are also applied to investigate the dynamic characteristics of EEG.…”

Section: Introductionmentioning

confidence: 99%

Improved GP algorithm for the analysis of sleep stages based on grey model

Li¹,

Xie²,

Zhao³

et al. 2017

ScienceAsia

View full text Add to dashboard Cite

Correlation dimension analysis of EEG signals is widely used to access sleep stages. However, the standard Grassberger-Procaccia (GP) algorithm used to calculate the correlation dimension is very time consuming. To overcome this problem, an algorithm that combines the grey model and GP algorithm (GM-GP) is proposed. The results show that the correlation dimensions computed from GP and GM-GP are highly correlated, and the significance between the CDs in different stages of GM-GP is similar to GP. Furthermore, the computation time of the proposed method is at most 5% of that of the GP. The proposed algorithm is suitable for the real-time monitoring of sleep stages, which can provide a deeper understanding of brain function.

show abstract

Polysomnography and Sleep Analysis

2020

Body Sensor Networking, Design and Algorithms

View full text Add to dashboard Cite

Automated identification of sleep states from EEG signals by means of ensemble empirical mode decomposition and random under sampling boosting

Cited by 222 publications

References 31 publications

Sleep Stages Classification Using Spectral Based Statistical Moments as Features

Sleep Stages Classification Using Spectral Based Statistical Moments as Features

Improved GP algorithm for the analysis of sleep stages based on grey model

Polysomnography and Sleep Analysis

Contact Info

Product

Resources

About