Circadian cycles: A time-series approach

The extraction of circadian cycles from experimental data can be interpreted as a specific case of time-series or signal analysis, but chronobiology and time-series analysis appear to have developed according to separate paths. Whereas some techniques such as continuous (CWT) and discrete wavelet analysis (DWT) are used frequently in rhythmobiology, other specialized methdos such as digital filters, nonlinear mode decomposition (NMD), singular spectrum analysis (SSA), empirical mode decomposition (EMD), ensemble empirical mode decomposition (EEMD) and complete ensemble empirical model decomposition with adaptive noise (CEEMDAN) have only occasionally been applied. No studies are available that compare the applicability between a wide variety of different methods or for different variables, and this is the purpose of the present contribution. These methods improve the goodness-of-fit of the circadian cycle with respect to the standard approach of cosinor analysis. They have the additional advantage of being able to quantify the day-to-day variability of the circadian parameters of mesor, amplitude, period and acrophase around their average values, with potential clinical applications to distinguish between healthy and unhealthy populations. Finally, the circadian parameters are interpreted within the context of homeostatic regulation with distinctive statistics for regulated and effector variables.