A real-time EEG-based classification system using Random Forest and Support Vector Machine models detected stress with accuracies of 71% and 67%, respectively.
Machine learning models applied to EEG signals can classify stress induced by different stimuli in real-time with up to 71% accuracy.
Absolute Event Rate: 71% vs 67%
The significance of understanding stress responses has gained increasing attention due to its profound impact on mental health and cognitive functioning. Prior studies have explored the potential of electroencephalography (EEG) in detecting stress, focusing on brain wave patterns like alpha and beta waves. There is a recognized need for the development of advanced methods that can offer real-time classification of stress induced by a wide range of stimuli. This research aims to develop a robust real-time EEG-based classification system to detect and analyze stress levels in response to various stress-inducing tasks. The methodology involved collecting EEG signals and analyzing them through signal processing and machine learning techniques. The Random Forest and Support Vector Machine (SVM) models were employed, achieving accuracies of 71% and 67% respectively. The model displayed a high level of precision in identifying stress. The results indicate that different stressors elicit distinct EEG patterns, with cognitive tasks engaging the frontal brain regions more intensely, while emotional tasks show reduced frontal activity. The model's performance highlights its potential for real-time applications in stress management and mental health monitoring. These findings underscore the effectiveness of EEG in real-time stress detection and pave the way for more adaptive and personalized stress management systems.
Mahmoudi et al. (Tue,) conducted a other in Stress. Real-time EEG-based classification using machine learning (Random Forest and SVM) was evaluated on Stress detection accuracy. A real-time EEG-based classification system using Random Forest and Support Vector Machine models detected stress with accuracies of 71% and 67%, respectively.