Optimal EEG channels selection for alcoholism screening using EMD domain statistical features and harmony search algorithm

Bavkar, Sandeep; Iyer, Brijesh; Deosarkar, S. B.

doi:10.1016/j.bbe.2020.11.001

Cited by 27 publications

(16 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…LS-SVM and KNN were used to classify the extracted features into alcoholic and normal signals. Bavkar et al, 2021) also applied empirical mode decomposition to classify alcoholic EEG signals. The extracted features using empirical mode decomposing were sent to the KNN classifier.…”

Section: Discussionmentioning

confidence: 99%

An Eigenvalues-Based Covariance Matrix Bootstrap Model Integrated With Support Vector Machines for Multichannel EEG Signals Analysis

Al‐Hadeethi

Abdulla

Diykh

et al. 2022

Front. Neuroinform.

View full text Add to dashboard Cite

Identification of alcoholism is clinically important because of the way it affects the operation of the brain. Alcoholics are more vulnerable to health issues, such as immune disorders, high blood pressure, brain anomalies, and heart problems. These health issues are also a significant cost to national health systems. To help health professionals to diagnose the disease with a high rate of accuracy, there is an urgent need to create accurate and automated diagnosis systems capable of classifying human bio-signals. In this study, an automatic system, denoted as (CT-BS- Cov-Eig based FOA-F-SVM), has been proposed to detect the prevalence and health effects of alcoholism from multichannel electroencephalogram (EEG) signals. The EEG signals are segmented into small intervals, with each segment passed to a clustering technique-based bootstrap (CT-BS) for the selection of modeling samples. A covariance matrix method with its eigenvalues (Cov-Eig) is integrated with the CT-BS system and applied for useful feature extraction related to alcoholism. To select the most relevant features, a nonparametric approach is adopted, and to classify the extracted features, a radius-margin-based support vector machine (F-SVM) with a fruit fly optimization algorithm (FOA), (i.e., FOA-F-SVM) is utilized. To assess the performance of the proposed CT-BS model, different types of evaluation methods are employed, and the proposed model is compared with the state-of-the-art models to benchmark the overall effectiveness of the newly designed system for EEG signals. The results in this study show that the proposed CT-BS model is more effective than the other commonly used methods and yields a high accuracy rate of 99%. In comparison with the state-of-the-art algorithms tested on identical databases describing the capability of the newly proposed FOA-F-SVM method, the study ascertains the proposed model as a promising medical diagnostic tool with potential implementation in automated alcoholism detection systems used by clinicians and other health practitioners. The proposed model, adopted as an expert system where EEG data could be classified through advanced pattern recognition techniques, can assist neurologists and other health professionals in the accurate and reliable diagnosis and treatment decisions related to alcoholism.

show abstract

Section: Discussionmentioning

confidence: 99%

An Eigenvalues-Based Covariance Matrix Bootstrap Model Integrated With Support Vector Machines for Multichannel EEG Signals Analysis

Al‐Hadeethi

Abdulla

Diykh

et al. 2022

Front. Neuroinform.

View full text Add to dashboard Cite

show abstract

“…However, despite the AUD burden, it has been less common for researchers to study and analyze EEG signals to select optimal channels. To the best of the authors' knowledge, only a few studies have investigated dimensionality reduction by selecting optimal EEG channels to address AUD concerns (Bavkar et al, 2021;Ong et al, 2006;Palaniappan et al, 2002;Shooshtari and Setarehdan, 2010;Zhu et al, 2014). The aforementioned EEG analysis studies reduced the dimension of the EEG dataset either by applying feature extraction techniques alone or by performing features or channel selection based on extracted features.…”

Section: Related Workmentioning

confidence: 99%

Discrete to Continuous Algorithm for Optimal Channel Selection to Detect Alcoholism through EEG Recordings

