Global Sensitivity Analysis of Left Atrial Electrophysiology Models

Personalisation of cardiac electrophysiology models remains an important challenge for development of digital twins that can be used to guide interventions in the clinic.Calibration of these models from data is a significant challenge, where establishing parameter identifiability is crucial, and global sensitivity analysis (GSA) an important first step.The study aimed to determine the sensitivity of a left atrial electrophysiology model to its parameters and how this sensitivity is influenced by left atrial size and anatomy.We used the monodomain equation with excitability described by the modified Mitchell-Schaeffer (mMS) model.An input dataset of 202 samples of model parameters was generated using Latin hypercube sampling.Each sample was used to run a simulation in each geometry, with 3S1 beats at 800 ms cycle length followed by an S2 beat with a coupling interval of 500 ms.Local activation time (LAT) and action potential duration (APD) were determined across the mesh.These outputs were then used to build Gaussian process emulators of LAT and APD, and used to determine first order and total effect sensitivity indices.LAT and APD were most sensitive to in , out and conductivity.The overall sensitivity indices were similar across geometries.