EEG-Based Emotion Recognition using Hybrid CNN-LSTM Model for Mental State Analysis

The progress of brain-computer interface (BCI) technology has allowed for real-time understanding of human emotions through electroencephalogram (EEG) signals. This research suggests a hybrid deep learning framework that combines Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks for emotion recognition using the DEAP dataset. EEG signals are processed to remove noise, normalized, and turned into spatial-temporal representations that are suitable for deep feature extraction. The CNN part captures spatial dependencies among EEG channels, while the LSTM captures temporal changes over time. The CNN-LSTM model showed higher classification accuracy in identifying emotional states across valence and arousal dimensions than traditional methods. The extended model includes an attention mechanism to improve feature weighting and make results easier to understand. This hybrid approach helps develop strong affective computing systems that can be used in mental health monitoring, adaptive learning, human- computer interaction, and emotion-aware virtual assistants. The results demonstrate the potential of combining CNN, LSTM, and attention-based deep architectures for effective emotion recognition from non-invasive EEG signals.