Identification of quantifiers related to circadian rhythm variations

Abstract : The aim of this study is to propose quantifiers allowing to characterize the adaptation speed of the circadian rhythm of a living being based on temperature measurements. While the most commonly used model to represent circadian rhythm is the COSINOR model, it is not adapted to analyze changes. An extended COSINOR model is then proposed here in order to take into account temperature disturbances related to the activity of the studied living being. The quantifiers are obtained in a two step process. First, the extended COSINOR parameters are determined with a modified recursive least square algorithm, which is modified in order to compensate for the lack of persistent excitation. The estimates of these parameters are further filtered in order to provide several quantifiers which provide useful information on the adaptation speed of the studied living being. The proposed methodology is evaluated on rats temperature measurements obtained from a laboratory experiment.