Uncertainty quantification for cardiac models

The heart is an electromechanical pump, where contraction is initiated and synchronised by a propagating wave of electrical excitation. The function of the heart can be modelled by systems of coupled and nonlinear differential equations, constrained by conservation laws. These models are multi-scale, ranging from representations of electrical activation and force generation in a single cell up to models solved on an anatomical mesh that represent the heart of an individual patient. There are a wide range of potential applications that include assessing the cardiac safety of new drugs, understanding the basic mechanisms of heart rhythm disorders, and providing guidance for clinical procedures on an individual patient. A major obstacle to progress is that the present generation of cardiac models do not take into account the natural variability of real cells and organs, and are difficult to calibrate from the noisy and incomplete images and data that are typically available in the clinical setting.