EEG-based Statistical Analysis on Determining the Stress Mental State on Police Personnel

Abstract

Suicide is a global phenomenon that occurs worldwide, including in Indonesia. This is due to complications from high and severe stress because of economic factors, family and environmental problems. High and severe stresses are not only experienced by the individual in a society, but also by government staff such as members of the police. The heavy workload, such as handling demonstrations with high escalation, is one of the factors that cause police members to experience high stress. Psychological assistance and healing such as in-depth interviews and psychological assessment tests have been carried out to map the psychological barriers and stressful conditions of police officers. The use of an Electroencephalogram (EEG) is one of the physiological signals that can be used to measure and recognize stress based on data on human brain activity. This research is exploring stress state by using eeg signal analysis in time domain. The analysis is done based on EEG time-domain features in the theta (4-8Hz), alpha (8-13Hz), and beta (13-30Hz) frequency bands from two different channels, namely F3 and F4 in the 10/20 EEG system. Twenty members of the state police (10 under stress conditions and 10 in normal conditions) are involved in this study. Statistical features such as Mean, Standard Deviation, and Zero Crossing are used to distinguish between stress and normal conditions. The experimental results showed that the Standard Deviation feature on the Alpha subband provided the highest difference in comparing between normal and stress conditions. In the classification of stress and normal conditions using several algorithms, SVM indicates the highest classification accuracy (88.90%), compared to other algorithms such as Random Forest (86.10%), K-NN (77.80%) and Decision Tree (77,80%).