Reducing complexity and unidentifiability when modelling human atrial cells.

Area of Science:

• Computational biology
• Cardiac electrophysiology
• Mathematical modeling

Background:

• Cardiac models are increasingly complex, especially ion channel gating kinetics.
• Accurate parameter estimation is crucial for personalized medicine applications.
• Understanding uncertainty in model parameters is essential for reliable clinical decision-making.

Purpose of the Study:

• To re-calibrate human atrial cell models using approximate Bayesian computation.
• To quantify uncertainty in ion channel gating kinetics parameter estimates.
• To investigate methods for reducing parameter uncertainty and assess their impact on action potential (AP) models.

Main Methods:

• Applied approximate Bayesian computation (ABC) to re-calibrate gating kinetics of four ion channels in two human atrial cell models.
• Investigated two uncertainty reduction strategies: using more complete datasets and employing less complex model formulations.
• Integrated re-calibrated models into full cell models to evaluate effects on AP.

Main Results:

• Re-calibration provided measures of parameter uncertainty, highlighting potential issues with unique value selection.
• More complete datasets did not eliminate parameter uncertainty.
• A less complex model formulation, especially for the fast sodium current, yielded better data fit, lower uncertainty, and faster computation.

Conclusions:

• Parameter uncertainty remains a challenge in cardiac modeling, even with more comprehensive data.
• Simplified model formulations can effectively reduce uncertainty and improve computational efficiency.
• This work contributes to uncertainty quantification in cardiac modeling for potential clinical applications.