目的：充足身體活動量 (sufficient physical activity) 與避免過長靜態時間 (prolonged sedentary time)對諸多健康指標有正向的益處。透過行為的客觀評估工具 (如加速規) 有助於提供身體活動及靜態時間的行為準則更有力的證據。但當前的測量工具仍有許多的差異，特別是活動量閾值的異質性可能會隨著身體活動量、強度及不同目標族群而有所不同，進而使研究證據較難提供更準確的身體活動與靜態時間指引。現行身體活動與靜態行為的客觀測量儀器－三軸加速規為目前較廣泛被使用。然而，三軸加速規儀器的設定程序與資料蒐集過程仍缺乏研究共識，容易限制研究證據的解釋性。方法：有鑑於現行使用三軸加速規之研究多以 ActiGraph (Pensacola, FL, USA) 公司所推出之三軸加速規 (型號：GT3X、GT3X+或 wGT3X-BT 等) 為主，因此本研究回顧三軸加速規使用之相關文獻與系統性文獻回顧將以較廣泛被使用的三軸加速規剖析在身體活動及靜態行為測量之研究程序與方法。結果：本研究回顧相關文獻顯示，在測量前須確認三軸加速規的主要配戴位置為腰部 (髖部)，並針對不同的族群 (學齡前兒童、青少年、成年人與高齡者) 給予相對應的加速規軟體設定內容 (如樣本頻率、epoch 長度)。結論：在研究進行中，應給與配戴者提醒；在資料處理時須使用合格資料篩選標準 (非穿戴時間定義與合格配戴天數) 以及注意不同族群有相對應的強度身體活動與靜態行為之切點。建議相關研究可依據本文彙整之三軸加速規測量方法選擇合適的設定方式，提高研究數據的準確性與可比較性。

Purposes: Sufficient physical activity and avoiding prolonged sedentary time could benefit many healthrelated outcomes. Using an objective assessment of movement behaviors (i.e., accelerometry) would robust the healthy behavioral guidelines of physical activity and sedentary time. However, there is a variety of approaches in measuring physical activity and sedentary time, especially the threshold would likely vary by level and intensity of physical activity or sedentary time and by population subgroups. Among measurements of movement behaviors, an objective measurement, tri-axial accelerometer, is the most frequently used by researchers. However, it remains lack of consensus regarding the processing and collection of accelerometer data which may limit the interpretation of the evidence. Methods: Since commercially available of accelerometers is mainly developed by the ActiGraph (Pensacola, FL, USA), this article reviewed literatures and studies which were related to the application of ActiGraph devices, such as GT3X, GT3X+, and wGT3X-BT. Results: According to previous studies, this review introduces mainstream data collection protocols of tri-axial accelerometer in measuring physical activity and sedentary behavior. Conclusion: Based on previous studies, we conclud that researchers should confirm appropriate placement (i.e., hip) of device before measuring, as well as remind subjects about principle of wearing the monitor during measurement. Additionally, using suitable setting of application (sampling frequency, epoch length), inclusion criteria (definition of non-wear-time and valid days of device wearing) and cut-off point of different intensity of physical activity according to different age groups in data processing are important. We suggest that researchers in related field consider valid data processing in accordance with research purpose and subjects to improve data accuracy.