An improved numerical method for strong coupling of excitation and contraction models in the heart

Quantifying the interactions between excitation and contraction is fundamental to furthering our understanding of cardiac physiology. To date simulating these effects in strongly coupled excitation and contraction tissue models has proved computationally challenging. This is in part due to the numerical methods implemented to maintain numerical stability in previous simulations, which produced computationally intensive problems. In this study, we analytically identify and quantify the velocity and length dependent sources of instability in the current established coupling method and propose a new method which addresses these issues. Specifically, we account for the length and velocity dependence of active tension within the finite deformation equations, such that the active tension is updated at each intermediate Newton iteration, within the mechanics solution step. We then demonstrate that the model is stable and converges in a three-dimensional rod under isometric contraction. Subsequently, we show that the coupling method can produce stable solutions in a cube with many of the attributes present in the heart, including asymmetrical activation, an inhomogeneous fibre field and a nonlinear constitutive law. The results show no instabilities and quantify the error introduced by discrete length updates. This validates our proposed coupling framework, demonstrating significant improvement in the stability of excitation and contraction simulations.