Research Progress on Sleep Screening Methods Using Actigraph(GT3X/±Series) Accelerometer

In the study of 24 h activity behaviors based on accelerometry, it is essential to distinguish accelerometer data into wake periods（for analyzing physical activity and sedentary behavior） and sleep periods（for examining sleep duration and quality） before calculating different activity behavior times. However, there is no widely accepted standard method for separating wake-sleep periods in accelerometer data, and a systematic review of these methods is lacking. This study identifies nine wake-sleep separation methods applicable under free-living conditions from 425 relevant papers. By elucidating the principles, advantages, and limitations of these methods, the study provides recommendations for researchers using GT3X/± accelerometers: 1) For selecting automatic sleep screening algorithms, wrist-worn devices should use the sleep diary/HDCZA algorithm built into GGIR combined with the van Hees 2015 algorithm, while waist-worn devices can use the Tracy ProChildren, Tudor-Locke, McVeigh, or Bellettiere-McVeigh algorithms; 2) when using Actilife for sleep screening, the Sadeh algorithm is recommended for individuals under 30 years old, and the Cole-Kripke algorithm for those aged 30 and above, both requiring the support of a sleep diary. This study provides data processing recommendations for 24 h activity behavior research and offers direction for developing new wake-sleep separation methods in the future. Nonetheless, more research is needed to compare and validate these sleep screening methods to promote the standardization of research methods.