18.02.02 · organismal-bio / cardiovascular

Cardiac action potentials, pacemaker physiology, and the ECG

shipped3 tiersLean: none

Anchor (Master): Bers, *Excitation-Contraction Coupling and Cardiac Contractile Force* (2nd ed., Kluwer 2001) — the canonical reference on cardiac calcium handling; Zipes, Jalife & Stevenson, *Cardiac Electrophysiology: From Cell to Bedside* (7th ed., Elsevier 2017); Noble 1962 (*J. Physiol.* 160, 317-352) — the first Purkinje action-potential model; DiFrancesco & Noble 1985 (*Phil. Trans. R. Soc. B* 307, 353-398) — the modern pacemaker model; Luo & Rudy 1991 (*Circ. Res.* 68, 1501-1526) and 1994 (*Circ. Res.* 74, 1071-1096; 1097-1113) — the dynamic ventricular model; Beeler & Reuter 1977 (*J. Physiol.* 268, 177-210); Maltsev & Lakatta 2009 (*Cardiovasc. Res.* 83, 367-374) — the coupled-clock pacemaker model

Intuition [Beginner]

The heart is a pump driven by a sequence of electrical pulses that march through the muscle in a stereotyped order, several dozen times a minute, for a lifetime. Each pulse is a cardiac action potential — a millisecond-scale change in the voltage across a heart-cell membrane, much like the neural action potential but with three twists that make the cardiac version distinct and that the rest of this unit is about.

The first twist is the plateau. Where a neuron spikes and snaps back down in about one millisecond, a ventricular heart cell spikes, then sits at a depolarised voltage for roughly two hundred milliseconds before repolarising. That long flat top of the cardiac waveform is what makes the heart different. It is built by a slow inward current of calcium ions through specialised channels that open during the spike and stay open for hundreds of milliseconds. The plateau lets calcium enter the cell in bulk, the calcium triggers contraction, and the cell stays inexcitable for the whole duration — long enough to push blood out before the next beat can start.

The second twist is autorhythmicity. A small specialised group of cells in the right atrium — the sinoatrial (SA) node — does not sit at a stable resting voltage between beats. After it repolarises, it slowly drifts back up toward the firing threshold under its own power, fires a new action potential, and the whole cycle repeats. That drift is the pacemaker depolarisation.

The drift is driven by an unusual ion channel that opens when the cell hyperpolarises; the current that flows through it pulls the cell back up again. DiFrancesco named the current the funny current, $I_{f}$ , because of its odd activation behaviour, and the name has stuck. The SA node sets the heart rate; the AV node, which sits between the atria and the ventricles, sets a backup rate in case the SA node fails.

The third twist is coupling. Heart muscle cells are wired together by special protein channels called gap junctions that let ions flow directly from one cell to the next. When one cell fires, its neighbours feel the current and fire too. The whole heart acts like a single coupled excitable medium, and the action potential spreads as a wave across the muscle, choreographing the contraction in a sequence that pushes blood out of the chambers in the right order.

Three places in the heart fire in three flavours of action potential, each tuned to its job. SA-node cells have small, slowly rising spikes driven by calcium rather than sodium, because they need to be controllable by the autonomic nervous system. Ventricular cells have huge fast spikes with a long plateau, because they need to drive a strong contraction. Specialised conduction cells in between — the AV node, the bundle of His, the Purkinje fibres — have intermediate properties chosen to route the wave correctly.

The electrocardiogram (ECG) is the macroscopic shadow of all of this on the body surface. As waves of action potentials sweep across millions of coupled heart cells, the moving boundary between depolarised and resting tissue acts like a tiny electrical dipole that moves through the chest. Electrodes on the skin measure the projection of that moving dipole onto the line between them, and the resulting tracing — the familiar P-QRS-T pattern — is the signature of the whole electrical sequence: atrial depolarisation (P), ventricular depolarisation (QRS), and ventricular repolarisation (T).

