Parameter estimation in cardiac ionic models

Area of Science:

• Computational biology
• Biophysics
• Cardiac electrophysiology

Background:

• Mathematical models of cardiac electrical activity are crucial for understanding excitable cells.
• The Beeler-Reuter model is a standard for ventricular action potential simulation.
• Parameter estimation is challenging due to the complexity of ionic current kinetics.

Purpose of the Study:

• To investigate parameter estimation for the Beeler-Reuter model.
• To determine if ionic currents can be reconstructed from action potential data.
• To assess the impact of experimental design on parameter identifiability.

Main Methods:

• Utilized the Beeler-Reuter (1977) model with 63 parameters.
• Employed a 'data-clamp' protocol fitting total membrane current to experimental data.
• Used action potential recordings perturbed by pseudo-random injection currents.
• Assessed local parameter identifiability using the reciprocal condition value (1/lambda) of the Hessian.

Main Results:

• Fitting to a single action potential resulted in an over-determined model (1/lambda ≈ 3.6e-14).
• Including 2 perturbed waveforms slightly improved identifiability (1/lambda ≈ 1.4e-10).
• Additional perturbed data enabled accurate reconstruction of all ionic currents.

Conclusions:

• Parameter estimation can reconstruct ionic current kinetics and amplitudes in the Beeler-Reuter model.
• Pseudo-random current injection is a viable experimental design for inferring membrane properties.
• Appropriate experimental design allows inference of underlying membrane currents from transmembrane potential waveforms.