This study utilizes the ERP model of aesthetic cognition to reveal the neural basis of aesthetic experience, pointing out the complex cognitive processes including emotional escalation. Advanced techniques are used to pre-process violin sound signals and multiple algorithms are applied to extract critical features for accurate emotion recognition. The developed DBM-LSTM model combines the advantages of multi-layer stacked RBM and LSTM, which substantially improves emotion recognition accuracy with a success rate of more than 0.85. Finally, through a questionnaire survey, it is found that the public’s aesthetic preference of violin performance is significantly correlated with emotion perception, indicating that the more the performance conforms to the public’s aesthetics, the more it can stimulate positive emotions. This study not only provides theoretical and practical guidance on the combination of violin performance art and information technology, but also opens up a new way for the emotional communication of artistic performances and enhancing public aesthetic experience.