So the cardiac action potential is the same general kind of object as the neural one, but specialised in three ways for the job of driving a pump that runs on schedule for a hundred years. The plateau gives contraction time. The funny current gives autorhythmicity. Gap junctions give the coupling that turns millions of single cells into a single coordinated organ.

Visual [Beginner]

The defining picture is a side-by-side comparison of three voltage traces. The neuronal action potential, on the left, is a sharp upward spike that lasts about one millisecond. The ventricular cardiac action potential, in the middle, has an equally sharp upstroke but stays elevated for a long plateau of about two hundred milliseconds before dropping back. The SA-node pacemaker action potential, on the right, has no flat resting voltage at all: between spikes it slowly drifts upward from about $- 60$ mV to the firing threshold near $- 40$ mV, fires, repolarises, and immediately starts drifting up again. The cycle is a continuous oscillation, not a sequence of spikes triggered from outside.

A second useful picture is the heart laid out as a conduction map. The SA node sits at the top of the right atrium and fires first. The wave spreads across both atria (producing the P wave on the ECG), reaches the AV node, slows down (the PR segment), travels through the bundle of His and bundle branches, and finally reaches the ventricles through the Purkinje network, where it depolarises the working myocardium rapidly (the QRS complex). The ventricles then sit in their plateau (the ST segment) and repolarise (the T wave). Then the whole heart rests during electrical diastole and the SA node fires again.

Worked example [Beginner]

Walk through a single beat as a sequence of events at the cellular level, in order.

Step 1. The SA node finishes a spike and repolarises to about $- 60$ mV. The funny current $I_{f}$ activates because the cell is now hyperpolarised. Sodium and potassium flow inward through the funny-current channel; the cell starts to depolarise slowly. At the same time, intracellular calcium release events from the sarcoplasmic reticulum add small inward currents through the sodium-calcium exchanger, helping the drift along. The voltage climbs from $- 60$ to $- 50$ to $- 40$ mV over about 500 ms.

Step 2. At the firing threshold near $- 40$ mV, L-type calcium channels open. Calcium flows in. The voltage rises sharply (slower than a neuron's sodium-driven upstroke, because L-type calcium activation is slower than fast sodium activation). The SA node has fired.

Step 3. The depolarisation spreads through gap junctions to neighbouring atrial muscle cells. Atrial cells have voltage-gated sodium channels, so their action potentials have rapid upstrokes once threshold is reached. The wave sweeps across both atria, producing the P wave on the surface ECG. Atrial contraction begins, topping off ventricular filling.

Step 4. The wave reaches the AV node. AV-node cells are slow: their action potential is calcium-driven like the SA node, and conduction velocity drops by roughly a factor of ten. This delay (about 100 ms) ensures the ventricles do not fire until the atria have finished contracting. The PR interval on the ECG is the timing of this delay.

Step 5. The wave exits the AV node into the bundle of His, fans out through the bundle branches, and reaches the Purkinje fibres — fast-conducting specialised cells that distribute the wave to the ventricular myocardium near-simultaneously across both ventricles. The ventricular wall depolarises within about 80 ms. This rapid activation produces the QRS complex on the ECG.

Step 6. Ventricular myocytes are now in their plateau. Each cell sits near $+ 10$ mV for about 200 ms. Calcium is flowing in through L-type channels, triggering further calcium release from the sarcoplasmic reticulum, which activates the actin-myosin machinery; the ventricle contracts and ejects blood into the aorta and pulmonary artery. The plateau is electrically silent on the ECG — the ST segment is flat — because all the ventricular cells are depolarised together and there is no moving boundary to project onto the body surface.

Step 7. Ventricular cells repolarise: delayed-rectifier potassium channels finish opening, L-type calcium channels inactivate, and the voltage falls back to about $- 85$ mV. Repolarisation does not happen everywhere at once — the inner endocardium repolarises after the outer epicardium because plateau durations differ across the wall — and that moving repolarisation boundary produces the T wave on the ECG.