Sedrati¹,

Rhalem²,

Aqili³

et al. 2023

Journal of Computer Science

View full text Add to dashboard Cite

Alcoholism is a serious public health issue, and early diagnosis of this brain disease can be performed by analyzing Electroencephalogram (EEG) signals. However, the high dimensionality of EEG datasets requires significant computational time and resources for the automatic processing of EEG signals. This study proposes a novel method to reduce EEG dataset dimensionality using a Discrete to Continuous algorithm (DtC) by selecting optimal EEG channels. The DtC approach compares alcoholic and nonalcoholic EEG signals as two-time series in a two-dimensional space based on a distance measurement between the two-time series. The Dynamic Time Warping (DTW) algorithm is used to compare the performance of the DtC approach. Classification performance metrics were evaluated for both the DtC and DTW algorithms. The optimal selected channels by our approach are the C3, CP5, PO7, and F8 channels with accuracy values of 100, 100, 94 and 81%, respectively. These findings are consistent with previous research on statistical analysis and machine learning methods and with the DTW algorithm results. Our findings are also in line with scientific evidence from clinical research. The DtC approach was efficient in selecting the best channels to reduce the EEG dataset dimensionality, allowing us to select four out of the 64 EEG channels (C3, CP5, PO7, and F8) that retain essential information related to alcoholism, which is useful in reducing computational time and resources during the classification task of alcoholic EEG.

show abstract

“…The excellent functional neuroimaging capabilities such as noninvasive nature, high temporal resolution, portability, inexpensiveness, easy accessibility, and safe nature make EEG a comprehensive tool for the brain imaging task as compared to that of functional magnetic resonance imaging (fMRI), positron emission tomography (PET), and magnetoencephalogram (MEG). 3 EEG based automated diagnosis using computer-assisted solutions have been applied to many clinical applications such as epileptic-seizure detection, [4][5][6][7][8] alcohol-use disorder (AUD), 9 multi-class classification of epileptic seizure types, 10 Parkinson's, 11 and depression. 12 Apart from the clinical applications, EEG based emotion recognition 13 and identity authentication 14 have been emerged as the interesting areas of research.…”

Section: Introductionmentioning

confidence: 99%

A novel method for EEG based automated eyes state classification using recurrence plots and machine learning approach

Khosla

Khandnor

Chand

2022

Concurrency and Computation

View full text Add to dashboard Cite

Automated eyes-state classification from the EEG signals using nonlinear analysis tools is a new area of research. Based upon the theory of nonlinear analysis, recurrence plots (RPs) and recurrence quantification analysis (RQA) are of greater significance that help in understanding the chaotic and recurrence behavior of the dynamically occurring complex physiological signals. In our study, a novel method is proposed combining the RPs with the machine learning based algorithms for automated classification of EEG signals into eyes-open and eyes-close states. A huge dataset of 109 subjects has been acquired from the PhysioNet database. Each of the six RQA-based measures (recurrence rate, determinism, entropy, laminarity, trapping time, and longest vertical line)has been extracted from 64 EEG channels. Feature selection has been performed using genetic algorithm. The selected features have been averaged and combined to form a six-dimensional input vector which shows statistically significant differences (p < 0.01) between the two states. It is fed to different machine learning based algorithms such as logistic-regression, support vector machine, random forest, k-nearest neighbor, Gaussian naïve Bayes, and adaptive boosting. Logistic regression achieves the highest performance results in terms of accuracy, F1 score, precision, recall, and specificity of 97.27%, 97.17%, 98.26%, 96.36%, and 98.18%, respectively, with the least testing time of the model as 2.52 ms. Therefore, our method might be of greater significance in the development of the practical applications.

show abstract

Optimal EEG channels selection for alcoholism screening using EMD domain statistical features and harmony search algorithm

Cited by 27 publications

References 28 publications

An Eigenvalues-Based Covariance Matrix Bootstrap Model Integrated With Support Vector Machines for Multichannel EEG Signals Analysis

An Eigenvalues-Based Covariance Matrix Bootstrap Model Integrated With Support Vector Machines for Multichannel EEG Signals Analysis

Discrete to Continuous Algorithm for Optimal Channel Selection to Detect Alcoholism through EEG Recordings

A novel method for EEG based automated eyes state classification using recurrence plots and machine learning approach

Contact Info

Product

Resources

About