Estimability Analysis and Optimal Design in Dynamic Multi-scale Models of Cardiac Electrophysiology Poster.pdf
pdf
Introduction: Multi-scale methods are used to study phenomena that occur over multiple temporal
or spatial scales and observations at different scales often require distinct experimental protocols. A problem in quantitative physiology is to develop the methodology for simultaneous modeling and assessment of data that arise from multiple, multi-scale experimental frameworks.
Materials and Methods: We present an applied approach to optimal experimental design and estimability analysis for mechanistic models of cardiac electrophysiology, by extending and improving on
existing computational and graphical methods. We used numerical approaches including the sensitivity matrix to study the identifiability and estimability of a novel multi-scale model of the cardiac action potential provided that information regarding the single cell and propagating action potential are known and measurable.
Results and Discussion: We demonstrate that the model is both identifiable and estimable. In light of our analysis, we identified two model modifications that improve parameter estimability, and show that the choice of optimality criterion has a profound effect on the influence of each experimental protocol.
Submitted by David Corman
on
Introduction: Multi-scale methods are used to study phenomena that occur over multiple temporal
or spatial scales and observations at different scales often require distinct experimental protocols. A problem in quantitative physiology is to develop the methodology for simultaneous modeling and assessment of data that arise from multiple, multi-scale experimental frameworks.
Materials and Methods: We present an applied approach to optimal experimental design and estimability analysis for mechanistic models of cardiac electrophysiology, by extending and improving on
existing computational and graphical methods. We used numerical approaches including the sensitivity matrix to study the identifiability and estimability of a novel multi-scale model of the cardiac action potential provided that information regarding the single cell and propagating action potential are known and measurable.
Results and Discussion: We demonstrate that the model is both identifiable and estimable. In light of our analysis, we identified two model modifications that improve parameter estimability, and show that the choice of optimality criterion has a profound effect on the influence of each experimental protocol.