高社會焦慮者的注意力偏誤與注意力控制之關係：以眼動作業為例

研究目的與背景：社會焦慮的認知病理模式顯示注意力偏誤對於社會焦慮症的維持和發展扮演重要角色，過往研究顯示高社會焦慮者相較低社會焦慮者，可能會對威脅刺激有促進的注意力投入或注意力脫離困難傾向。近幾年有研究者指出社會焦慮者的注意力控制缺陷可能與注意力偏誤背後的機制有關。根據注意力控制理論，焦慮會減損個體的注意力控制能力，特別是抑制與轉換功能，因此推測社會焦慮者可能存在注意力控制缺陷，進而導致並維持社會焦慮者的注意力偏誤。然而過去研究較少同時探討社會焦慮者的注意力偏誤（注意力過度投入與注意力難以脫離）與注意力控制功能之行為表現（抑制與轉換功能），以及兩者與社會焦慮之間的關聯性。基於上述，本研究的目的主要是透過行為測量，探討社會焦慮者是否存在抑制與轉換的注意力控制功能缺陷，以及注意力過度投入與注意力難以脫離的注意力偏誤現象，並進一步檢驗社會焦慮、注意力控制與注意力偏誤之間的關係。研究方法：本研究共有114名參與者（男46人，女68人；平均年齡為20.44歲，標準差為1.34）完成實驗，其中包含高社會焦慮組40人，低社會焦慮組38人，中間組36人。所有參與者均完成混合反跳視作業、注意力投入-脫離作業，並完成測量包含社會焦慮、特質與情境焦慮、注意力控制與憂鬱等變項的各項自陳量表。研究結果：在混合反跳視作業上，高社會焦慮組在反跳視平均跳視潛伏期和跳視錯誤率上與低社會焦慮組無明顯差異，然而高社會焦慮組的反跳視耗損大於低社會焦慮組。其次，兩組的轉換耗損無顯著差異。注意力投入-脫離作業的結果顯示，高社會焦慮組在面對威脅刺激時，注意力投入與注意力脫離的時間均與低社會焦慮組無顯著差異。此外，所有參與者對於威脅臉孔的注意力投入時間都比高興臉孔短，然而對於威脅臉孔的注意力脫離時間都比高興臉孔長。相關分析的結果顯示，所有參與者的注意力控制量表分數（主觀評估的注意力控制能力）與反跳視耗損和轉換耗損（注意力控制的行為指標）均無顯著相關；反跳視耗損或是轉換耗損與威脅相關的注意力投入和注意力脫離的時間（注意力偏誤指標）都沒有顯著的相關性。迴歸分析的結果顯示，反跳視耗損無法預測對威脅刺激的注意力投入以及脫離時間，轉換耗損也無法預測對威脅刺激的注意力脫離時間；注意力控制相關指標未能調節社會焦慮與注意力偏誤相關指標的關係；注意力偏誤相關指標未能中介注意力控制相關指標與社會焦慮的關係。研究結論：本研究結果顯示，高社會焦慮者相較低社會焦慮者，在注意力抑制效能有缺損，但在轉換功能沒有缺損；其次，高社會焦慮者相較低社會焦慮者，在面對威脅刺激時，並未出現明顯的注意力過度投入或是注意力難以脫離的注意力偏誤。不過，本研究結果也顯示，整體而言，威脅刺激相較於正向刺激，更能吸引參與者的注意力。本研究結果也指出自陳量表所測量的注意力控制能力與行為測量的注意力控制表現之間沒有明顯相關，反映出自陳量表與行為測量所測得的注意力控制構念可能不同。此外，本研究並未發現注意力控制在社會焦慮與注意力偏誤之間的調節效果，也未發現注意力偏誤在注意力控制與社會焦慮之間的中介效果。

關鍵字

社會焦慮；注意力偏誤；投入；脫離；注意力控制；抑制；轉換

並列摘要

Research purpose and background: Cognitive-behavioral models of social anxiety assume that attentional bias plays a critical role in the development and maintenance of social anxiety disorder. Previous studies have shown that high socially anxious individuals compared with low socially anxious individuals, may demonstrate facilitated engagement with and difficulty in disengagement from threat. Recent studies indicated that attentional control deficits in socially anxious individuals may be related to the mechanisms underlying attentional bias. According to attentional control theory, anxiety impairs the performance of attentional control functions, particularly in inhibition and shifting. Therefore, socially anxious individuals are assumed to exhibit attentional control deficits and these deficits may cause and maintain threat-related attentional bias. However, there have been few studies investigating the relationship among attentional bias (facilitated engagement and difficulty in disengagement), attentional control (the inhibition and shifting functions), and social anxiety. To sum up, the present study aimed to use behavioral measures to investigate whether high socially anxious individuals show impairments in attentional control (inhibition and shifting), and show attentional bias for threats (facilitated engagement and difficulty in disengagement). Furthermore, this study aimed to examine the relationship among social anxiety, attentional control, and attentional bias. Method: A total of 114 participants (46 males, 68 females, average age=20.44, SD=1.34), including 40 high socially anxious participants, 38 low socially anxious participants, and 36 participants with an intermediate levels of social anxiety completed the experiment. All participants completed the mixed-antisaccade task, the engagement-disengagement task, and self-reported scales for measuring social anxiety, trait and state anxiety, attentional control, and depression. Results: In the mixed-antisaccade task, high socially anxious participants and low socially anxious participants did not differ in their antisaccade latencies or in their error rates. However, high socially anxious participants had larger antisaccade costs than low socially anxious participants. The two groups did not differ in their latency switch costs. In the engagement-disengagement task, the two groups did not differ in the time to engage attention to threatening faces or in the time to disengage from threatening faces. In addition, the time to engage attention to threatening faces was shorter than happy faces, but the time to disengage attention from threatening faces was longer than happy faces in all participants. Correlation analyses showed that all participants’ scores on attentional control scale (subjective evaluation of their own attentional control ability) was not significantly associated to antisaccade costs and shift costs (behavioral measures of attentional control). Furthermore, participants’ antisaccade costs and shift costs did not significantly correlate with the time to engage to threat and the time to disengage from threat (indics of attentional bias). Regression analyses showed that antisaccade costs did not significantly predict the time to engage to and disengage from threats; shift costs also did not significantly predict the time to disengage from threats. In addition, the results showed that attentional control did not moderate the relationship between social anxiety and attentional bias; attentional bias did not mediate the relationship between attentional control and social anxiety. Conclusions: This study suggests that high socially anxious individuals compared with low socially anxious people, demonstrate diminished efficiency of inhibition function but show no significant impairment of shifting function. High socially anxious individuals show no significant threat-related attentional bias compared with low socially anxious individuals. However, this study also reveals that threat stimuli are more likely to attract attention than positive stimuli for all participants. This study also indicates that there is no significant correlation between self-reported and behavioral measures of attentional control, suggesting that the construct of attentional control measured by self-reported scales may differ from that measured by behavioral tasks. In addition, this study suggests that attentional control does not moderate the relationship between social anxiety and threat-related attentional bias; threat-related attentional bias does not mediate the relationship between attentional control and social anxiety.