Step 8. The whole heart sits at rest for a few hundred milliseconds until the SA node fires the next beat.

What this tells us: the heartbeat is a choreographed sequence of three distinct cardiac action potential shapes (pacemaker, atrial, ventricular) firing in a fixed spatial order, with the timing set by the SA node's autorhythmicity and the spatial order set by the wiring of the conduction system. Each part of the surface ECG corresponds to a specific phase of this sequence.

Check your understanding [Beginner]

Exercise (easy, multiple choice).

The plateau phase of a ventricular cardiac action potential is sustained primarily by:

A. A persistent inward sodium current. B. A persistent inward calcium current through L-type calcium channels, balanced against an outward potassium current. C. A pure absence of any current — the membrane is sitting at the sodium equilibrium potential. D. The funny current $I_{f}$ .

Hint

The plateau is what makes the cardiac action potential different from the neural one. Which ion species crosses into the cell in bulk during the plateau and links the electrical signal to contraction?

Answer

B. The plateau is a balance between an inward L-type calcium current (the slow inward current of Beeler-Reuter and Luo-Rudy) and an outward delayed-rectifier potassium current. The calcium influx during the plateau is what triggers calcium-induced calcium release from the sarcoplasmic reticulum and drives contraction. A persistent inward sodium current (the late sodium current $I_{N a L}$ ) contributes secondarily in some cell types but is not the primary plateau driver. The funny current $I_{f}$ acts in pacemaker cells during diastolic depolarisation, not during the ventricular plateau.

Exercise (easy, multiple choice).

On the surface ECG, the QRS complex corresponds to:

A. Atrial depolarisation. B. Ventricular depolarisation — the rapid spread of the activation wavefront through the ventricular myocardium. C. Ventricular repolarisation. D. The plateau phase of the ventricular action potential.

Hint

The QRS is the large, sharp deflection. It marks the start of ventricular contraction.

Answer

B. The QRS complex is the body-surface signature of the rapid spread of the depolarisation wave through the ventricular myocardium. The wave is large and fast because the ventricular muscle mass is large, conduction through the Purkinje system is rapid, and the activation produces a large moving dipole. The P wave is atrial depolarisation; the T wave is ventricular repolarisation; the plateau itself is electrically silent on the surface ECG (the ST segment).

Exercise (medium, numeric).

A patient has a resting heart rate of $60$ beats per minute. Roughly how long does one full cardiac cycle last, and what fraction of that cycle is occupied by the plateau of a typical ventricular action potential (duration about $250$ ms)?

Hint

One beat per second at $60$ bpm. The plateau is a fixed-duration absolute refractory period that sets a floor on how short the cycle can become.

Answer

One full cycle lasts $1000$ ms; the plateau occupies about $250/1000 = 25%$ of the cycle. At $60$ bpm, the interval between beats is exactly $1$ second. The ventricular plateau of about $250$ ms means that the ventricles are electrically inexcitable for a quarter of every cycle. This long absolute refractory period is the structural reason a healthy heart cannot tetanise the way a skeletal muscle can: a second beat triggered too soon will encounter inexcitable tissue and fail to propagate. At higher heart rates the plateau shortens (rate-dependent action-potential-duration adaptation), preserving the safety margin.

Formal definition [Intermediate+]

Cardiac myocytes are excitable cells whose plasma membrane separates an intracellular compartment from the extracellular fluid; concentrations and equilibrium potentials follow the same Nernst / Goldman-Hodgkin-Katz framework laid out in 17.09.02 pending for neural cells. Typical mammalian ventricular concentrations are $[K^{+}]_{i} \approx 140$ mM, $[K^{+}]_{o} \approx 4$ mM, $[Na^{+}]_{i} \approx 10$ mM, $[Na^{+}]_{o} \approx 145$ mM, $[Ca^{2 +}]_{i} \approx 1 0^{- 4}$ mM (diastolic) rising to $\approx 1 0^{- 3}$ mM at the peak of the calcium transient, and $[Ca^{2 +}]_{o} \approx 1.8$ mM. At $T = 310$ K the equilibrium potentials are $E_{K} \approx - 94$ mV, $E_{Na} \approx + 71$ mV, $E_{Ca} \approx + 130$ mV.

