Synchronized Memory-Dependent Intracellular Oscillations for a Cell-Bulk ODE-PDE Model in $\mathbb{R}^2$

TL;DR

This paper develops a hybrid asymptotic-numerical approach to analyze synchronized intracellular oscillations in a cell-bulk PDE model, introducing a new integro-differential system and a fast numerical scheme to efficiently study complex oscillatory behaviors.

Contribution

It introduces a novel integro-differential ODE system for cell-bulk dynamics and a fast sum-of-exponentials based numerical scheme for efficient long-term simulations.

Findings

Demonstrates various oscillatory behaviors including synchronization and quorum-sensing.

Provides a computationally efficient algorithm that outperforms traditional PDE solvers.

Validates the reduced model against full PDE simulations with significant speedup.

Abstract

For a cell-bulk ODE-PDE model in $\mathbb{R}^2$, a hybrid asymptotic-numerical theory is developed to provide a new theoretical and computationally efficient approach for studying how oscillatory dynamics associated with spatially segregated dynamically active ``units" or ``cells" are regulated by a PDE bulk diffusion field that is both produced and absorbed by the entire cell population. The study of oscillator synchronization in a PDE diffusion field was one of the initial aims of Yoshiki Kuramoto's foundational work. For this cell-bulk model, strong localized perturbation theory, as extended to a time-dependent setting, is used to derive a new integro-differential ODE system that characterizes intracellular dynamics in a memory-dependent bulk-diffusion field. For this nonlocal reduced system, a novel fast time-marching scheme, relying in part on the \emph{sum-of-exponentials method} to numerically treat convolution integrals, is developed to rapidly and accurately compute numerical solutions to the integro-differential system over long time intervals. For the special case of Sel'kov reaction kinetics, a wide variety of large-scale oscillatory dynamical behavior including phase synchronization, mixed-mode oscillations, and quorum-sensing are illustrated for various ranges of the influx and efflux permeability parameters, the bulk degradation rate and bulk diffusivity, and the specific spatial configuration of cells. Results from our fast algorithm, obtained in under one minute of CPU time on a laptop, are benchmarked against PDE simulations of the cell-bulk model, which are performed with a commercial PDE solver, that have run-times that are orders of magnitude larger.