Stress detection from EEG using power ratio

Abstract

Stress correlates itself as a mental conscious and emotion within a person that influences mental ability and decision-making skills, which results in an inappropriate work. Studies have recently developed to detect the stress in a person while performing different tasks. One of the methods is through Electroencephalograph (EEG). These are the bioelectrical signals generated in a human body while performing the tasks and thus describes the activity of the brain. Any action taken by a person changes the properties of these signals. This present work focuses on the classification of baseline (relax) and stress detection using EEG sub-band power ratio as features. Support vector machine (SVM) classifier with different kernel function parameters and K-nearest neighbor (KNN) classifier with a different number of neighbors with holdout and 10-fold crossvalidation technique were used to classify power ratio features in order to detect stress. To evaluate the classifier performance various performance metrics were used. It is observed that KNN with a number of neighbors as one, with Euclidean distance gives better performance in both validation techniques and also anterior frontal channel Fp1 that is placed at the left side of the brain itself gives a good accuracy of 99.42%. The performance of the proposed method is verified on a publicly available mental arithmetic dataset where stress is induced while performing the mental cognitive workload i.e., mental serial subtraction.

Stress Classification by Multimodal Physiological Signals Using Variational Mode Decomposition and Machine Learning

“…Most stress related studies reported for specific/limited channels [ 37 – 39 ]. The proposed approach for the stress classification has outperformed other existing methods [ 3 , 40 – 43 ] by achieving 100% accuracy with minimum time of mental arithmetic activity and has also given an insight for the identification of marker lobewise (frontal, temporal, central, and occipital) rather than selection of channels in a generalized way. Our brain activity consists of interchange of ions between neurons which results into current flow through synaptic mode.…”

Neural Decoding of EEG Signals with Machine Learning: A Systematic Review

“…Our investigation revealed that FFT, PCA, WT, and AR were the most widely used and effective methods for feature extraction among the reviewed articles in this domain; these methods were reported as superior 31%, 19%, 15%, and 15% of the time, respectively. For example, FFT has been implemented by researchers in several published studies [ 81 , 196 , 211 , 239 , 240 , 241 , 242 ] to extract features on the basis of the frequency of the EEG signals; however, another study has confirmed AR as one of the most reliable methods [ 243 ]. We also identified several essential feature extraction techniques that were less frequently applied in MWL tasks, including entropy and HHT, which elicited features of both nonlinear and non-stationary signals.…”

“…Likewise, Shah and Ghosh [ 55 ] have developed a real-time classification system by using PCA and a simple KNN classification algorithm. Interestingly, the studies in references [ 196 , 241 ] have proposed using FFT to evaluate the PSD on the basis of the time domain features incorporated in different ML models for classification. The results indicated the highest accuracy with KNN, at 99.42% and 90.5%, respectively.…”

“…Seven types of ML and DL classification algorithms were used to study MWL tasks: CNN, LSTM, SVM, KNN, ANNs, RF, and XGBoost. Two studies [ 196 , 241 ] have used KNN classifiers with the feature extraction method FFT and compared the performance of KNN versus other algorithms. Both studies have reported that KNN with FFT results in the most accurate performance.…”

“… Sleep-EDF database Entropy RF SVM DT Accuracy = 97.8 [ 180 ] 2018 SD 25 subj. 100 channels BONN database PCA RBF-SVM Accuracy = 100 [ 185 ] 2021 SD BONN database VMD RF Accuracy = 98.7–100 [ 204 ] 2020 SS 2 channels Sleep-EDF database FFT RBF-SVM Accuracy = 87.8 [ 196 ] 2020 MWL 36 subj. 19 channels PhysioBank database FFT KNN SVM Accuracy = 99.4 [ 154 ] 2020 Stress 33 subj.…”