A novel approach to calibrate populations of models to data distributions: new electrophysiological insights into atria remodelling

Prof Kevin Burrage ( Department of Computer Science, University of Oxford and Queensland University of Technology, Brisbane, Australia )

The populations of models approach is becoming a powerful modelling approach for quantifying and capturing uncertainty in experimental and observational data. Fundamental to this approach  are two issues: how to sample parameter space effectively and how to calibrate the models. We first give an overview of these issues and then present a new approach based on Sequential Monte Carlo and Simulated Annealing to calibrate a model against the underlying distributions in the biomarker data, as opposed to the standard approach of calibrating against ranges.  We then illustrate these new ideas using the atrial action potential Courtemanche model and a data set of two cohorts of people: sinus rhythm and those with chronic atrial fibrillation.  We attempt to describe any new insights that this approach reveals.