Abstract
BACKGROUND：A 24 -hour periodic regression analysis is often conducted for the humans’ rest-activity rhythm
indexed by activity amount. However, this analysis cannot be said to accurately capture the rest-activity rhythm of
inpatients. OBJECTIVE：We preliminarily developed a new evaluation method to accurately visualize the complex
24 -h rest-activity rhythms of inpatient life. METHODS：Physical activity of one male inpatient was measured. The
periodic regression analysis in a 24 - or 12-h cycle as well as a synthetic 24 - and 12-h cycle were applied to the activity
amount data. RESULTS：A periodic regression analysis with only a 24 -h period resulted in y = 1.19352 + 0.115537
× cos（15・t-183.662）. The multiple contribution ratio（R２）was 0.76. A periodic regression analysis with only a 12-h
period resulted in y = 1.19352 + 0.0444696 × cos（30・t-273.522）. The R２ was 0.11. The results of the synthetic
periodic regression curve combining the 24 - and 12-h periodic components were y = 1.19352 + 0.115537 × cos（15・
t-183.662）+ 0.0444696 × cos（30・t-273.522）. The R２ was 0.87. The synthetic periodic regression curves were able
to capture the subject’s rest-activity rhythms more accurately. CONCLUSION：The synthetic periodic regression
analysis on the activity amount was able to capture more accurately the rest-activity rhythm of the subject. In the
future, it is necessary to examine the reliability and validity of the rest-activity rhythm based on synthetic periodic
regression analysis by increasing the number of subjects.
