Mental workload vs. stress differentiation using single-channel EEG

• Published in: CMBEBIH 2017 pp. 511 - 515
• Main Authors: Secerbegovic, A., Ibric, S., Nisic, J., Suljanovic, N., Mujcic, A.
• Format: Book Chapter
• Language: English
• Published: Singapore Springer Singapore
• Series: IFMBE Proceedings
• Subjects: Electrodermal Activity; Empirical Mode Decomposition Algorithm; Ensemble Empirical Mode Decomposition; Hilbert Huang Transform; Mental Workload
• ISBN: 9811041652; 9789811041655
• ISSN: 1680-0737; 1433-9277
• DOI: 10.1007/978-981-10-4166-2_78

The emergence of wearable low-cost wireless devices has allowed for continuous acquisition of physiological signals. Recently number of studies have applied these acquisition systems in different types of health monitoring. Since continuous elevation of stress hormones can have negative impact on individuals’ health, it is important to recognize and possibly prevent stress episodes in working environments. In this paper, we have tested if single-channel electroencephalography (EEG) signals can be utilized in assessment of different levels of mental workload and stress. Experimental study was conducted in laboratory settings with nine participants. In addition to EEG signals, we have acquired electrocardiogram (ECG) and electrodermal activity (EDA) recordings during all stages. Two scenarios are tested: first group of participants was introduced to only mental workload assignments, while second group was tested with mental workload and public speaking task as an stress inducing assignment. The experimental results show that EEG features have an acceptable separation ability between investigated states, where best classification accuracy, obtained between relaxed and high mental workload states, was 86.66%. Compared to only ECG or EDA features, EEG-based classification accuracy is higher in both scenarios, but lower in comparison with combined features from all three physiological signals.