Xianwei Lin ¹ Xinyue Wu ² Zhengting Cai ^1* Zihan Zhang ^1* Guangdong Xie ^1* Lianxin Hu ^1* Zefeng Wang ^1* Laurent Peyrodie ^3*

• ¹ Huzhou University, Huzhou, China
• ² Beijing University of Chinese Medicine, Beijing, Beijing Municipality, China
• ³ Junia School of Engineering, Catholic University of Lille, Lille, Nord-Pas-de-Calais, France

Music profoundly influences human emotions, capable of evoking a wide array of emotional responses. This ability has been effectively applied in the field of music therapy. Given the close relationship between music and language, researchers have focused on exploring how music affects brain activity and cognitive processes, integrating artificial intelligence with advancements in neuroscience. In this study, 120 subjects aged 19 to 26 years were recruited, and each subject listened to six 1-minute music segments representing different emotions, while speaking at the 40-second mark. The results showed that the differences in EEG signals between various emotional states during speech were more pronounced than those in a quiet state. The study also established an emotion recognition model based on EEG signals under different emotional music stimuli. The classification performance of deep neural networks was compared with other machine learning algorithms in classifying the EEG signals. Using deep neural network algorithms, classification accuracies reached 95.84% for speech and 96.55% for quiet states. The findings suggest that musicinduced emotional stimuli lead to distinct differences in EEG signals during speaking and quiet states, with deep neural networks outperforming other machine learning techniques in EEG classification.