This paper presents a machine learning-based approach for detecting stress using electroencephalogram (EEG) signals. The dataset consists of EEG data recorded from 40 subjects (26 males, 14 females) with a mean age of 21.5 years, during the performance of various cognitive tasks such as the Stroop color-word test, arithmetic problem-solving, and mirror image identification, as well as during a relaxation state. The EEG signals were recorded using a 32-channel Emotiv Epoc Flex gel kit and were segmented into 25-second non-overlapping epochs. The data was preprocessed to remove baseline drifts using a Savitzky-Golay filter and to eliminate artifacts through wavelet thresholding. Machine learning models are then employed to identify stress patterns in the EEG data, aiming to facilitate the development of brain-computer interfaces and improve stress detection techniques. The findings from this study may provide valuable insights into stress identification and its potential applications in various domains, including healthcare and human-computer interaction.
Stress is a prevalent issue in modern society, significantly impacting both physical and mental health across various demographics. The World Health Organization (WHO) identifies stress as a major contributor to the global burden of disease, with prolonged or chronic stress being linked to serious health complications such as cardiovascular diseases, depression, anxiety disorders, and cognitive impairments. Chronic stress disrupts homeostasis and can weaken the immune system, impairing the body's ability to fight off infections and increasing vulnerability to illnesses. Given its far-reaching effects, early detection and effective management of stress are essential for improving quality of life and reducing healthcare costs. Consequently, stress detection has become a growing area of interest in the fields of healthcare, psychology, and neuroscience, with researchers exploring innovative ways to accurately assess stress levels in real-time environments.
Traditional methods for stress assessment often rely on self-report questionnaires, physiological measurements such as heart rate variability (HRV) and cortisol levels, and behavioral observations. While these methods provide valuable insights, they have limitations, including subjectivity, the need for invasive procedures, or the inability to capture real-time changes. In contrast, Electroencephalogram (EEG) signals, which measure the brain's electrical activity through electrodes placed on the scalp, offer a non-invasive and objective approach to understanding stress patterns. EEG-based stress detection leverages brainwave frequencies (delta, theta, alpha, beta, and gamma) that change in response to cognitive and emotional stressors. For example, increased beta activity and decreased alpha waves are often associated with heightened stress and mental effort. This modality enables continuous, real-time monitoring of brain activity, making it highly suitable for applications in brain-computer interfaces (BCI), human-computer interaction (HCI), and healthcare monitoring systems.
Hardware:
1.PC
Software:
1.Python idle 3.8