A conductance-based cardiac model specialises the Kirchhoff-current-law cable framework of 17.09.02 pending to the ionic-current ensemble of a particular cell type. For a ventricular myocyte at the cellular (space-clamped) level,

C_{m} \frac{d V _{m}}{d t} = - (I_{Na} + I_{Ca, L} + I_{to} + I_{Kr} + I_{Ks} + I_{K1} + I_{NaK} + I_{NaCa} + I_{Ca, b} + I_{Na, b}) + I_{stim},

with the fast sodium current $I_{Na} = \overset{g}{ˉ}_{Na} m^{3} hj (V_{m} - E_{Na})$ (an $m^{3} hj$ form generalising the neural $m^{3} h$ by adding a slow inactivation gate $j$ ), the L-type calcium current $I_{Ca, L}$ with activation $d$ , voltage inactivation $f$ , and calcium-dependent inactivation $f_{C a}$ , the transient-outward potassium current $I_{to}$ (the early repolarisation notch), the rapid and slow delayed-rectifier potassium currents $I_{Kr}$ and $I_{Ks}$ (the major plateau-terminating currents), the inward-rectifier $I_{K1}$ (sets the resting potential at $- 85$ mV near $E_{K}$ ), the electrogenic Na+/K+ pump current $I_{NaK}$ , the Na+/Ca2+ exchanger current $I_{NaCa}$ (electrogenic 3:1 stoichiometry), and small background leaks. Each gating variable $x \in {m, h, j, d, f, f_{C a}, \dots}$ obeys first-order kinetics $d x / d t = α_{x} (V_{m}) (1 - x) - β_{x} (V_{m}) x$ with empirically fit voltage-dependent rate constants. The Luo-Rudy LR2 dynamic model ^{[Luo-Rudy 1994]} couples these currents to intracellular calcium dynamics through the SERCA pump, the ryanodine-receptor release flux, and calcium buffers.

The action-potential phases are conventionally numbered $0$ through $4$ following Hoffman-Cranefield. Phase $0$ is the rapid upstroke ( $I_{Na}$ dominant in ventricular cells, $I_{Ca, L}$ dominant in nodal cells). Phase $1$ is the early repolarisation notch ( $I_{to}$ activation, brief). Phase $2$ is the plateau ( $I_{Ca, L}$ inward against $I_{Kr} / I_{Ks}$ outward, $200$ – $300$ ms). Phase $3$ is final repolarisation ( $I_{Kr} / I_{Ks}$ dominant as $I_{Ca, L}$ inactivates). Phase $4$ is the resting (or, in pacemaker cells, the diastolic-depolarising) phase: in ventricular cells set by $I_{K1}$ at $V_{m} \approx E_{K}$ ; in SA-node cells dominated by $I_{f}$ (the funny current), background calcium influx through T-type and L-type channels, and the calcium-clock-driven Na+/Ca2+ exchanger inward current.

The funny current $I_{f}$ is a mixed Na+/K+ inward current activated by hyperpolarisation:

I_{f} = \overset{g}{ˉ}_{f} y (V_{m} - E_{f}), E_{f} \approx - 25 mV,

where the gating variable $y$ has an activation curve $y_{\infty} (V_{m})$ shifted to hyperpolarised potentials — $y$ activates as the cell hyperpolarises, the opposite direction from the standard $m$ -gate of $I_{Na}$ . The channel carrying $I_{f}$ is the HCN family (hyperpolarisation-activated, cyclic-nucleotide-gated; molecularly cloned by DiFrancesco and others in the 1990s). The reversal potential $E_{f} \approx - 25$ mV lies between $E_{K}$ and $E_{N a}$ because the channel is non-selectively permeable to both ions, so opening of $I_{f}$ at hyperpolarised voltages drives the cell inward toward $E_{f}$ , producing the diastolic depolarisation ^{[DiFrancesco-Noble 1985]}.

