LSTM Algorithm for the Detection of Mental Stress in EEG

Abstract

Stress is a prevalent mental health issue that can lead to severe consequences if not addressed properly. In recent years, electroencephalography (EEG) signals have gained attention for stress detection. However, most existing approaches rely on pre-processed features, which can be time-consuming and may not capture all the relevant information in the EEG signals.In this paper, we proposed a novel deep-learning approach for real-time stress detection using raw EEG signals. Our approach utilizes a long short-term memory (LSTM) network to automatically capture features and classify the stress level. Our method allows for capturing all the relevant information in the EEG signals, without the need for manual feature engineering.We evaluated our approach on the DEAP dataset, which includes EEG signals from 32 subjects under various emotional states. Experimental results demonstrate that our approach achieves state-of-the-art performance in stress detection, with an accuracy of approximately 94%. Our proposed approach has the potential for real-world applications, such as stress management in the workplace and mental health monitoring in clinical settings.