Fig. 2

Visualization of the operationalized diurnal rest-activity rhythm parameters: mesor, amplitude, acrophase, circadian quotient, dichotomy index, and 24-h autocorrelation coefficient, based on fictitious accelerometer data. A Example of a day the accelerometer was worn by a participant. Activity counts of the participant in each 1-min interval of a 24-h day were available. Based on how these activity counts were distributed across the 24-h day, the best cosine function was fitted according to the following formula: \((t)\hspace{0.17em}=\hspace{0.17em}M\hspace{0.17em}+\hspace{0.17em}Acos (2\uppi t/\mathrm{\rm T}\hspace{0.17em}+\hspace{0.17em}\upphi )\hspace{0.17em}+\hspace{0.17em}e(t)\). In this formula, M is the mesor, A is the amplitude, ϕ is the acrophase, Τ is the period, and e(t) is the error term. For each day the accelerometer was worn, a cosine function was fitted to determine the diurnal rest-activity parameters. Afterwards, these values were averaged across all available accelerometer days. B The dichotomy index was calculated as the percentage of in-bed (I) activity counts, expressed per 1-min interval, that were less than the median of out-of-bed (O) activity counts. The dichotomy index was calculated for each day the accelerometer was worn, and these values were averaged across all available days. C Overlap is shown between two different fictitious days the accelerometer was worn. The 24-h autocorrelation coefficient describes the average correlation between activity counts in 1-min intervals separated by 24 h. For the 24-h autocorrelation coefficient only one value was calculated, as the autocorrelation across all available accelerometer days. In this example, only day 1 and 2 are shown; however, activity counts separated by 24 h between day 2 and day 3, day 3 and 4 etc., are also included in the calculation