Key result
A proposed method based on singular value decomposition demonstrated that small singular values signify a lack of parameter uniqueness in cardiac action potential models.
Why the study?
It was unclear whether maximum conductances in cardiac action potential models can change without appreciable changes in the action potential, potentially undermining parameter uniqueness.
A singular value decomposition method can quantify the identifiability of maximum conductances in cardiac action potential models, highlighting potential issues with parameter uniqueness when evaluating drug effects.
Answers are grounded in this paper and cite their sources.
Method enables checks on conductance parameterization uniqueness in cardiac models; leaves open reliability for clinical simulations.
Mathematical models describing the dynamics of the cardiac action potential are of great value for understanding how changes to the system can disrupt the normal electrical activity of cells and tissue in the heart. However, to represent specific data, these models must be parameterized, and adjustment of the maximum conductances of the individual contributing ionic currents is a commonly used method. Here, we present a method for investigating the uniqueness of such resulting parameterizations. Our key question is: Can the maximum conductances of a model be changed without giving any appreciable changes in the action potential? If so, the model parameters are not unique and this poses a major problem in using the models to identify changes in parameters from data, for instance, to evaluate potential drug effects. We propose a method for evaluating this uniqueness, founded on the singular value decomposition of a matrix consisting of the individual ionic currents. Small singular values of this matrix signify lack of parameter uniqueness and we show that the conclusion from linear analysis of the matrix carries over to provide insight into the uniqueness of the parameters in the nonlinear case. Using numerical experiments, we quantify the identifiability of the maximum conductances of well-known models of the cardiac action potential. Furthermore, we show how the identifiability depends on the time step used in the observation of the currents, how the application of drugs may change identifiability, and, finally, how the stimulation protocol can be used to improve the identifiability of a model.
No takes yet. Share an insight, caveat, or question.
Jæger et al. (2019) studied Cardiac action potential modeling. Singular value decomposition method was evaluated on Identifiability of maximum conductances. A proposed method based on singular value decomposition demonstrated that small singular values signify a lack of parameter uniqueness in cardiac action potential models.
Synapse has enriched 10 related papers. Consider them for comparative context: