特質焦慮對情緒臉孔處理歷程的影響： 事件關聯電位研究

特質焦慮被認為會影響情緒臉孔的處理。然而，目前行為研究對於焦慮影響是源自知覺因素還是決策因素的結論仍不一致。本研究旨在系統性地探討特質焦慮如何影響情緒臉孔辨識，特別關注其在知覺敏感度（d'）和反應標準（β）上的差異，並透過事件關聯電位（ERP）技術來揭示其背後的神經機制。本實驗招募了 40 名受試者，根據中文版本的狀態-特質焦慮量表（STAI-T）將其分為高特質焦慮組（18 人）和低特質焦慮組（22 人）。受試者完成情緒偵測任務，判斷屏幕上顯示的臉孔是否包含情緒。臉孔刺激包括四種情緒（快樂、悲傷、害怕、生氣）和一種中性表情，並透過 FanTaMorph 5 軟體生成十種不同情緒強度的臉孔（0%、6%、12%、24%、36%、48%、60%、72%、80%、100%）。行為數據紀錄正確率並採用訊號偵測理論（Signal Detection Theory, SDT）分析，同步紀錄 EEG / ERP，聚焦在P200 和晚期正成分（LPP）。研究結果顯示，高特質焦慮組對情緒刺激的判斷閾值（反應標準β）較低，更傾向將臉孔判定為情緒臉孔，而其知覺敏感度（d'）卻低於低焦慮組，且此差異在部分情緒低強度（6% - 24%）時已顯現。此外，ERP分析結果顯示高焦慮組的 P200 及 LPP 成分的振幅顯著高於低焦慮組，表示他們對情緒臉孔的神經反應在早期知覺與持續性注意階段都更強烈。綜合而言，本研究支持特質焦慮在行為和神經層面均影響情緒臉孔的辨識，特別是在模糊情緒刺激下，高特質焦慮者的反應標準較低，導致其更可能將刺激判斷為情緒性臉孔。這些發現有助於理解焦慮如何影響情緒處理，並期望未來在介入式療法中可以針對焦慮者的反應標準進行訓練，以減少對情緒訊號的反應偏誤。

Anxiety has been shown to affect emotional face recognition. However, current behavioral research remains inconsistent regarding whether the impact of anxiety originates from perceptual or decisional factors. This study systematically investigates how trait anxiety affects emotional face recognition, specifically focusing on its differences in perceptual sensitivity (d') and decision bias (response criterion β), and aims to reveal the underlying neural mechanisms through event-related potential (ERP) techniques. For this purpose, 40 participants were recruited and divided into high trait anxiety (n = 18) and low trait anxiety (n = 22) groups based on the Chinese version of the State-Trait Anxiety Inventory (STAI-T). Participants completed an emotion detection task, judging whether faces displayed on the screen contained emotion. The facial stimuli included four emotional expressions (happiness, sadness, fear, anger) and a neutral expression (0% intensity). Using FanTaMorph 5 software, ten graded levels of emotional intensity were generated: 0%, 6%, 12%, 24%, 36%, 48%, 60%, 72%, 80%, and 100%. Behavioral accuracy was analyzed using signal detection theory (SDT), while EEG/ERP recordings focused on the P200 and late positive potential (LPP) components.The results showed that the high trait anxiety group had a lower response criterion β, showing a greater tendency to classify faces as emotional, although their overall perceptual sensitivity (d') was lower than that of the low anxiety group; this difference was already evident at some lower emotional intensities (6% - 24%). Furthermore, ERP analysis revealed that the amplitudes of P200 and LPP components were significantly higher in the high anxiety group, indicating stronger neural responses to emotional faces during both early perception and sustained attention stages. In conclusion, this study supports that trait anxiety influences emotional face recognition at both behavioral and neural levels, especially in the context of ambiguous emotional stimuli, where high trait anxiety individuals exhibit a lower response criterion, leading to a greater likelihood of judging stimuli as emotional. These findings enhance our understanding of how anxiety impacts emotional processing and suggest that future interventional therapies could focus on training the response criteria of anxious individuals to reduce their bias in responding to emotional signals.