Estimating Emotions from EEG Signals Using Gated Recurrent Unit Architecture

The classification of emotional states using electroencephalogram (EEG) signals has emerged as a critical area of research for understanding neural mechanisms underlying emotions and advancing mental health applications. This study presents a comparative analysis of two models that were implemented both employing Gated Recurrent Units (GRU) for classification of three emotional states: negative, neutral, and positive. The models were evaluated by using accuracy, precision, and robustness in analyzing EEG signals recorded using standard electrode placements. The experimental results indicate that Reddy’s model outperformed Atkins’s model by achieving a classification accuracy of 98.438%, while Atkins’s model attained an accuracy of 96.875%. This result demonstrates the superior performance of Reddy’s model across all emotional states. The findings underscore the potential of GRU-based architectures for emotion classification and highlight the importance of model selection in EEG signal processing for applications in neuroscience and artificial intelligence.