Skip to main content

m12 · The surface ECG and the forward problem

Warning

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

In 60 seconds

The surface ECG is the heart’s current sources seen through the torso. Activation fronts carry dipole moments (m01); the torso is a volume conductor (m02); each lead is a weighted difference of electrode potentials. Together these make every lead voltage a projection: for one equivalent dipole, the heart vector \(\mathbf p\), a lead records \(\mathbf L\cdot\mathbf p\), where the lead vector \(\mathbf L\) depends on the electrodes and the torso. In a homogeneous spherical torso with the heart at the center, the three limb leads have lead vectors 60° apart, the Einthoven triangle; because they share electrodes, lead I plus lead III always equals lead II. The augmented leads fill in the six-axis frontal reference frame, and the frontal axis follows from leads I and aVF. Real hearts sit off center in an inhomogeneous torso, and the lead vectors bend. Computing body-surface potentials from cardiac sources is the forward problem; m07 and m32 invert it.

From the field

  • Lead vectors depend on where the heart sits. In a spherical torso with the heart 0.3 of the radius left of center, lead I’s vector turns by 22° and the textbook axis formula reads a 60° heart vector as 78° (lab part E). Burger and van Milaan described the heart vector and its leads in these terms (Burger and Van Milaan 1947).
  • Standards fix the technology. A scientific statement reviews how the modern ECG is derived, recorded and displayed, including lead placement and digital processing (Kligfield et al. 2007).
  • Some lead systems aim at the heart vector itself. Frank designed a lead system for spatial vectorcardiography, the recording of the heart vector in three dimensions (Frank 1956).
  • Small vectors make noisy axes. With noise of 0.01 (relative units) on each lead, the axis error spread grows from 1.5° to 7.9° when the heart vector shrinks fivefold (lab part C).

Derivations

Symbols used throughout. \(\mathbf p\) dipole moment (heart vector); \(\sigma\) torso conductivity; \(R\) torso radius; \(\hat{\mathbf r}\) unit vector to a surface point; \(\theta\) angle between \(\mathbf p\) and \(\hat{\mathbf r}\); \(\mathbf r_0\) dipole position; \(I\) a point current; \(\gamma\) angle between \(\mathbf r_0\) and the surface point; \(P_l\) Legendre polynomials. \(\Phi_R\), \(\Phi_L\), \(\Phi_F\) potentials at the right arm, left arm and foot; I, II, III, aVR, aVL, aVF the limb leads; WCT the Wilson central terminal; \(\mathbf L\) lead vector; \(\alpha\) frontal axis (degrees, 0° toward the left arm, +90° toward the feet). Frontal-plane coordinates: \(x\) to the patient’s left, \(y\) inferior.

A dipole in a spherical torso

Inside an insulated, homogeneous sphere, the potential of a dipole at the center is the infinite-medium dipole plus a uniform-field correction that makes the normal current vanish at the surface:

\[ \varphi(r, \theta) = \frac{p\cos\theta}{4\pi\sigma}\left(\frac{1}{r^2} + \frac{2r}{R^3}\right), \qquad \varphi(R, \theta) = \frac{3\,\mathbf p\cdot\hat{\mathbf r}}{4\pi\sigma R^2} \tag{1}\]

The insulating boundary triples the surface potential compared with an infinite medium (lab part A). For a source away from the center, expand \(1/\lvert\mathbf x - \mathbf r_0\rvert\) in Legendre polynomials and add, term by term, the harmonic that cancels the normal current at \(R\):

\[ \varphi(R, \gamma) = \frac{I}{4\pi\sigma}\sum_{l \ge 1}\frac{2l + 1}{l}\,\frac{r_0^l}{R^{l+1}}\,P_l(\cos\gamma) + \text{const} \tag{2}\]

A dipole is a source–sink pair, so the constant cancels. Rudy and Plonsey used eccentric spheres to study the role of geometry and inhomogeneities in electrocardiography (Rudy and Plonsey 1979).

Key takeaways

  • A bounded torso changes surface potentials: three times the infinite-medium value for a centric dipole in a sphere.
  • An off-center source needs the series; it converges for any source inside the sphere.

