A Parameter Representing Missing Charge Should Be Considered when Calibrating Action Potential Models.

Area of Science:

• Computational electrophysiology
• Biophysics
• Mathematical modeling

Background:

• Computational models are essential for understanding cell membrane electrical potential and action potentials (APs).
• These models, typically differential equations, can be algebraically reformulated using the parameter Γ₀, representing unmodeled charge concentrations.
• Previous research has focused on Γ₀'s long-term effects, with limited exploration of its impact on model fitting and physiological properties.

Purpose of the Study:

• To investigate the influence of the parameter Γ₀ on action potential duration restitution.
• To examine the consequences of incorrect Γ₀ specification during computational model calibration.
• To assess the impact of fitting Γ₀ as a separate parameter on model accuracy.

Main Methods:

• Analysis of computational models of membrane electrical potential.
• Investigation of the parameter Γ₀'s effect on action potential duration restitution.
• Model calibration with varying Γ₀ values, including fitting Γ₀ as a parameter.

Main Results:

• Physiologically plausible concentration ranges yield orders of magnitude differences in Γ₀, leading to divergent model predictions.
• Incorrectly specified Γ₀ during model calibration results in biased parameter estimates.
• Fitting Γ₀ as a separate parameter restores the predictive power of calibrated models.

Conclusions:

• Explicitly defining Γ₀ in computational models is valuable for addressing uncertainty in initial concentrations.
• Understanding and accounting for Γ₀ is crucial for accurate electrophysiology model predictions and calibration.
• This work emphasizes the importance of considering unmodeled charges for robust computational neuroscience.