A Markov chain approximation of switched Fokker-Planck equations for a model of on-off intermittency in the postural control during quiet standing

TL;DR

This paper develops a numerical method using Markov chains and finite element analysis to simulate switched Fokker-Planck equations, modeling postural control dynamics during quiet standing, and validates it against Monte Carlo simulations.

Contribution

It introduces a comprehensive numerical recipe for approximating switched Fokker-Planck equations with Markov chains, extending to cases with boundary discontinuities, and applies it to postural sway analysis.

Findings

Markov chain model accurately simulates postural sway dynamics

Stationary probability density functions vary with model parameters

Validation confirms the numerical method's effectiveness

Abstract

The intermittent on-off switching of feedback control is considered as a major mechanism of postural stabilization during human quiet standing, which can be modeled by switched-type hybrid stochastic delay differential equations with unstable subsystems. Dynamics of the model can also be described by the corresponding switched-type Fokker-Planck (FP) equations. Here, we develop a comprehensive numerical recipe to simulate switched-type FP equations in the case that the probability current is conserved at the switching boundary, as is the case for noise-free models exhibiting C^0-continuity for solutions at the boundary, but in a way extendable to cases with discontinuous jump. Specifically, the FP equations are approximated by a finite state Markov chain model using the finite element method. Then, dynamics of the Markov chain model, including time evolution of probability density function (PDF), stationary PDF, and power spectrum of postural sway are analyzed. We further investigate how the stationary PDF alters as values of important parameters of the model change. Dynamics of the Markov chain model are compared with Monte Carlo-based dynamics of the model, by which the developed numerical recipe is validated. The obtained Markov chain model forms a basis of our future investigations of the intermittent postural control as a Markov decision process.