Scholarlink

<p>EEG shows that when a person listens to music, their emotional response is related to how the brain processes music. The electrodes that are placed on the scalp allow researchers to get very good data about a person's brain at different locations. This information can be used to develop customized music therapy techniques according to each individual's specific emotional regulation needs. Features for label-free emotional state inference and unsupervised clustering were developed based on an analysis of power spectral density in the Delta, Theta, Alpha, Beta, and Gamma bands. This objective, data-driven approach will serve as a basis for developing custom applications based on improving mental health in both every day and clinical settings. The EEG data showed stable patterns of activity in different brainwave bands indicating that they had established a good starting point from which to assess changes in emotions and thinking following the music intervention. Focused states (Cluster 0, α/β = 0.98, θ/β ≈ 1.09) and Lightly Stressed states (Cluster 2, α/β ≈ 1.00, θ/β ≈ 1.00) (Clarke et al. in Appl Psychophysiol Biofeedback 44:123-129 (2019), 10.1007/s10484-018-09428-6) dominated computational predictions, although band ratio analysis mostly showed Neutral or Relaxed responses. However, the affective dispersion was wider in the post-music data. Numerous examples were nevertheless classified as either Focused (Cluster 0, α/β ≈ 0.985, θ/β ≈ 1.043) or Lightly Stressed (Cluster 2, α/β ≈ 1.010, θ/β ≈ 1.009) by computational predictions. The band ratios revealed widespread patterns of emotion leading to rich emotional states (i.e., calm &amp; relaxed, excited (α/β ≈ 1.018; θ/β ≈ 1.081) and stressed (α/β ≈ 0.934; θ/β ≈ 1.049)) from an initial and single directed state of focus. This evidenced the effectiveness of using music to shift from an initial and singular focus to a wider array of emotional states and increase an individual’s ability to adapt to change. In conclusion, this study demonstrates that custom-created music and therapeutic interventions derived from neurologic data may be feasible for improving mental-emotional health in therapeutic or everyday settings.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

EEG-based happiness analysis and emotion classification through machine learning using personalized music therapy

  • 期刊论文

  • 发表日期：

    2026年6月23日

查看全文
  • Paromita Das,
  • Biswarup Neogi,
  • Somsubhra Gupta

摘要

EEG shows that when a person listens to music, their emotional response is related to how the brain processes music. The electrodes that are placed on the scalp allow researchers to get very good data about a person's brain at different locations. This information can be used to develop customized music therapy techniques according to each individual's specific emotional regulation needs. Features for label-free emotional state inference and unsupervised clustering were developed based on an analysis of power spectral density in the Delta, Theta, Alpha, Beta, and Gamma bands. This objective, data-driven approach will serve as a basis for developing custom applications based on improving mental health in both every day and clinical settings. The EEG data showed stable patterns of activity in different brainwave bands indicating that they had established a good starting point from which to assess changes in emotions and thinking following the music intervention. Focused states (Cluster 0, α/β = 0.98, θ/β ≈ 1.09) and Lightly Stressed states (Cluster 2, α/β ≈ 1.00, θ/β ≈ 1.00) (Clarke et al. in Appl Psychophysiol Biofeedback 44:123-129 (2019), 10.1007/s10484-018-09428-6) dominated computational predictions, although band ratio analysis mostly showed Neutral or Relaxed responses. However, the affective dispersion was wider in the post-music data. Numerous examples were nevertheless classified as either Focused (Cluster 0, α/β ≈ 0.985, θ/β ≈ 1.043) or Lightly Stressed (Cluster 2, α/β ≈ 1.010, θ/β ≈ 1.009) by computational predictions. The band ratios revealed widespread patterns of emotion leading to rich emotional states (i.e., calm & relaxed, excited (α/β ≈ 1.018; θ/β ≈ 1.081) and stressed (α/β ≈ 0.934; θ/β ≈ 1.049)) from an initial and single directed state of focus. This evidenced the effectiveness of using music to shift from an initial and singular focus to a wider array of emotional states and increase an individual’s ability to adapt to change. In conclusion, this study demonstrates that custom-created music and therapeutic interventions derived from neurologic data may be feasible for improving mental-emotional health in therapeutic or everyday settings.