Dynamical System Modeling of Self-Regulated Systems Undergoing Multiple Excitations: First Order Differential Equation Approach. (5th August 2021)

• Record Type: Journal Article
• Authors: Mongin, Denis; Caparros, Adriana Uribe; Gateau, Julien; Gencer, Baris; Alvero-Cruz, Jose Ramon; Cheval, Boris; Cullati, Stéphane; Courvoisier, Delphine S.
• Is Part Of: Multivariate behavioral research. Volume 56:Number 4(2021)
• Page Start: 649
• Page End: 668
• Publication Date: 2021-08-05
• Subjects: Dynamical system modeling -- longitudinal data -- two-step approach -- differential equation -- self-regulated system -- system identification -- homeostasis -- DOREMI; Psychometrics -- Periodicals; Psychology, Experimental -- Periodicals; Psychology, Experimental; Psychometrics; Periodicals; 150.15195
• DOI: 10.1080/00273171.2020.1754155
• Languages: English
• ISSNs: 0027-3171

Abstract

Abstract: This article proposes a dynamical system modeling approach for the analysis of longitudinal data of self-regulated homeostatic systems experiencing multiple excitations. It focuses on the evolution of a signal (e.g., heart rate) before, during, and after excitations taking the system out of its equilibrium (e.g., physical effort during cardiac stress testing). Such approach can be applied to a broad range of outcomes such as physiological processes in medicine and psychosocial processes in social sciences, and it allows to extract simple characteristics of the signal studied. The model is based on a first order linear differential equation with constant coefficients defined by three main parameters corresponding to the initial equilibrium value, the dynamic characteristic time, and the reaction to the excitation. Assuming the presence of interindividual variability (random effects) on these three parameters, we propose a two-step procedure to estimate them. We then compare the results of this analysis to several other estimation procedures in a simulation study that clarifies under which conditions parameters are accurately estimated. Finally, applications of this model are illustrated using cardiology data recorded during effort tests.