Main Article Content
Stress is a common issue in modern society and can lead to various health problems when left unaddressed. Accurate stress detection is, therefore, crucial in order to provide effective interventions and improve overall well-being. This study presents the implementation of a Long Short-Term Memory (LSTM) model to detect stress using electroencephalogram (EEG) signals. EEG signals were collected from a sample of participants while they were exposed to stress-inducing tasks and control tasks. The data was pre-processed using filtering and artifact removal techniques to ensure high quality and reliability. The pre-processed EEG signals were then used to extract relevant features, such as spectral power and coherence, which served as inputs to the LSTM model. A deep learning architecture was developed, incorporating the LSTM layers and other components to optimize the model's performance. The LSTM model was trained and validated using the available dataset. The results showed that the LSTM model significantly outperformed the other algorithms in terms of accuracy, sensitivity, and specificity. Furthermore, the model demonstrated robustness in detecting stress across various tasks and EEG channels. These findings suggest that LSTM-based models have the potential to be used as effective tools for stress detection in real-life scenarios, and can contribute to the development of more personalized stress management interventions. Future research should focus on refining the model and exploring its applicability in different populations and settings.