Sensitivity and uncertainty analysis of cardiac cell models Event as iCalendar

(Seminars)

22 August 2017

4 - 5pm

Venue: Ground floor seminar room (G10)

Location: 70 Symonds St, Auckland Central

An ABI seminar by Dr Richard Clayton, Professor of Computational Physiology, Insigneo Institute of in-silico Medicine and Department of Computer Science,University of Sheffield, UK

Abstract

Physiome models are often biophysically detailed, and are computationally demanding to solve. As a consequence it is not easy to determine how model parameters influence outputs. Neither is it easy to calibrate models against measurements that may be noisy or uncertain. One approach to this problem is to replace the detailed model with an emulator, also known as a surrogate model or meta-model. Gaussian processes are tools developed for statistical machine learning, which can be used to emulate complex models. A particular advantage of this approach is that model parameters can be represented as random variables instead of fixed numbers, providing a way to represent the intrinsic variability of quantities such as ion channel densities. I will discuss recent work on using this approach to dissect and compare models of the human atrial action potential.

About the Speaker

Richard Clayton is Professor of Computational Physiology in the Insigneo Institute of in-silico Medicine and the Department of Computer Science at the University of Sheffield in the UK. He worked from 1990-1998 in hospital Medical Physics, undertaking cardiology research funded by British Heart Foundation (BHF) fellowships. From 1998-2003 he worked at the University of Leeds, working on models of cardiac arrhythmias, again funded by the BHF. He moved to Sheffield in 2003, with promotion to personal chair in 2014. His research interests are multi-scale modelling of electrical activity in the heart, and acquisition and processing of data from patients. His current focus is on sensitivity analysis and parameter inference for cardiac models, to better understand how parameter uncertainty and variability influence model behaviour.