Lead vectors

Potentials are linear in the sources, so every lead is linear in the dipole:

\[ V = \mathbf L\cdot\mathbf p, \qquad \mathbf L_{ab} = \frac{3}{4\pi\sigma R^2}\left(\hat{\mathbf a} - \hat{\mathbf b}\right) \text{ for a centric dipole} \tag{3}\]

where \(\hat{\mathbf a}\) and \(\hat{\mathbf b}\) point to the two electrodes. The lead-vector view goes back to Burger and van Milaan’s work on the heart vector and leads (Burger and Van Milaan 1947); reciprocity (m02, eq-m02-reciprocity) gives the same vector as the field the lead would make if it injected current. With the heart off center the lead vectors change length and direction: lead I’s vector turns by 22° when the heart moves 0.3 of the radius to the left.

Einthoven’s law and the augmented leads

The limb leads are differences of three electrode potentials, so they are not independent:

\[ \text{I} = \Phi_L - \Phi_R, \quad \text{II} = \Phi_F - \Phi_R, \quad \text{III} = \Phi_F - \Phi_L \;\Longrightarrow\; \text{I} + \text{III} = \text{II} \tag{4}\]

The Wilson central terminal averages the three, and the augmented leads compare each limb with the mean of the other two:

\[ \text{WCT} = \frac{\Phi_R + \Phi_L + \Phi_F}{3}, \qquad \text{aVR} = -\frac{\text{I} + \text{II}}{2}, \quad \text{aVL} = \frac{\text{I} - \text{III}}{2}, \quad \text{aVF} = \frac{\text{II} + \text{III}}{2} \tag{5}\]

so aVR + aVL + aVF = 0. Precordial leads are chest electrodes measured against the WCT; a scientific statement sets their placement and the recording standards (Kligfield et al. 2007).

The frontal axis

In Einthoven’s idealization the lead vectors of I and aVF are perpendicular, aVF’s being \(\sqrt3/2\) as long, so the frontal angle of the heart vector is:

\[ \alpha = \operatorname{atan2}\!\left(\frac{2\,\text{aVF}}{\sqrt3},\ \text{I}\right) \tag{6}\]

It is exact for a centric dipole in a sphere (lab part C). It becomes noisy when the vector is small, and biased when the lead vectors are not where the formula assumes (lab part E).

Exact

Assumptions

  • Model: a unit dipole at the center of a homogeneous, insulated sphere (Equation 1), limb electrodes at Einthoven’s equilateral positions; leads by Equation 4 and Equation 5; axis by Equation 6. Tested against epml.ecg.
  • Parameters: relative units; the vector’s angle is adjustable.
  • What it ignores: torso shape and inhomogeneity, heart position, distributed sources and time.

Key takeaways

  • Each lead records the projection of the heart vector on its lead vector.
  • I + III = II always; aVR + aVL + aVF = 0 always.

The forward problem

Many sources add: the body-surface potential is a sum of lead-vector projections over all active sources, here written for the frontal leads and a heart vector that changes with time,

\[ V(t) = \mathbf L\cdot\mathbf p(t), \qquad \mathbf p(t) = \int \text{dipole density of the activation fronts at time } t \tag{7}\]

A planar front’s dipole moment is proportional to its area (eq-m01-moment). Sweep a front across a ball of tissue 30 mm in radius at 0.6 m/s (illustrative) and the heart vector rises and falls along the front’s direction for \(2a/\text{CV}\) = 100 ms; the axis computed from the areas of leads I and aVF equals the front’s frontal angle (lab part D). Detailed forward models replace the sphere with the patient’s torso and the single dipole with the whole heart surface; m32 inverts them.

Exact

Assumptions

  • Model: lead vectors of the limb leads for a dipole off the center of a homogeneous, insulated sphere, by Equation 2 (120 terms, dipole as a source–sink pair) and Equation 3. Tested against epml.ecg.
  • Parameters: offsets as fractions of the torso radius, free choices for illustration.
  • What it ignores: lungs, blood and other inhomogeneities, the torso’s real shape and electrode placement errors.

