以眼動儀探討罹患自閉症類群障礙症 之兒童對自然情境圖片中社會訊息之 凝視型態：ASD 自然情境圖片眼動研究

Investigating Gaze Pattern of Children with Autism Spectrum Disorder toward Social Information in Naturalistic Pictures with Eye Tracker

本研究探討自閉症類群障礙症（Autism Spectrum Disorder， ASD）患者幼年 時期社會認知缺陷本質。針對部分研究結果支持ASD 患者較少觀看人物眼睛， 仍有爭議，例如：ASD 患者於動／靜態刺激的表現尚未有清楚穩定的結果。此 外，眾多研究將焦點放在ASD 成人患者。本研究使用眼動儀探索ASD 患童在 靜態高生態效度的社會刺激圖片上之凝視型態，共招募25 位自閉症類群障礙症 ASD 患童為實驗組，25 位配對性別、年齡及智力的典型發展兒童為控制組。操 弄變項為圖片人數（單人、雙人）及圖片類型（社交、非社交），依變項為注 視時間及第一凝視點。以注視時間為依變項所發現的結果，除了與過去研究大 致相符外（ASD 患童在人物、頭部及眼睛區域之注視時間顯著低於典型發展組， 於非人物區域及嘴巴區域顯著高於典型發展組），亦發現ASD 患童的社會注意 力缺陷或許不在於對社會訊息的凝視組型，而在於對社會訊息凝視偏好的程度。 更重要的是，以第一凝視點為依變項所發現的結果顯示，第一凝視點是呈現出 組間差異之敏感指標；第一凝視點分析結果顯示ASD 患童在人物、頭部及眼睛 區域的凝視優先性顯著低於典型發展組，而於嘴巴及頭部其他區域則高於典型 發展組。本研究貢獻在於嘗試以較具外在效度、具自然情境脈絡的取向來探索 ASD 患者之凝視行為，並採用更敏感的眼動指標，即第一凝視點。同時，本研 究以學齡兒童為受試者，故能探索更早期的凝視型態。再者，本研究同時考量 實驗素材之人數與互動性，亦控制人物的臉部表情，故能有較細緻的結論。

Purpose: The topic of social cognition is essential to exploring the nature of autism spectrum disorder (ASD). Compared with typically developing participants， individuals with ASD tend to gaze at a human’s mouth more than his/her eyes. However， this conclusion is controversial， especially when different processing mechanisms related to static and dynamic task materials are considered. Methods: To resolve this controversy， eye-tracking techniques were applied to investigate the gaze pattern of children with ASD， and a precise and sensitive indicator， namely the first fixation point， was drawn from the observed eye movements. Both fixation time and first fixation point were measured as the participants were viewing ecologically valid social information pictures. In total， 25 children with ASD (the experimental group， EG) and 25 typically developing children (the control group， CG) were recruited. The two groups were matched by gender， age， and IQ. Two within-participants independent variables were used， namely the number of people in the materials (one or two people) and the type of action the people exhibited (social or nonsocial); thus， there were four types of materials， and each comprised five pictures. Results/Findings: Most of the fixation time results are consistent with those in the literature. Specifically， the total fixation time of the EG was significantly shorter than that of the CG in the people， head and eyes areas and was significantly longer in the non-people and the mouth areas. Moreover， we found out that the deficit of social interaction of children with ASD might not be related to the pattern but to the intensity of their interest to those critical areas. Notably， our findings based on the first fixation points indicate that the gaze priority of the people， head， and eye areas of the EG was significantly lower than that of the CG， and also that the gaze priority of the mouth and the areas of the other parts of the head of the EG was significantly higher than that of the CG. These findings also suggest that the cause of the impairments in social attention of children with ASD may not be the overall gaze pattern but the intensity of preference toward social information areas. Our findings also support the first fixation point as a precise and sensitive indicator to detect the differences between the two groups. Conclusions/Implications: The main contributions of this study include (1) using ecologically valid materials to increase the degree of external validity， (2) measuring the first fixation point and proving it is a sensitive dependent variable， (3) recruiting children as participants to investigate their developmental pattern， and (4) controlling the number of people and their social interaction nature and facial expression in the materials. Accordingly， we conclude that the stable developmental differences of the gaze pattern can be reliably observed between the two groups of children.