目的：由於中風患者的高發生率及遠端手功能復健的迫切性，目前雖有研究顯示機器輔助治療能增進中風患者手功能，但仍缺乏針對遠端手功能之機器輔助治療的系統性回顧，來整合相關研究證據與療效程度，因此本文透過系統性回顧，統整機器輔助治療對中風患者手功能復健的相關論文，統整機器輔助治療於中風患者手功能復健的成效，以做為臨床與研究的參考。方法：經由PubMed、Medline、Scopus之國際電子資料庫收尋引擎，鍵入關鍵字：cerebrovascular accident/stroke、robot-assisted/robotics、hand/distal/finger function，搜尋文獻由2010年1月至2017年10月為止，並以機器輔助治療為手功能介入方案，且療效評量須包含上肢的動作表現、動作功能或動作基本能力為篩選條件。結果：共收錄5篇文獻，受試者皆為慢性，中至重度損傷的中風病患，利用機器輔助療法來促進中風患者手功能的恢復，顯示在身體結構層級與活動層級皆有正面療效，但目前證據力層級較高之研究仍較少。結論：機器輔助治療可增進中至重度動作損傷之單側偏癱中風患者的上肢功能以及遠端動作，然而因研究限制，其概化至不同族群之效果有限，但整體而言，中風患者手功能的機器輔助治療的成效是正向的，可為復健帶來新的前景。

Objective: The incidence of stroke high and stroke is among the leading causes of the adult disability including hand function deficits in the world. The recent literatures showed that robot- assisted training can improve on hand functions in patients with stroke. However, systematic review that synthesizes and integrates the effects of robot- assisted training on hand functions relative to control treatment remain limited. This systematic review was to synthesize studies that examined the effects of robot-assisted training on hand functions and provided information on clinical rehabilitation and research. Methods: articles published between January 2010 and October 2017 were retrieved from PubMed, Medline, and Scopus databases. The key words were cerebrovascular accident/stroke, robot-assisted/robotics, and hand/distal/finger function. The inclusion criteria in this study were the inclusion of robot-assisted training as an intervention approach in the empirical study and the assessment had to include upper limb movement performance, motor function and basic movement capability. Results: Six eligible studies were included and analyzed. Patients with moderate to severe motor impairment of upper limb at the chronic stage can benefit from robot-assisted training. In general, the patients who received robot-assisted training had significant improvements not only on the level of body structure and function but also on the aspects of activity of daily living. However, level of evidence of the reviewed studies was low. Conclusion: A robot-assisted training could improve hand functions in stroke patients with moderate to severe motor impairment of upper limb at the chronic stage but generalizability to different populations is limited. To sum up, the results of robot-assisted training on hand functions is positive and it can bring new prospects for hand rehabilitation after stroke in the future.