Electroencephalography Changes in Emotional Recognition: A Signal Processing Approach Using Wavelet Features

This is the authors' abstract. We don't add key points for this paper.

Background motion recognition using electroencephalography (EEG) has gained significant attention in the fields of affective computing and human–computer interaction due to its ability to capture real-time neural responses. EEG signal variations associated with emotional states can be analysed using wavelet-based signal processing within the arousal–valence framework. Purpose This study aims to analyse EEG signal variations associated with emotional states using a wavelet-based signal processing approach within the arousal–valence framework. Methods A prospective observational study was conducted on 48 healthy participants. Emotional states were induced using image, video and music stimuli categorised under the arousal–valence model. EEG signals were recorded using the standard 10–20 electrode placement system. The acquired signals were pre-processed and analysed using discrete wavelet transform (DWT) to extract time–frequency features, including energy and entropy. Statistical analysis was performed to evaluate hemispheric differences and gender-based variations. Results Differences in EEG activity were observed between the left and right hemispheres across multiple emotional states (p < .05). Video and music stimuli produced stronger and more distinguishable EEG responses compared to image-based stimuli. Negative emotions demonstrated higher discriminative patterns than positive emotions. Gender-based variations were also observed in EEG responses. Conclusion Wavelet-based EEG feature analysis provides meaningful and discriminative information for emotion recognition and may provide a robust feature set for future machine learning-based emotion recognition systems. Further studies incorporating classification algorithms and larger data sets are required to validate their predictive performance.