Key takeaways

  • The ECG is a forward problem: sources, torso, leads.
  • The textbook triangle is a model; an off-center heart bends 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. The 12-lead ECG frames every EP case before a catheter is placed, and its recording is standardized: lead placement, digital acquisition and computer measurement (Kligfield et al. 2007). The lead-vector view explains why leads relate as they do (Burger and Van Milaan 1947), why a lead system can be designed to record the heart vector itself (Frank 1956), and why torso geometry matters (Rudy and Plonsey 1979). ECG-based localization of arrhythmia origins (m41), AI on the ECG (m37) and ECG imaging (m32) all build on this forward model.

What the evidence is. A scientific statement (Kligfield et al. 2007), classic lead-vector and lead-system work (Burger and Van Milaan 1947; Frank 1956) and a torso-model study (Rudy and Plonsey 1979).

What a reader should ask:

  • Were the electrodes where the analysis assumes? Lead vectors change with placement (Equation 3).
  • Is a derived quantity, such as the axis, robust to a small heart vector? (Equation 6)
  • What torso model does a forward or inverse method use? (Equation 2)

Visual summary

Schematic

+− torso as a volume conductor RLF Einthoven: I + III = II off-center heart bends the leads
Figure 1: The torso as a volume conductor, the Einthoven triangle, and its distortion.

Assumptions

How experts think

Before reading meaning into a lead or a derived quantity, ask:

  1. What does this lead project onto? Its lead vector (Equation 3).
  2. Which leads are independent? Only two of the six limb leads (Equation 4, Equation 5).
  3. Is the heart vector large enough for its direction to mean something? (Equation 6)
  4. How far is this torso from the model? Position and inhomogeneity bend lead vectors (Equation 2).
  5. For inverse methods: is the forward model good enough to invert? (m07, m32)

One sentence to keep: every ECG lead is a projection, and the torso decides the direction.

Cheat sheet

Quantity Result
Centric dipole in a sphere (Equation 1) surface \(3\,\mathbf p\cdot\hat{\mathbf r}/(4\pi\sigma R^2)\): three times the infinite medium
Off-center source (Equation 2) \(\frac{I}{4\pi\sigma}\sum_{l\ge1}\frac{2l+1}{l}\frac{r_0^l}{R^{l+1}}P_l(\cos\gamma)\)
Lead vector (Equation 3) \(V = \mathbf L\cdot\mathbf p\); centric: \(3(\hat{\mathbf a} - \hat{\mathbf b})/(4\pi\sigma R^2)\)
Einthoven (Equation 4) I + III = II
Augmented leads (Equation 5) aVR = −(I + II)/2, aVL = (I − III)/2, aVF = (II + III)/2
Frontal axis (Equation 6) \(\operatorname{atan2}(2\,\text{aVF}/\sqrt3, \text{I})\)
Forward problem (Equation 7) \(V(t) = \mathbf L\cdot\mathbf p(t)\)

Values used in this module

Value Source
Front ball radius 30 mm, velocity 0.6 m/s, noise 0.01, heart offsets illustrative
Lead-vector angles, axis errors, QRS duration 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. Why does an insulated torso triple the surface potential of a centric dipole?

Model answer

Current cannot leave, so a uniform-field correction is added to cancel the normal current at the surface; at the surface it adds twice the infinite-medium potential (Equation 1).

Q2. What is a lead vector?

Model answer

The vector \(\mathbf L\) such that the lead’s voltage is \(\mathbf L\cdot\mathbf p\) for a dipole \(\mathbf p\) at a given place; it depends on the electrodes and the torso (Equation 3) (Burger and Van Milaan 1947).

Q3. Prove Einthoven’s law.

Model answer

\((\Phi_L - \Phi_R) + (\Phi_F - \Phi_L) = \Phi_F - \Phi_R\) (Equation 4). It holds for any sources, because the leads share electrodes.

Q4. How many independent limb leads are there?

Model answer

Two; the other four follow from them (Equation 4, Equation 5).

Q5. What is the Wilson central terminal, and why is it not zero potential?

Model answer

The average of the three limb potentials, used as a reference. In the centric sphere model with equilateral electrodes it is exactly zero; with the heart off center it is not (Equation 5).

Q6. Lead I is +0.5 and aVF is 0. Frontal axis?

Model answer

0°, along lead I (Equation 6).

Q7. Why can the axis be unreliable in a low-voltage ECG?

