Skip to main content

m09 · Propagation: cable theory, monodomain and bidomain, anisotropy

Warning

Education only. Not for diagnosis, treatment or any clinical decision.

In 60 seconds

Couple excitable cells through conductances and the action potential travels. Below threshold, a fiber is a leaky cable: a held voltage decays over a length constant set by membrane and axial resistances. Above threshold, each depolarized region pulls its neighbor past threshold and a wave moves at a steady conduction velocity. The bidomain model keeps the intracellular and extracellular spaces separate; when both are equally anisotropic it collapses to the simpler monodomain model, which is what most simulations use. Velocity scales with the square root of conductivity, and because cardiac fibers conduct better along their length, waves from a point spread as ellipses. Curved fronts are slower than flat ones, and a source too small to excite the tissue around it fails. Mapping systems measure the outcome of all of this, and what they show depends on the direction they measure in.

From the field

  • Fibers make velocity directional. Canine epicardial excitation spread 2.4 times faster along fibers than across them (Roberts et al. 1979); in calf trabeculae the ratio was 3.0 (Clerc 1976). A velocity measured along a line that is not aligned with the fibers lies in between (demo m09-d2).
  • Low-voltage tissue conducts slowly. In persistent atrial fibrillation patients, left atrial velocity averaged 1.0 m/s in normal-voltage areas of the anterior wall and 0.6 m/s in low-voltage areas (Heil et al. 2023).
  • Monodomain is usually enough. In organ-scale models of the human heart, propagation was only 2% faster in the bidomain model than in the monodomain model, and electrograms were visually indistinguishable when no current was applied (Potse et al. 2006).
  • Curvature matters at small scales. Wavefront curvature governs propagation from small electrodes, out of narrow structures and around sharp obstacles (Fast and Kléber 1997); it is one reason a stimulus can capture in one place and not in another.

Derivations

Symbols used throughout. \(V\) transmembrane voltage; \(\varphi_e\) extracellular potential; \(r_m\) membrane resistance times length, \(r_i\), \(r_e\) intracellular and extracellular axial resistances per length, \(c_m\) membrane capacitance per length; \(\lambda\) length constant (mm), \(\tau_m\) membrane time constant (ms). \(\sigma_i\), \(\sigma_e\) conductivity tensors, \(\sigma_m\) monodomain conductivity; \(\chi\) membrane area per volume; \(C_m\) capacitance per area; \(I_{\text{ion}}\) ionic current per area; \(D = \sigma_m/(\chi C_m)\) diffusion coefficient (mm²/ms). CV conduction velocity (mm/ms = m/s); subscripts L and T along and across fibers; \(\theta\) angle to the fibers; \(\kappa\) front curvature (1/mm), \(c_0\) plane wave velocity.

The passive cable

A fiber with leaky membrane and axial resistance obeys the cable equation:

\[ \lambda^2\frac{\partial^2 V}{\partial x^2} = \tau_m\frac{\partial V}{\partial t} + V, \qquad \lambda = \sqrt{\frac{r_m}{r_i + r_e}}, \quad \tau_m = r_m c_m \tag{1}\]

Hold one end at \(V_0\) and the steady state is \(V_0e^{-x/\lambda}\): 37% of the voltage remains one length constant away. The cable equation is the subthreshold part of propagation; it says how far a depolarized region can pull its neighbors toward threshold.

Key takeaways

  • Subthreshold voltage spreads over a length constant \(\sqrt{r_m/(r_i + r_e)}\).
  • Lower axial resistance means a longer reach.

Bidomain and monodomain

Tissue is two interpenetrating conductors, the cells and the space around them, coupled through the membrane. Current conservation in each gives the bidomain equations:

\[ \nabla\cdot\big(\sigma_i\nabla(V + \varphi_e)\big) = \chi\Big(C_m\frac{\partial V}{\partial t} + I_{\text{ion}}\Big), \qquad \nabla\cdot\big((\sigma_i + \sigma_e)\nabla\varphi_e\big) = -\nabla\cdot(\sigma_i\nabla V) \tag{2}\]

In 1D, with no net current along the fiber, the second equation gives \(\varphi_e = -\sigma_iV/(\sigma_i + \sigma_e)\) plus a constant. Substituting into the first leaves one equation for \(V\):

\[ \chi\Big(C_m\frac{\partial V}{\partial t} + I_{\text{ion}}\Big) = \nabla\cdot(\sigma_m\nabla V), \qquad \sigma_m = \sigma_i(\sigma_i + \sigma_e)^{-1}\sigma_e \tag{3}\]

The same reduction holds in 2D and 3D when \(\sigma_e\) is a multiple of \(\sigma_i\), that is, when both domains have the same anisotropy ratio. They do not: from Clerc’s resistance ratios in calf trabeculae, the intracellular domain is 9.4 times more conductive along fibers than across them, the extracellular domain 2.7 times (Clerc 1976). Even so, organ-scale comparisons find the monodomain model accurate for propagation without applied currents (Potse et al. 2006). Stimulation and defibrillation, where current is injected, need the bidomain model.

Key takeaways

  • The bidomain model has two potentials; the monodomain model has one, with conductivity \(\sigma_i\sigma_e/(\sigma_i + \sigma_e)\).
  • The reduction is exact in 1D and for equal anisotropy ratios, and a good approximation for propagation in general.

Conduction velocity

Divide the monodomain equation by \(\chi C_m\): conductivity appears only through \(D = \sigma_m/(\chi C_m)\). Stretching all lengths by \(\sqrt k\) turns \(D\) into \(kD\) and leaves the equation otherwise unchanged, so:

\[ \text{CV} \propto \sqrt{D}, \qquad \frac{\text{CV}_L}{\text{CV}_T} = \sqrt{\frac{\sigma_{m,L}}{\sigma_{m,T}}} \tag{4}\]

m33 uses the same scaling to tune conductivities (eq-m33-cv-scaling). On a Mitchell–Schaeffer cable, quadrupling \(D\) doubles the velocity to within 0.2%, and \(D\) = 0.86 mm²/ms gives 1.0 m/s (lab part B). Clerc’s resistance ratios give monodomain conductivities whose square-root ratio is 2.82, against a measured velocity ratio of 3.0 (Clerc 1976). Velocity also depends on the grid: a 0.4-mm grid computes a wave 2.7% slower than a 0.05-mm grid.

Simulation

Assumptions

  • Model: a 20-mm Mitchell–Schaeffer cable (Equation 3) by forward Euler, steps of 0.005 ms or less, no-flux ends, stimulus on the first 0.5 mm for 1 ms; velocity from a fit over the middle third. Tested against epml.propagation.
  • Parameters: membrane constants as in m08 (Nieto Ramos et al. 2023); \(D\) and the grid spacing are free choices for illustration.
  • What it ignores: discrete cells and gap junctions, fiber structure, curvature and the extracellular space.

Key takeaways

  • Velocity scales as the square root of conductivity; anisotropy in velocity is the square root of anisotropy in conductivity.
  • Coarse grids slow simulated waves; check convergence before trusting a simulated velocity.

Anisotropy

With fibers along one axis, a wave from a point eventually travels at \(\text{CV}_L\) along and \(\text{CV}_T\) across them, so its far-field isochrones are ellipses:

\[ T(\mathbf x) = \sqrt{\frac{x_L^2}{\text{CV}_L^2} + \frac{x_T^2}{\text{CV}_T^2}}, \qquad c(\theta) = \frac{\text{CV}_L\,\text{CV}_T}{\sqrt{\text{CV}_T^2\cos^2\theta + \text{CV}_L^2\sin^2\theta}} \tag{5}\]

\(c(\theta)\) is the apparent speed along a ray at angle \(\theta\) to the fibers. With a ratio of 3 and 1.0 m/s along fibers, a mapping line at 30° to them sees 0.58 m/s and one at 60° sees 0.38 m/s. In a simulated sheet with \(D_T = D_L/9\), the measured velocity ratio rises from 2.85 to 2.99 between 5 and 19 mm from the stimulus as the front flattens (lab part D).