The conduction velocity of an excitation wave on a cable of myocytes follows from the active cable equation $$ \frac{a}{2 r_a} \partial_{xx} V_m ;=; C_m \partial_t V_m + I_{\rm ion}(V_m, \mathbf{g}), $$ which is the same form as the neural cable equation of 17.09.02 pending but with the cardiac $I_{ion}$ rather than the squid-axon HH ionic current. The axial resistance $r_{a}$ depends crucially on the gap-junction conductance $g_{j}$ between adjacent myocytes; intracellular cytoplasm contributes a small fraction. In two and three dimensions, the cable equation generalises to the bidomain model (Tung 1978, Henriquez 1993), in which intracellular and extracellular potentials are coupled by membrane current with anisotropic conductivity tensors reflecting fibre orientation in the ventricular wall.

The electrocardiogram is the body-surface manifestation of the cardiac excitation. In the dipole approximation, each patch of myocardium at the boundary between depolarised and resting tissue acts as a small electrical dipole oriented along the local activation gradient; the volume-conductor integral of these dipoles over the entire heart at instant $t$ gives a single heart vector $H (t)$ . Voltage difference between two body-surface electrodes is approximately $$ V_{\rm electrode}(t) ;=; \mathbf{H}(t) \cdot \mathbf{e}{AB}, $$ where $\mathbf{e}{AB} $i s t h e u ni t v ec t or a l o n g t h e l e a d a x i s b e tw ee n e l ec t r o d es$ A $an d$ B $. E in t h o v e n^{'} s t h r ees t an d a r d bi p o l a r l imb l e a d s (I = l e f t a r m - r i g h t a r m; I I = l e f tl e g - r i g h t a r m; I I I = l e f tl e g - l e f t a r m) d e f in e * * E in t h o v e n^{'} s t r ian g l e * *, w h ose t h r ee l e a d v ec t or s l i er o ug h l y in t h e f r o n t a l pl an e a t$ 0° $,$ 60° $, an d$ 120° $; t h e E C G co m pl e x o n e a c h l e a d i s t h e p r o j ec t i o n o f$ \mathbf{H}(t)$ onto the corresponding lead axis.

Counterexamples to common slips

The plateau is not a passive state. It is a dynamic balance between large inward and outward currents; small shifts in $I_{C a, L}$ or $I_{K r}$ produce large changes in action-potential duration. Long-QT syndromes from $I_{K r}$ mutations exploit this sensitivity.
The funny current is not an outward potassium current. It is a mixed-cation inward current at hyperpolarised potentials, with reversal $E_{f} \approx - 25$ mV. At a typical diastolic voltage of $- 60$ mV it drives the cell inward (depolarising), opposite the action of $I_{K 1}$ which fixes the ventricular resting voltage.
The SA node does not fire because $I_{f}$ alone drives it. The modern coupled-clock model (Maltsev-Lakatta 2009) has the funny current cooperating with rhythmic intracellular calcium release events that produce inward Na+/Ca2+ exchanger current; pharmacologic block of either clock degrades but does not abolish pacemaking.
The QRS amplitude does not reflect contraction force. The QRS is determined by the volume of myocardium activated and the synchrony of activation; a hypertrophied but poorly contracting ventricle can have a tall QRS. ECG voltage encodes electrical events, not mechanical events directly.
Ventricular repolarisation order is reversed from depolarisation order. The epicardium repolarises before the endocardium because its plateau is shorter, even though the endocardium depolarises first. This reversal is why the T wave has the same polarity as the QRS in normal ECGs.

