Stress Management Using Physiological Signals and Audio Stimulation

Abstract

Stress has a significant role in the development of a wide variety of mental, psychological, emotional, behavioral, and physical illnesses. Additionally, there is substantial evidence in the literature that stress impairs vigilance. Thus, early stress detection, vigilance enhancement, and stress mitigation may aid in the prevention of a wide range of diseases and improve human health. The purpose of this thesis is to examine the effects of binaural beat stimulation (BBs) on increasing alertness and reducing mental stress in the workplace. We devised an experiment in which participants were subjected to time pressure and negative feedback while completing the Stroop Color-Word Task (SCWT). Then, we used 16 Hz BBs to improve vigilance and reduce stress levels. Functional Near-Infrared Spectroscopy (fNIRS), salivary alpha-amylase, behavioral data, and subjective reactions were used to determine the levels of stress. We quantified the level of stress using statistical analysis, functional connectivity based on Partial Directed Coherence (PDC), Graph Theory Analysis (GTA) and Convolution Neural Network (CNN). We discovered that BBs substantially increased target detection accuracy by 11.05% (p<0.001), decreased effort and temporal demand, boosted performance, and decreased cortisol levels. The deep learning results indicated that the CNN technique combined with PDC features is capable of discriminating between four distinct mental states (vigilance, enhancement, stress, and mitigation) with an average accuracy of 70.62%, a sensitivity of 68.39%, and a specificity of 90.76%.