音樂情緒及其腦電波頻譜動態之探討

Exploring EEG Spectral Dynamics of Music-Induced Emotions

葉人慈(Ren-Cih Ye)

指導教授 ： 林進燈 ； 黃志方

國立交通大學/工學院/聲音與音樂創意科技碩士學位學程/碩士(2011年)

https://doi.org/10.6842/NCTU.2011.00945

摘要

本研究使用Vieillard等人所設計出的56個可以引發情緒的音樂片段，分析其MIDI音樂以得到音樂特徵；在實驗中，請17位受測者聆聽這些音樂並為自己聽音樂的情緒作二個維度的評量：正負向（valence）與強弱向（arousal）；同時並量測受測者的腦電波（Electroencephalography），以探討聽音樂時的腦電波頻譜動態現象。本研究中使用Gomez與Danuser的迴歸方式，歸納五個音樂特徵與受測者的自我情緒評量之間的關係。音樂特徵分析的結果顯示，不同音樂特徵與聽者的情緒存在某些相關性，例如大調的音樂使人有正向感覺、小調的音樂則令人感到低落；音樂速度的快慢也影響聽者情緒的激昂或平靜。但較複雜的音樂特徵，例如節奏、旋律變化甚至和聲組成，則較難用簡單的音樂計算得到有效的情緒音樂特徵。本研究將聽音樂時的腦電波分為5個頻帶delta (1~3 Hz), theta (4~7 Hz), alpha (8~12 Hz), beta (13~30 Hz), 及gamma (31~50 Hz)，並先以獨立成份分析(ICA)分出不同訊號源，再依訊源特徵對訊號分群，所得的腦波訊號大小（EEG power）再經ANOVA分析，檢驗特定腦區頻帶的作用是否與受測者的情緒正負向或情緒強弱有關係。研究發現，大腦的前額葉，特別是兩側腦半球現象的不對稱性，在左邊alpha及右邊gamma頻帶都分別與情緒的正負向相關；而在額葉中線的alpha、左側軀體運動區的delta以及枕葉區的theta頻帶都分別發現與情緒強度相關的現象。由本研究可再次證實，大腦兩邊的不對稱現象是情緒正負趨向的指標；而大腦在許多區域都對聽音樂引發的情緒有反應。

關鍵字

情緒 ； 音樂特徵 ； 情緒正負向 ； 情緒強弱向 ； 腦電波 ； 獨立成份分析 ； 頻譜動態 ； 不對稱性

並列摘要

Vieillard et al. (2008) has designed 56 emotional musical excerpts that convey four different emotions. The MIDI formats of the musical excerpts were analyzed to obtain their musical features. In addition, 17 participants were invited to listen to the musical excerpts and to assess their emotions by two evaluative dimensions: positive/negative valence and high/low arousal. The regression model developed by Gomez and Danuser was used to discover the relationships between the assessments by the participants and the five musical features. The result of the musical feature analysis indicated a correlation between musical features and the emotions of the listener. For example, major musical keys created positive feelings, minor musical keys created depressed feelings, and music tempo made the participants feel excited or peaceful. Rhythm, melody, and harmony (i.e. more complex musical features) were rather difficult to calculate with simple musical features to determine the relationship of complex musical features with emotional state. In part two of the study, the participants’ electroencephalography (EEG) was measured while they listened to musical excerpts to investigate their EEG spectral dynamics. The five frequency bands for EEG were delta (1~3 Hz), theta (4~7 Hz), alpha (8~12 Hz), beta (13~30 Hz), and gamma (14~50 Hz). First, the Independent Component Analysis (ICA) was completed to distinguish the source of different signals. Then, all signals were classified based on the feature of the sources. The obtained EEG power of the signals was analyzed with ANOVA to determine if frequency bands of specific brain regions were related to participants’ positive/negative valences and high/low arousals. The results demonstrated that the frontal lobe was characterized by hemispheric asymmetry with the left-frontal alpha and the right-frontal gamma indicating a relationship with positive/negative valences. Furthermore, the frontal midline alpha, the left somatomotor delta, and the occipital theta were found to be related to high/low arousal. This study reaffirmed the phenomenon of hemispheric asymmetry as a good indicator for positive/negative valences. Furthermore, responses of many brain regions have an observable relationship with music-induced emotions.

並列關鍵字

emotion ； musical feature ； valence ； arousal ； Electroencelography (EEG) ； Independent Component Analysis (ICA) ； spectral dynamic ； asymmetry

參考文獻

[1] P. N. Juslin and P. Laukka, "Expression, perception, and induction of musical emotions: A review and a questionnaire study of everyday listening," Journal of New Music Research, vol. 33, pp. 217-238, Sep 2004.

[2] P. N. Juslin and D. Vastfjall, "Emotional responses to music: the need to consider underlying mechanisms," Behavioral and Brain Sciences, vol. 31, pp. 559-75; discussion 575-621, Oct 2008.

[4] G. A. and J. P, "Emotional expression in music," in Handbook of affective sciences, R. Davidson, K. R. Scherer, and H. H. Goldsmith, Eds., ed New York: Oxford University Press, 2003, pp. 503-534.

[5] Y. H. Yang and H. H. Chen, "Music Emotion Ranking," 2009 Ieee International Conference on Acoustics, Speech, and Signal Processing, Vols 1- 8, Proceedings, pp. 1657-1660, 2009.

[6] L. Lu, D. Liu, and H. J. Zhang, "Automatic mood detection and tracking of music audio signals," Ieee Transactions on Audio Speech and Language Processing, vol. 14, pp. 5-18, Jan 2006.