RESUMO
Mental stress has been identified as a significant cause of several bodily disorders, such as depression, hypertension, neural and cardiovascular abnormalities. Conventional stress assessment methods are highly subjective and tedious and tend to lack accuracy. Machine-learning (ML)-based computer-aided diagnosis systems can be used to assess the mental state with reasonable accuracy, but they require offline processing and feature extraction, rendering them unsuitable for real-time applications. This paper presents a real-time mental stress assessment approach based on convolutional neural networks (CNNs). The CNN-based approach afforded real-time mental stress assessment with an accuracy as high as 96%, the sensitivity of 95%, and specificity of 97%. The proposed approach is compared with state-of-the-art ML techniques in terms of accuracy, time utilisation, and quality of features.
RESUMO
In mental stress studies, cerebral activation and autonomic nervous system are important distinctly. This study aims to analyze disparities associated with scalp potential, which may have impact on autonomic activation of heart during mental stress. Ten healthy subjects participated in this study that performed arithmetic tasks in stress and control environment. Task difficulty was calculated from their correct responses. During the experiment, electroencephalogram (EEG) and electrocardiogram (ECG) signals were recorded concurrently. Sympathetic innervation of heart was estimated from heart rate (HR), which is extracted from the ECG. The value of theta Fz/alpha Pz was measured from EEG scalp potential. The results show a significant surge in the value of theta Fz/alpha Pz in stress as compared to baseline (p<0.013) and control (p<0.042). The results also present tachycardia while in stress as compared to baseline (p<0.05). Task difficulty in stress is also considerably higher than control environment (p<0.003).