Key theorem with proof [Intermediate+]

Theorem (cable-equation conduction velocity for a uniform excitable cable). Consider an infinite cylindrical cable of cardiac myocytes carrying a propagating action potential with active cable equation

\frac{a}{2 r _{a}} \partial_{xx} V_{m} = C_{m} \partial_{t} V_{m} + I_{ion} (V_{m}, g) .

Suppose a traveling-wave solution $V_{m} (x, t) = U (ξ)$ exists with $ξ = x - θ t$ and $θ$ the propagation speed. Then $θ$ satisfies the eigenvalue problem

\frac{a}{2 r _{a}} U^{''} (ξ) + θ C_{m} U^{'} (ξ) - I_{ion} (U (ξ), g (ξ)) = 0,

together with the gating ODEs $θ g^{'} (ξ) = - α (U) \cdot (g - g_{\infty} (U)) / τ (U)$ . The eigenvalue $θ$ scales as $θ \propto a / r_{a}$ at fixed ionic kinetics, and reducing the gap-junction conductance (increasing the effective $r_{a}$ ) decreases $θ$ — the conduction slowing seen in fibrosis and aged myocardium.

Proof. Substitute $V_{m} (x, t) = U (x - θ t)$ into the active cable equation. The time derivative is $\partial_{t} V_{m} = - θ U^{'} (ξ)$ and the spatial Laplacian is $\partial_{xx} V_{m} = U^{''} (ξ)$ . The cable equation becomes the second-order ODE $$ \frac{a}{2 r_a} U''(\xi) = -\theta C_m U'(\xi) + I_{\rm ion}(U, \mathbf{g}), $$ i.e. $(a /2 r_{a}) U^{''} + θ C_{m} U^{'} - I_{ion} = 0$ . The gating variables under the traveling-wave ansatz become $g (x, t) = g (ξ)$ and their kinetic equations $d g / d t = (g_{\infty} (V_{m}) - g) / τ (V_{m})$ transform to $- θ g^{'} (ξ) = (g_{\infty} (U) - g) / τ (U)$ . Boundary conditions are $U (- \infty) = V_{rest}$ (resting in the wake), $U (+ \infty) = V_{rest}$ (resting ahead), with the action-potential excursion in between.

To extract the scaling, non-dimensionalise: let $\hat{ξ} = ξ / λ$ with $λ = R_{m} \cdot a / (2 r_{a})$ (the membrane length constant for a cable of radius $a$ ), and $\hat{t} = t / τ_{m}$ with $τ_{m} = R_{m} C_{m}$ (membrane time constant; $R_{m}$ is the resting membrane resistance per unit area). Then $θ = λ / τ_{m} \cdot \hat{θ}$ where $\hat{θ}$ is a dimensionless eigenvalue depending only on the non-dimensional ionic kinetics. Substituting $λ = R_{m} a / (2 r_{a})$ and $τ_{m} = R_{m} C_{m}$ , $$ \theta ;=; \frac{1}{C_m}\sqrt{\frac{a}{2 r_a R_m}} \cdot \hat\theta ;\propto; \sqrt{a / r_a}, $$ holding $R_{m}$ and the ionic kinetics fixed. The propagation speed therefore scales as the square root of the cable radius and as the square root of the inverse axial resistance. Doubling $r_{a}$ (e.g., by halving gap-junction conductance) reduces $θ$ by a factor $2$ — substantial conduction slowing. The existence and stability of the traveling-wave solution for a Hodgkin-Huxley-type kinetic system is the FitzHugh-Nagumo / Keener-Sneyd theorem (asymptotic analysis valid in the singular limit of well-separated time scales). $□$

