Detection of mental stress using EEG signal and classifier based on reward and punishment processing

Abstract

Mental stress is the main well being problem worldwide today. Mental stress is by far the most common, responsible for most of all mental-brain diseases. The high incidence, high impairment and heavy risk of illness make mental stress a major health issue facing the planet and is predicted to become the most prevalent disorder. Additionally, most of the time, suicidal people also hide true feelings and fail to communicate their psychiatric problems to physicians. Mental stress is a common as well as complex condition that has a limited explanation of its origin. Explanations regarding disease causes are centered both on their psychology and behavioral symptoms. Both factors that have led are currently being related by experiments investigating and how depressive individuals handle reward and punishment. Many studies have been proposed that people with such a disorder will be unable to utilize e ective knowledge to guide actions. The speci c issues that need to be addressed are, how to nd an easy, reliable and realistic way to diagnose mental stress and recognize mental stress early in order to keep it from becoming a serious and irreversible condition. To avoid diseases including health issues, the primary prevention of mental stress utilizing machine learning classi ers based on reward and punishment processing is important. The nervous system of man is the primary subject of mental stress. For all of this purpose, a machine learning framework is applied to evaluate the electroencephalogram signals for fty individuals in our proposed model. For a successful mental stress detection application, this layout has implemented a combination of features that supplies nine ML classi ers which are Support Vector Machine, Random Forests, K-Nearest Neighbors, Decision Tree classi er, AdaBoost classi er, Extra Trees classi er, Bagging classi er, Gradient Boosting classi er, Gaussian Na ve Bayes Classi er to identify Comorbidity based on reward and punishment processing. The experimental results indicate that Extra Trees classi er, Gaussian Na ve Bayes Classi er and Random Forests have higher success predicting the presence of mental stress. This implemented classi er based on reward and punishment processing systems using EEG signals has the ability to statistically detect mental stress.