過動兒行為處理策略人工智慧查詢系統試用體驗初探

A Preliminary Study The Investigation of Artificial Intelligence Query System–Providing Strategies on Children with ADHD Behaviors

本研究目的在建置以行為評估與行為治療模式、適用於國小時期過動兒的人工智慧查詢系統，並初探此系統在實務領域，對聚焦問題行為、提供處理策略、並依據內容教養過動兒的應用分析。本研究之研究團隊依據精神疾病診斷準則、行為評估與行為治療原則，開發人工智慧查詢系統，之後邀請已事先訓練七位兒童臨床心理師試用此系統，並試用此系統於32位過動兒、共進行44次查詢，且於結束後進行試用體驗訪談。本研究之結果顯示人工智慧查詢系統以完整行為理論為架構，提供問題行為查詢流程、症狀表現與策略，可聚焦問題並提供多種策略，可適用於國小學齡過動兒童。且此系統適用於初次會談評估並可增進專業知能，可針對需求，提供個別化處理策略，達到提升照顧者專業知能並初步處理困擾目的。但此系統語用過於專業、內容不易理解，致策略不易直接應用，目前仍無法由過動兒照顧者直接使用。本研究初步討論為人工智慧查詢系統具資訊資料處理優點，可簡易、快速進行行為評估並提供適用策略，有利於增進專業知能與新手學習，亦有重複學習之優點，但缺乏ADHD共病症狀與伴隨困擾之解決策略，且未達照顧者可自學之人工智慧系統目標。

The purpose of this study was to construct an artificial intelligence query system, which is suitable for elementary school children with attention-deficit/hyperactivity disorder （ADHD）based on observed behavioral assessment and behavioral theory. The system was assessed for its ability to analyze problem behavior, provide and advise care-givers on treatment strategies and whether it is applicable for elementary school children with ADHD in a clinical setting. The research team designed this artiflcial intelligence query system based on Diagnostic and Statistical Manual of Mental Disorders, behavioral assessment and behavioral theory. Then, seven clinical child psychologists were invited to test the system and applied it on 32 ADHD children. A total of 44 trials were performed. All participants were interviewed at the end of this study. The results of this study indicates that the artificial intelligence query system was based on behavioural theory, provides a problem behavior checking procedure, symptoms of ADHD and treatment strategies. It can focus on the problems that were identifled and provide a variety of treatment strategies which can be applied to elementary school aged children with ADHD. Also, this system can be used during an intake session to provide an individualized training strategy to deal with and reduce the child’s behavior problem at the initial stage. However, the explanation and description of strategies were too difflcult to understand and translate into the daily life for a child. Conclusions of this study shows that this query system can take advantage of the stableness and quickness of the computer’s function to support individualized problem-strategy checking system. Also, this system can enhance professional knowledge and is good for new-learners, and also has the advantages of repeated learning. However the system lacks the symptoms of ADHD comorbidity and solutions of associated problem behaviors limits its capabilities. It is not able to become an automatic training technology in its current design form.