Human Mental Stress Evaluation Using Wearable EEG and ECG Sensors

Mental stress could pose a negative impact on an individual's health and affects their daily life. This study aims to analyse human mental stress behaviour using wearable biosensors, i.e. electrocardiogram (ECG) and electroencephalogram (EEG). We also investigated the validity of a heart rate monitor (HRM) by comparing its measurements against the ECG. A total of 17 participants (9 males and 8 females) were recruited and they were required to perform relax and stress activities (incongruent and congruent Stroop Colour Word tests) in the experimental set-up. All the participants also completed the given questionnaire to rate their stress scores. Our results showed that there are significant differences (p < 0.05) in distinguishing mental stress state from relaxation using EEG power ratio of alpha/beta (stress: 1.06 ± 0.35 vs. relax: 1.44 ± 0.59). In the validation study, there is an agreement between ECG and HRM, in which high correlations of mean R-R were found both at rest (r = 0.999, p < 0.001) and stress sessions (r = 0.994, p < 0.001) meanwhile strong agreement was also noted from the individual sessions (ρc for rest sessions: R1 = 0.983, R2 = 0.968 and R3 = 0.975; ρc for stress sessions: S1 = 0.980 and S2 = 0.981). For the validity study on HRV features between the two devices, the highest correlations exist in relax sessions for pNN50 (r = 0.988, d = 0.02) and NN50 (r = 0.987, d = 0.10), subsequently followed by SDNN (r = 0.804, d = 0.09) and RMSSD (r = 0.684, d = 0.09).