Exact

Assumptions

  • Model: Equation 5 in uniform tissue, isochrones every 5 ms; the readout is \(c(\theta)\) for the horizontal line. Tested against epml.propagation.
  • Parameters: 1.0 m/s along fibers, reported for normal-voltage left atrium (Heil et al. 2023); the velocity ratio (default 3.0 (Clerc 1976)) and the fiber angle are adjustable.
  • What it ignores: curvature near the stimulus (part D of the lab), fiber rotation through the wall, boundaries and heterogeneity.

Key takeaways

  • Point stimuli give elliptical isochrones elongated along the fibers.
  • A velocity measured along an arbitrary line depends on its angle to the fibers.

Curvature

An expanding front feeds more tissue ahead of it than a flat one, so it slows:

\[ c(\kappa) \approx c_0 - D\,\kappa \tag{6}\]

For a circle of radius \(r\), \(\kappa = 1/r\). In a radially symmetric Mitchell–Schaeffer simulation with \(D\) = 0.1 mm²/ms the front speed matches \(c_0 - D/r\) within 2% from 3 mm outward (lab part E). Shrink the stimulated disk and the wave eventually fails: it launched from a 0.5-mm disk and failed from 0.3 mm, the same order as \(D/c_0\) = 0.29 mm. Curvature and this source–sink balance shape propagation from electrodes, out of narrow channels and around obstacles (Fast and Kléber 1997; Kléber and Rudy 2004).

Key takeaways

  • Convex fronts are slower by about \(D\kappa\).
  • A source must excite a large enough region to overcome the sink around it.

Clinical correlate

NotePending clinical review

This section describes published findings only. It is not guidance for patient care, and it has not yet been reviewed by an EP physician.

Where this shows up in EP. Velocity maps (m25) summarize propagation, and reported values differ with tissue state: 1.0 m/s in normal-voltage and 0.6 m/s in low-voltage left atrial areas in one study of persistent atrial fibrillation (Heil et al. 2023). Velocity also depends on direction: experimental velocity ratios along and across fibers of 2.4 and 3.0 (Roberts et al. 1979; Clerc 1976) mean that a value measured along an arbitrary line mixes tissue properties with geometry. Slowing and block at curved fronts and narrow structures (Fast and Kléber 1997; Kléber and Rudy 2004) are the substrate for the reentry mechanisms of m10.

What the evidence is. Experimental measurements in animal tissue (Clerc 1976; Roberts et al. 1979), a clinical mapping study (Heil et al. 2023), a modelling comparison (Potse et al. 2006) and reviews (Fast and Kléber 1997; Kléber and Rudy 2004).

What a reader should ask:

  • In which direction was velocity measured, relative to the fibers? (Equation 5)
  • How close to the pacing site? Near the source, curvature slows the front (Equation 6).
  • For simulated velocities: what grid, and was convergence checked? (Equation 4)

Visual summary

Schematic

λ cable: decay over λ anisotropy: fibers → slow fast curvature: c ≈ c0 − Dκ
Figure 1: Subthreshold spread, anisotropic spread and the cost of curvature.

Assumptions

How experts think

Before trusting a conduction velocity, ask:

  1. Along which direction? Compare it with the fibers (Equation 5).
  2. How far from the source? Near the pacing site the front is curved and slower (Equation 6).
  3. Plane wave or not? Velocity formulas assume a locally flat front.
  4. For a model: monodomain or bidomain, and why? Applied currents need the bidomain (Equation 2).
  5. For a simulation: grid converged? (Equation 4)

One sentence to keep: velocity is a property of the wave as much as of the tissue.

Cheat sheet

Quantity Result
Cable (Equation 1) \(\lambda = \sqrt{r_m/(r_i + r_e)}\); steady decay \(e^{-x/\lambda}\)
Bidomain (Equation 2) two potentials, two conductivity tensors
Monodomain (Equation 3) \(\sigma_m = \sigma_i(\sigma_i + \sigma_e)^{-1}\sigma_e\); exact in 1D or for equal anisotropy ratios
Velocity (Equation 4) \(\text{CV} \propto \sqrt D\); \(\text{CV}_L/\text{CV}_T = \sqrt{\sigma_{m,L}/\sigma_{m,T}}\)
Anisotropy (Equation 5) elliptical isochrones; \(c(\theta)\) between \(\text{CV}_T\) and \(\text{CV}_L\)
Curvature (Equation 6) \(c \approx c_0 - D\kappa\)