Model answer

The angle of a small vector is sensitive to noise on the two leads; in the lab a fivefold smaller vector gave a fivefold larger spread (Equation 6).

Q8. How does heart position affect the limb leads?

Model answer

It changes the lead vectors’ lengths and directions; with the heart 0.3 of the radius to the left, lead I’s vector turned 22° in a sphere model (Equation 2, Equation 3).

Q9. What is the forward problem of electrocardiography?

Model answer

Computing body-surface potentials from known cardiac sources and a torso model (Equation 7); ECG imaging inverts it (m32).

Q10. Why does a front’s dipole moment grow and shrink as it crosses the heart?

Model answer

It is proportional to the front’s area, which grows then shrinks as the front crosses the tissue (Equation 7; m01).

Q11. What does a lead system for vectorcardiography need?

Model answer

To read the heart vector’s three components directly, three leads whose lead vectors are perpendicular and equally long (Equation 3). Frank designed a system for spatial vectorcardiography (Frank 1956).

Q12. You compute an axis from a digital ECG and it disagrees with the machine. What would you check?

Model answer

Which interval or measure each used (peak, area, mean), the lead definitions and filtering, and whether the machine’s algorithm follows the standards (Kligfield et al. 2007).

Code lab

The lab builds the ECG from a dipole in a sphere.

Part What you do Data
A A dipole in a spherical torso (Equation 1, Equation 2) synthetic
B Lead vectors, Einthoven’s law and the augmented leads (Equation 3, Equation 4, Equation 5) synthetic
C The frontal axis, with noise (Equation 6) synthetic
D A QRS from a front crossing the heart (Equation 7) synthetic
E An off-center heart (Equation 2, Equation 3) 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. I = 0.6, III = 0.3. Lead II?

Answer

0.9 (Equation 4).

S2. I = 0.6, II = 0.9. aVR?

Answer

−0.75 (Equation 5).

S3. aVL = 0.2, aVF = −0.1. aVR?

Answer

−0.1 (Equation 5).

S4. I = 0 and aVF positive. Axis?

Answer

+90° (Equation 6).

S5. A centric dipole’s surface potential in a sphere, against the same distance in an infinite medium?

Answer

Three times larger (Equation 1).

S6. A front crosses a 40-mm-wide region at 0.5 m/s. How long does its dipole last?

Answer

80 ms (Equation 7).

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.

  • Heart vector and leads. Burger and van Milaan’s analysis of leads in terms of the heart vector (Burger and Van Milaan 1947). This module: lead vectors.
  • A lead system for vectorcardiography. Frank’s clinically practical system for spatial vectorcardiography (Frank 1956). This module: lead vectors and the clinical correlate.
  • Eccentric spheres. Rudy and Plonsey’s model of the heart inside the torso, with geometry and inhomogeneities (Rudy and Plonsey 1979). This module: a dipole in a spherical torso.

Further reading

References

Burger, H C, and J B Van Milaan. 1947. “Heart-Vector and Leads. Part II.” British Heart Journal 9 (3): 154–60. https://doi.org/10.1136/hrt.9.3.154.
Frank, E. 1956. “An Accurate, Clinically Practical System for Spatial Vectorcardiography.” Circulation 13 (5): 737–49. https://doi.org/10.1161/01.cir.13.5.737.
Kligfield, Paul, Leonard S Gettes, James J Bailey, Rory Childers, et al. 2007. “Recommendations for the Standardization and Interpretation of the Electrocardiogram: Part I: The Electrocardiogram and Its Technology.” Circulation 115 (10): 1306–24. https://doi.org/10.1161/CIRCULATIONAHA.106.180200.
Rudy, Y., and R. Plonsey. 1979. “The Eccentric Spheres Model as the Basis for a Study of the Role of Geometry and Inhomogeneities in Electrocardiography.” IEEE Transactions on Biomedical Engineering 26 (7): 392–99. https://doi.org/10.1109/tbme.1979.326417.
Wagner, Patrick, Nils Strodthoff, Ralf-Dieter Bousseljot, et al. 2020. “PTB-XL, a Large Publicly Available Electrocardiography Dataset.” Scientific Data 7 (1): 154. https://doi.org/10.1038/s41597-020-0495-6.