Mental Stress Quantification Using EEG Signals

Abstract

Mental stress has been identified as one of the major contributing factors that leads to various diseases such as heart attack and stroke. To avoid this, stress quantification is very important for clinical intervention and disease prevention. In this study, we investigate the feasibility of exploiting Electroencephalography (EEG) signals to discriminate stress from rest state in mental arithmetic tasks. The experimental results showed that there were significant differences between the rest state and under stress at three levels of arithmetic task levels with p-values of 0.03, 0.042 and 0.05, respectively. We thus confirm the feasibility of EEG signals in detecting mental stress levels. Using support vector machine (SVM) we could detect mental stress with an accuracy of 94%, 85%, and 80% at level one, level two and level three of arithmetic problem difficulty respectively.