JSAI2025
May 27 - May 30, 2025Osaka International Convension Center + Online
JSAI2025
May 27 - May 30, 2025Osaka International Convension Center + Online
[3K6-IS-2c-02]Music Therapy for Dialysis Stress Relief and Machine Learning-Based Classification
〇JIYUAN ZHANG1, Ken-ichi Fukui1,6, Tsukasa Kimura1, Noriko Otani3, Tetsuya Ohira2, Masayuki Numao4,5(1. The University of OSAKA, 2. Fukushima Medical University, 3. Tokyo City University, 4. Kyoto Tachibana University, 5. SANKEN, The University of OSAKA, 6. Faculty of Business Data Science, Kansai University)
[[Online]]
Keywords:
Machine Learning,EEG analysis,Music Therapy
This study investigates the effects of brAInMelody (AI-generated music) and normal music on stress reduction in dialysis patients using power spectral density (PSD) and phase synchronization index (PSI) analysis. While brAInMelody has reduced negative emotions in healthy individuals, its impact on dialysis patients is unexamined. The SE-VAE model was used for stress detection and therapy.
The study found that brAInMelody significantly improved valence, indicating reduced stress compared to normal music. PSI analysis showed variations in electrode connectivity, providing insights into neural synchronization under different music types. Additionally, SE-VAE outperformed other models (VAE, LSTM, and Transformer) with the highest AUC score (0.7458±0.0188), demonstrating its effectiveness for stress detection.
In conclusion, brAInMelody can reduce stress in dialysis patients, and SE-VAE proves reliable for stress identification.
The study found that brAInMelody significantly improved valence, indicating reduced stress compared to normal music. PSI analysis showed variations in electrode connectivity, providing insights into neural synchronization under different music types. Additionally, SE-VAE outperformed other models (VAE, LSTM, and Transformer) with the highest AUC score (0.7458±0.0188), demonstrating its effectiveness for stress detection.
In conclusion, brAInMelody can reduce stress in dialysis patients, and SE-VAE proves reliable for stress identification.