Values used in this module

Value Source
Velocity ratio 3.0 and resistance ratios 3.6, 12.6, 9.4, 2.7 (calf trabeculae) Clerc (1976)
Velocity ratio 2.4, resistivity ratio 3.2 (canine epicardium) Roberts et al. (1979)
1.0 m/s normal-voltage, 0.6 m/s low-voltage left atrium Heil et al. (2023)
Bidomain propagation 2% faster than monodomain Potse et al. (2006)
Membrane constants Nieto Ramos et al. (2023), based on Mitchell and Schaeffer (2003)
Velocities, grid errors, curvature fits, launch radii our own code

Interview questions

Try each question before you open its model answer. In a real interview, say which assumptions you are making.

Q1. What is the length constant, and what sets it?

Model answer

The distance over which a held voltage decays to \(1/e\) in a passive cable, \(\sqrt{r_m/(r_i + r_e)}\) (Equation 1).

Q2. Write the bidomain equations in words.

Model answer

Current leaving the intracellular space crosses the membrane as capacitive plus ionic current; total current (intracellular plus extracellular) is conserved (Equation 2).

Q3. When does the bidomain model reduce exactly to the monodomain model?

Model answer

In 1D, or when the extracellular conductivity tensor is a multiple of the intracellular one (equal anisotropy ratios); then \(\sigma_m = \sigma_i(\sigma_i + \sigma_e)^{-1}\sigma_e\) (Equation 3).

Q4. When would you insist on a bidomain model?

Model answer

When current is applied, as in pacing-field or defibrillation studies. For propagation alone, organ-scale comparisons found differences of about 2% (Potse et al. 2006).

Q5. You double the conductivity. What happens to the velocity?

Model answer

It rises by \(\sqrt 2\), about 41% (Equation 4).

Q6. Velocity ratio along/across fibers is 3. What is the conductivity ratio?

Model answer

9, for the monodomain conductivity (Equation 4).

Q7. Why are isochrones from a point elliptical?

Model answer

Velocity is highest along fibers and lowest across them, so the far-field activation time is \(\sqrt{x_L^2/\text{CV}_L^2 + x_T^2/\text{CV}_T^2}\) (Equation 5).

Q8. A map reports 0.5 m/s along a line in tissue with 1.0 m/s along fibers and a ratio of 3. What else could explain it?

Model answer

The line may be oblique to the fibers: at about 40° the apparent speed is near 0.5 m/s (Equation 5). Curvature near the source and the method of estimation also matter.

Q9. Why do waves slow near a small pacing electrode?

Model answer

The front is strongly curved, so each bit of excited tissue must depolarize more tissue ahead of it: \(c \approx c_0 - D\kappa\) (Equation 6) (Fast and Kléber 1997).

Q10. What is a source–sink mismatch?

Model answer

Too little depolarizing current to bring a larger region of resting tissue to threshold, as when a wave leaves a narrow channel or a stimulus excites too small a region (Equation 6) (Fast and Kléber 1997).

Q11. Your simulated velocity changes when you halve the grid spacing. What does that mean?

Model answer

The grid is too coarse for the front: refine until the velocity stops changing, or tune conductivity at the grid you will use (Equation 4).

Q12. Why might the extracellular potential matter even if the monodomain model is enough for propagation?

Model answer

Electrograms are extracellular signals. They can be computed from the monodomain solution with a separate volume-conductor step (m02, m03).

Code lab

The lab goes from a passive cable to a curved wave.

Part What you do Data
A The passive cable and its length constant (Equation 1) synthetic
B Velocity on an excitable cable: square-root scaling and grid error (Equation 4) synthetic
C From bidomain to monodomain; anisotropy from measured resistances (Equation 2, Equation 3) published ratios
D Spread from a point in anisotropic tissue (Equation 5) synthetic
E Curved fronts and the stimulus that fails (Equation 6) synthetic

Each part ends with checks: lines that pass with the code as given. If you change something and a check fails, that is information, not an error to hide.

