EEG-Based brain-machine interface for categorizing cognitive sentimental emotions

Abstract

The development of brain-machine interfaces (BMIs) has revolutionized the study of neuroscience by making it possible for the brain to communicate directly with outside objects. In this study, EEG brainwave data is used to categorize emotional experiences using both individual and group methods. We employ a four-electrode resolution (TP9, AF7, AF8, and TP10) commercial MUSE EEG headband. Film clips with clear emotional content elicit both good and negative emotions. For one minute per session, neutral resting data is also obtained without external stimuli. To do this, we use machine learning algorithms to decode and interpret participant EEG data as they perform activities that evoke a range of emotional reactions. Relevant characteristics are collected from the EEG signals using intensive data preprocessing, feature extraction, and selection approaches to identify the underlying patterns of cognitive sentimental feelings. Following that, the development and assessment of classification models, such as Gradient Artificial Neural Networks (G-ANN), use the retrieved features as input. In conclusion, this study presents an EEG-based BMI system for categorizing cognitive sentimental emotions. The proposed G-ANN achieves a high accuracy of 98.59%, demonstrating superior performance compared to existing methodologies.