*2.5. Statistical Analysis*

The diurnal data are time series data, and single cosine analysis was employed to fit the data. The data was considered as diurnal oscillation by a zero-amplitude test, with a *p*-value less than 0.05. All experiments were assessed by comparing two groups mean scores from control animals and diabetic animals using a Student's *t*-test, plus ANOVA for multiple comparisons by using Prism® statistical software ver. 5 (GraphPad, La Jolla, CA, USA) with differences of *p* value < 0.05 were considered statistically significant. Results are expressed as mean ± SEM.
