研究目的：欲瞭解自閉症類疾患兒童，橫跨幼兒期、學齡前期至學齡期的自閉症狀發展軌跡，探究當中是否存有不同的發展軌跡，並檢視不同發展軌跡的群體在認知功能與適應行為隨時間的變化型態是否有差異，以及不同群體在早期的社會溝通能力或療育經驗上是否有差異。 方法：本研究有37名自閉症類與24位非自閉症類受試者納入分析，共測量三次：時間點一平均生理年齡為2歲半、時間點二為4歲、時間點三為8歲半。於三個時間點皆以自閉症診斷觀察量表評估自閉症狀嚴重度、以及穆林發展量表或魏氏兒童智力量表第四版評估認知功能，在時間點二與時間點三以文蘭適應行為量表第二版評估適應行為；另外，在時間點一以台灣版兩歲期自閉症篩檢工具評估社會溝通能力，以及在時間點一至時間點二之間，以電話訪問的方式紀錄兒童每週參與療育的時數。 結果：由階層式集群分析結果顯示，社交情感嚴重度校正分數可將自閉症類與非自閉症類兒童分群，並於自閉症類兒童中可再分為維持高嚴重度組與退步組。兩組自閉症類兒童於學齡時期在社交情感症狀皆呈現嚴重度明顯上升的趨勢，在認知功能與適應行為隨時間的變化型態相似，不過維持高嚴重度組在認知功能與適應行為有較低的分數（但未達統計顯著）。此外，兩組自閉症類兒童於時間點一的意圖溝通能力即有差異，不過僅共享式注意力可預測日後社交情感症狀發展軌跡之分群。 總結：由結果可見自閉症狀維持高嚴重度組，於認知功能與適應行為有較明顯缺損，推論自閉症類兒童的自閉症狀嚴重度與認知功能及適應行為並非完全獨立的變項。兩組自閉症類兒童於學齡期症狀嚴重度上升，而此趨勢在退步組兒童更為明顯。雖然本研究並未發現早期療育經驗對於社交情感症狀發展的影響，不過共享式注意力對於日後社交情感症狀有預測力。進一步討論此結果在理論與臨床實務上的應用。 / Purposes: the purpose of the study was to plot longitudinal developmental trajectories of autism symptom severity in the children with autism spectrum disorders (ASDs) from toddler age to school age. And to examine whether these different trajectories are associated with the cognitive function, adaptive behavior, early social communication skill, and early intervention experience. Methods: Thirty-seven children with ASDs and 24 children with developmental delay participated at time 1 (mean chronological age was 2.5 years old), and then followed at time 2 (mean chronological age was 4 years old) and time 3 (mean chronological age was 8.5 years old). The study assessed the autism symptom severity by ADOS and cognitive function by MSEL or WISC-Ⅳ at three time points; adaptive behavior by VABS-II at time 2 and time 3; and early social communication skills by T-STAT at time 1 and also used telephone interviews to record participants’ weekly intervention hours during time 1 to time 2. Results: In hierarchical cluster analysis, social affect calibrated severity score (CSS) could discriminate ASDs and Non-ASDs groups, also could divided ASDs into two groups, called persistent high and worsening groups. The social affect symptom severity increased at school age in both groups. On the other hand, the change of cognitive function and adaptive behavior with time showed similar pattern in the two groups. However, participants in the persistent high group have relatively lower scores of cognitive function and adaptive behavior, although did not rearch statistically significant. Furthermore, two ASDs groups at toddler showed difference in intention communication skills; however, only joint attention could predict social affect trajectory subgrouping. Conclusions: In summery, the study showed that persistent high group has more cognitive and adaptive function impairments than the worsening group in the children with ASDs. It seemed that the autistic symptom are not entirely independent with cognitive function and adaptive behavior. Moreover, the symptom severity increased at school age in both the persistent high and the worsening groups, and this trend was more significant in the worsening group. Additionally, the experience of early intervention showed no significant effects on developmental trajectories of social affect symptom; however, the joint attention was found to be an index to predict the development of social affect symptom in ASDs groups. The theoretical and clinical implications were discussed.