兒童情緒臉孔辨識發展與社會適應之探討

目的：臉部情緒辨識（Facial emotion recognition）能力在兒童社會適應方面至關重要。同儕是社會化的重要媒介之一，但較少有研究使用兒童的臉部表情作為刺激探討個體情緒辨識能力。本研究旨在採用自編的華人兒童臉孔情緒來探討5-12歲兒童臉部情緒辨識與社會適應行為（Adaptive social behaviour）之關聯性。方法：研究對象為90名就讀普通班的5-12歲兒童，分為三個組別「幼兒園組」（n=30；Male 14；Meanage=5.53歲；SD=0.32）、「低年級組」（n=30；Male 13；Meanage=7.48歲；SD=0.40）和「中高年級組」（n=30；Male 16；Meanage=10.07歲；SD=1.03）。採用修訂畢保德圖晝詞彙測驗（Peabody Picture Vocabulary Test-Revised Mandarin version，簡稱PPVT-R）量測接收性語言能力，皆高於80。由家長填寫「適應行為評量系統第二版中文版（Adaptive Behaviour Assessment System–II Taiwanese version，簡稱ABAS-II）」蒐集兒童社會適應行為表現，包括休閒和社交領域，再計算休閒技能量表分數（Leisure skill scaled score，簡稱LIESS）與社交技能量表數分（Social skill scaled score，簡稱SOCSS）並換算為社會知能組合分數（Social composite score，簡稱SOCS）。另採用自編的臉部情緒辨識測驗蒐集兒童的臉部情緒辨識能力，題項共50題，再計算各情緒類型的命中率（Hit rate，簡稱HR）、誤警率（False alarm rate，簡稱FA）、敏感度（Sensitivity 亦即d'）和反應偏向（Response bias亦即C）。結果：（1）以HR、FA、d' 與C的情緒辨識變項為依變項，並以混合設計5情緒（開心、生氣、害怕、難過和厭惡）x 3相似程度（高、中和低水準）x 3年級（幼兒園、低年級和中高年級組）三因子變異數分析考驗兒童臉部情緒辨識能力的發展， 結果表示無三因子交互作用關係存在，但發現情緒與相似程度二因子交互作用，以及具年級主要效果；（2）以年級為獨變項，LIESS、SOCSS與SOCS為依變項進行單因子變異數分析，並以Bonferroni法進行事後比較來考驗兒童社會適應行為的發展；（3）再以LIESS、SOCSS與SOCS為依變項，HR、FA、d' 與C為預測變項並控制年級、性別和接收性語言進行多元迴歸分析，來檢視情緒辨識變項是否能預測兒童的適應行為。結論：情緒類型、相似程度和年級都影響兒童臉部情緒辨識。年級對兒童的休閒適應技能有影響，幼兒園組的表現比中高年級組優。兒童的臉部情緒辨識能力與其適應性社會行為存在顯著關聯性。中等相似度生氣的命中率下降，與中等相似度生氣和低等相似度開心的反應偏差下降，均影響兒童的休閒適應技能。此外，兒童的整體社會適應能力會因中等相似度生氣的命中率和敏感度降低而提高。

Purpose: Facial emotion recognition in children is critical to adaptive social functioning. Peers are one of the primary mediators of socialization, but few studies have used child emotional face stimuli to examine children's emotion recognition abilities. This study examined the relationship between facial emotion recognition and adaptive social behaviour in children aged 5 to 12 by using self-edited faces of Asian children.Method: Experiments were conducted with ninety children aged 5 to 12, and were divided into three groups: Kindergarten (n=30; Male 14; Meanage=5.53; SD=0.32), Lower grade (n=30; Male 13; Meanage=7.48; SD=0.40), and Middle or upper grade (n=30; Male 16; Meanage=10.07; SD=1.03). The Mandarin version of the Peabody Picture Vocabulary measured the participants' receptive language skills, with the standard scores all more than 80. Parents completed the Taiwanese version of the Adaptive Behaviour Assessment System-II forms to evaluate participants' adaptive skills in the social domain, including leisure and social skills. Leisure skill area scaled score (LIESS) and social skill area scaled score (SOCSS) were combined to obtain the social composite score (SOCS). In addition, a Self-edited Face Emotion Recognition Test was used to measure the participants' ability to identify five emotions in facial expressions. A total of fifty items were assessed to calculate the scores of hit rate (HR), false alarm rate (FA), sensitivity (d'), and response bias (C) for each emotion type. Results: (1) Four dependent emotion recognition variables, scores of HR, FA, d' and C, were entered separately into a three-way mixed ANOVA. To examine the development of children's facial emotion recognition ability, we used 5 emotion (happiness, anger, fear, sadness, and disgust) x 3 similarity (high, moderate, and low) x 3 grade (kindergarten, lower grade, middle or upper grade). The results indicated that no three-way interaction existed, but did show a two-way interaction of emotion type and similarity level, and a main effect of grade level on variables (2) Three dependent variables, LIESS, SOCSS, and SOCS, were entered separately into one-way ANOVA with grade levels as independent variables. Follow-up pairwise comparisons were done for investigating the effect of grade level on each adaptive skill dimension. (3) Multivariate regression analyses were conducted using LIESS, SOCSS, and SOCS as dependent variables, HR, FA, d' and C as predictor variables, and controlling for grade, gender, and receptive language skills to investigate the role of emotion recognition variables in predicting social adaptive function.Conclusions: Emotion type, similarity, and grade level all affected children's facial emotion recognition. Grade had an effect on children’s leisure adaptive skills, and kindergarten students performed better than middle upper students. There is a significant association between children's facial emotion recognition ability and adaptive social behaviour. A decrease in moderate-similarity anger face of hit rate, along with an increase in moderate-similarity anger face, and a decrease in low-similarity happy face of response bias, influenced children's leisure adaptive skills. Additionally, it was found that a decrease in both hit rate and sensitivity of moderate-similarity anger faces would enhance children's social adaptation.