DOI: 10.6129/CJP.20120607
中華心理學刊 民101，54卷，4期，527-560
Chinese Journal of Psychology 2012, Vol.54, No.4, 527-560

李囿運（國立成功大學心理學系）；謝淑蘭（國立成功大學心理學系；國立成功大學健康照護科學研究所）；翁嘉英( 國立中正大學心理學系）；孫蒨如（國立政治大學心理學系）；梁庚辰（國立台灣大學心理學系）

摘要

本研究旨在探討不同情緒感受是否引發特定模式的自律神經反應。一百位正常之年輕實驗參與者於實驗中觀賞引發中性、生氣、難過、驚訝、滿足、噁心、害怕與快樂等八種不同情緒的短片，同時接受自律神經系統生理量測，收集其心電圖、皮膚導電反應、皮膚表面溫度與臉部肌電圖等訊號，並於觀看每部影片後填寫情緒自評量表。單因子變異數分析的結果指出：多數情緒引發之自律神經相關生理指標變化在不同情緒下有顯著的差異。本研究進一步利用主成份分析法分析不同生理指標在各主成份下之權重，藉以找出較具代表性的生理指標，並使用六種樣型識別分析法嘗試區隔並分類不同情緒的自律神經反應模式。實驗分析結果發現在大多數的分類法則中，各種情緒皆能藉由篩選出之生理指標的變化正確的被區分。因此本研究的結果支持不同情緒各自具有特定的自律神經反應模式。

關鍵詞：主成份分析、情緒特定性、樣型識別分析法

Patterns of Autonomic Activity Differentiate Emotions Induced by Film Clips

You-Yun Lee（Department of Psychology, National Cheng Kung University）；Shulan Hsieh（Department of Psychology, National Cheng Kung University；Institute of Allied Health Sciences, National Cheng Kung University）；Chia-Ying Weng（Department of Psychology, National Chung Cheng University）；Chien-Ru Sun（Department of Psychology, National Chengchi University）；Keng-Chen Liang（Department of Psychology, National Taiwan University）

Abstract

The present study aimed to investigate whether the response pattern of the autonomic nervous system (ANS) is
emotion specific. One hundred young healthy participants viewed film clips that evoked the following emotions: neutral, anger, sad, surprise, content, disgust, fear and amusement. ANS responses were measured through electrocardiogram, skin conductance, finger temperature, and facial electromyography. Following each film clip, participants were asked to self-report their subjective affect. The results of a one-way analysis of variance indicated that the ANS response change is significantly different among the various emotions. ANS-related variables were further analyzed by principal component analysis. Variables with higher weighting in each principal component were selected as emotional indicators. The emotion-specific ANS response patterning was detected by pattern classification analysis, with 6 different classification algorithms tested. The results indicated that the classification rate is higher than the chance level among most of the classification algorithms. The current findings thus support the hypothesis of emotion-specific ANS response patterning. 

Keywords: emotion specificity, pattern classification analysis, principal component analysis