前言：關節鬆動術學習難度頗高，然過去研究多著眼於施力準確度之掌握，很少人運用認知想 像介入技巧來促進學習。目的：檢定強化專注力回饋技巧（attention-reinforcement feedback, ARF) 於肩關節鬆動術學習之成效。方法：本研究延攬自願者38人，參與肩關節鬆動術課程（含8項技 巧）。每位參與者均需隨機選擇4項技巧介入ARF學習模式，並選擇其餘4項技巧接受傳統教學。 各項技巧隨機分為有介入與無介入ARF兩組（各組19筆資料），最後於學習結束一週後接受技術 檢定並填寫學習滿意度問卷。結果：ARF組於肱骨頭下滑關節鬆動技巧之成績顯著優於非ARF組 (^=0.019)'其中又以於活動末段位置進行者成效較佳。介入ARF相對於無ARF組，落在低分區的 人次少5人，落在高分區的人次多8人。學習滿意度調查顯示所有學習者均偏好ARF學習法，90% 以上認為於傳統教學中介入ARF模式有助於增加學習專心程度與降低學習難度。結論：ARF回饋 學習法可應用於操作治療等技術性課程，藉由強化學習者於練習過程之敏感度與自覺能力，有助 於提昇學習成效。（物理治療2012;37(4):338-346 )

Background: The effectiveness of knowledge of results (KR) applied to joint manipulative technique (JMT) learning was examined mainly focusing on the precision of force level with a kinetic simulator or performance feedback. Attention-gated reinforcement learning (ARL) has been proved to improve motor skill acquisition but to date its effect has not been investigated in learning JMT. Purpose: We examined the effectiveness of ARL in combination with providing quantitative KR (attention-reinforcement feedback, ARF) on learning JMT. Methods: Thirty-eight physiotherapy college students were recruited for learning shoulder JMT. In this study JMT included 2 levels (basic and advanced) of 4 techniques namely 8 skills in total. For each skill all students were randomly assigned to ARF or traditional learning (TL) groups. ARF referred to a survey analysis of immediate student recall after each skill performed by his partner using a visual analog scale (VAS) as quantitative KR to respond the force level or direction, hand placement, positioning, or comfort. The outcome of learning was evaluated objectively with a 4-point scoring system by 2 senior physiotherapists and subjectively with a learning satisfaction questionnaire. Results: ARF group exhibited significantly higher learning score than TL group when performing advanced inferior glide JMT skill (/?=0.019). The results of homogeneity tests further demonstrated the effectiveness of ARF in learning inferior glide JMT (%2=6.23， p<0.05), especially for the occurrence rate of highest performance score (ARF: 11/38, TL:
3/38). Overall satisfaction for ARF up to 97.4% was also found and more than 90% of students considered ARF as a useful augmentation to improve concentration and reduce difficulties during learning. Conclusions: The current study has provided a convenient and innovated method to apply both ARL and KR in manipulative techniques teaching. Most of all, the learners have played an active instead of a passive role during this learning process, and the learning was considered to be fun with higher satisfaction. (FJPT 2012;37(4):338-346)