Self-check

Short checks you can do in your head. Open each answer after you have tried it.

S1. \(r_m\) = 4, \(r_i + r_e\) = 1 (consistent units). Length constant?

Answer

2 (Equation 1).

S2. \(\sigma_i\) = 1, \(\sigma_e\) = 3 along a fiber. Monodomain conductivity?

Answer

0.75 (Equation 3).

S3. Velocity 0.5 m/s at \(D\); what at \(4D\)?

Answer

1.0 m/s (Equation 4).

S4. \(\text{CV}_L\) = 1.0, \(\text{CV}_T\) = 0.5 m/s. Apparent speed at 90° to the fibers?

Answer

0.5 m/s (Equation 5).

S5. \(c_0\) = 0.34 m/s, \(D\) = 0.1 mm²/ms. Speed of a circular front of radius 5 mm?

Answer

About 0.34 − 0.02 = 0.32 m/s (Equation 6).

S6. Which model would you use to simulate a defibrillation shock?

Answer

The bidomain model (Equation 2).

Literature figures

These figures are worth looking at alongside this module. We link to them instead of copying them: some licenses do not allow reuse here, and others we could not confirm. Each entry says what to look for and where this module explains it.

  • Elliptical spread on the epicardium. Roberts and colleagues show isochrones elongated along the fibers (Roberts et al. 1979). This module: anisotropy.
  • Curvature and block. Fast and Kléber illustrate fronts emerging from narrow structures and around obstacles (Fast and Kléber 1997). This module: curvature.
  • Monodomain against bidomain. Potse and colleagues compare activation and electrograms from both models (Potse et al. 2006). This module: bidomain and monodomain.

Further reading

References

Clerc, L. 1976. “Directional Differences of Impulse Spread in Trabecular Muscle from Mammalian Heart.” The Journal of Physiology 255 (2): 335–46. https://doi.org/10.1113/jphysiol.1976.sp011283.
Fast, V G, and A G Kléber. 1997. “Role of Wavefront Curvature in Propagation of Cardiac Impulse.” Cardiovascular Research 33 (2): 258–71. https://doi.org/10.1016/s0008-6363(96)00216-7.
Heil, Emanuel, Jin-Hong Gerds-Li, Julian Keznickl-Pulst, Vesna Furundzija-Cabraja, et al. 2023. “Left Atrial Conduction Times and Regional Velocities in Persistent Atrial Fibrillation Patients with and Without Fibrotic Atrial Cardiomyopathy.” Heart and Vessels 38 (10): 1277–87. https://doi.org/10.1007/s00380-023-02282-5.
Kléber, André G, and Yoram Rudy. 2004. “Basic Mechanisms of Cardiac Impulse Propagation and Associated Arrhythmias.” Physiological Reviews 84 (2): 431–88. https://doi.org/10.1152/physrev.00025.2003.
Mitchell, Colleen C, and David G Schaeffer. 2003. “A Two-Current Model for the Dynamics of Cardiac Membrane.” Bulletin of Mathematical Biology 65 (5): 767–93. https://doi.org/10.1016/S0092-8240(03)00041-7.
Nieto Ramos, Alejandro, Christopher L. Krapu, Elizabeth M. Cherry, and Flavio H. Fenton. 2023. Employing Gaussian Process Priors for Studying Spatial Variation in the Parameters of a Cardiac Action Potential Model. arXiv preprint arXiv:2311.10114. https://doi.org/10.48550/arXiv.2311.10114.
Potse, Mark, Bruno Dubé, Jacques Richer, Alain Vinet, and Ramesh M Gulrajani. 2006. “A Comparison of Monodomain and Bidomain Reaction-Diffusion Models for Action Potential Propagation in the Human Heart.” IEEE Transactions on Biomedical Engineering 53 (12 Pt 1): 2425–35. https://doi.org/10.1109/TBME.2006.880875.
Roberts, D E, L T Hersh, and A M Scher. 1979. “Influence of Cardiac Fiber Orientation on Wavefront Voltage, Conduction Velocity, and Tissue Resistivity in the Dog.” Circulation Research 44 (5): 701–12. https://doi.org/10.1161/01.res.44.5.701.