Table 1.

Different types of evidence relevant to the credibility of a cardiac EP model, with ion channel, cell, and organ-level examples.

Category	Type of credibility evidence	Examples
		Ion channel	Cell model	Organ-level model
Category 1	Evidence regarding validity of model assumptions or supporting the model formulation	Successes of Hodgkin-Huxley formulation for modeling ion channels—see section Ion channel models	Evidence supporting the formulation of cell membrane as a parallel resistor-capacitor electric circuit	The successes of the bidomain equations, in particular predictions made that were later experimentally observed—see section Organ-level models
	Evidence regarding accuracy/fidelity of model parameters/inputs	Evidence supporting accuracy of steady-state inactivation parameters—see section Ion Channel Models	Rationale behind standard choice of membrane capacitance equal to 1 uF/cm2.	Evidence on fidelity of geometry used and on fidelity of fiber/sheet specification—discussed in section Organ-Level Models.
Category 2	Calibration results	Results showing agreement between ion channel model and experimentally recorded current-voltage relationship when ion channel parameters are calibrated using this data	Results showing agreement between the model action potential and experimental recordings when maximal conductances are tuned to achieve the match	Results showing activation patterns match experiment if fast sodium current maximal conductance (which controls conduction velocity) chosen to maximize agreement
Category 3	Reproduced (emergent) phenomena	Simulation results demonstrating that a rapid sodium current model can exhibit damped oscillations	Simulation results demonstrating that a cell model reproduces action potential spike and dome morphology	Simulation results demonstrating that ECG predicted by a heart and torso model exhibits realistic-looking QRS complex and T wave
	General validation results	Comparison of a general-purpose ion channel model predictions to new voltage-clamp data not used in the construction of the model.	Comparisons of model results with experimental data for a novel general-purpose cell model, e.g., all such results in O'Hara et al. (2011). Discussed in detail in section Cell Models	Excitation patterns of general purpose bi-ventricular model compared to experimental/clinical data. ECG of general-purpose heart and torso model compared to experimental/clinical data.
	COU-driven validation results	Evaluation of a hERG model to predict pharmaceutical pro-arrhythmic risk	Evaluation of a cell model-based biomarker to predict pharmaceutical pro-arrhythmic risk (e.g., CiPA, discussed in section Cell Models)	Number of phase singularities during ventricular fibrillation (VF) compared to clinical data, when the model will be used to understand mechanisms behind VF—see section Organ-Level Models. Clinical evaluation of a whole-heart model which uses patient-specific information to predict optimal ablation targets to terminate arrhythmias—see section Organ-Level Models