Bridge. The square-root scaling builds toward the reentry analysis in the Master tier below: when gap junctions deteriorate, conduction slows, and the wavelength $λ_{w} = θ \cdot τ_{refractory}$ shrinks. Reentry requires the wavelength to fit within an available anatomical pathway — a shorter wavelength allows reentry on smaller circuits, which is the foundational reason atrial fibrillation becomes harder to terminate as the atrium remodels. This is exactly the link between cellular-scale conduction parameters and organ-scale arrhythmia substrate, and the same wavelength-vs-pathway argument generalises from the cable to two-dimensional spirals and three-dimensional scroll waves. The cable theorem appears again in 17.09.02 pending as the parent passive-cable result for the squid axon, and putting these together identifies cardiac propagation with the same dynamical-systems object — a traveling-wave eigenvalue problem on an excitable medium — that organises neural conduction.

Exercises [Intermediate+]

Exercise 2 (easy, multiple choice).

Which current is principally responsible for setting the resting membrane potential of a ventricular myocyte near $- 85$ mV?

A. The funny current $I_{f}$ . B. The fast sodium current $I_{Na}$ . C. The inward-rectifier potassium current $I_{K1}$ . D. The L-type calcium current $I_{Ca, L}$ .

Hint

At rest, ventricular myocytes sit near $E_{K}$ . Which channel is open at rest with high enough conductance to clamp $V_{m}$ there?

Answer

C. $I_{K1}$ (the inward-rectifier potassium current carried by Kir2.x channels) is the dominant resting conductance in ventricular myocytes; its high conductance at hyperpolarised potentials and rectifying property at depolarised potentials together produce both the stable resting potential near $E_{K} = - 94$ mV and the maintenance of the plateau. SA-node and AV-node cells have very little $I_{K1}$ , which is precisely why they have no stable resting potential and can pacemake.

Exercise 3 (medium, numeric).

The conduction velocity of an excitation wave in normal ventricular myocardium along the fibre direction is approximately $θ_{L} \approx 50$ cm/s; across fibres it is approximately $θ_{T} \approx 17$ cm/s. The ratio $θ_{L} / θ_{T} \approx 3$ . Using the theorem above (with $θ \propto 1/ r_{a}$ ), what ratio of effective axial-resistance values explains this anisotropy?

Hint

If $θ \propto 1/ r_{a}$ at fixed ionic kinetics, then $(θ_{L} / θ_{T})^{2} = r_{a, T} / r_{a, L}$ .

Answer

$(θ_{L} / θ_{T})^{2} = 3^{2} = 9$ . So the transverse axial resistance is roughly nine times the longitudinal one: $r_{a, T} / r_{a, L} \approx 9$ . This matches measured gap-junction-distribution anisotropy in ventricular myocardium — gap junctions are concentrated at the end-to-end intercalated discs (along the fibre axis) and sparser at side-to-side contacts (across fibres). The square-root-of-axial-resistance scaling is the foundational reason cardiac conduction is anisotropic and why fibre-orientation reconstruction matters for predicting reentry pathways.

Exercise 4 (medium, symbolic).

Write the SA-node pacemaker dynamics as a minimal three-variable conductance-based model with state $(V_{m}, y, d)$ , where $y$ is the funny-current activation and $d$ is the L-type calcium activation. Assume the only currents are $I_{f}$ , $I_{Ca, L}$ , a constant background outward potassium current $I_{K, b}$ , and ignore inactivation gates and intracellular calcium dynamics. Identify what makes this system autonomous (no external input current needed for oscillation).

Hint

The right-hand side of $d V_{m} / d t$ should be a sum of three current terms with no $I_{stim}$ . The system is autonomous because $I_{f}$ activates at hyperpolarised potentials.

Answer

C_{m} \dot{V}_{m} \overset{y}{˙} \dot{d} = - \overset{g}{ˉ}_{f} y (V_{m} - E_{f}) - \overset{g}{ˉ}_{C a} d (V_{m} - E_{C a}) - \overset{g}{ˉ}_{K, b} (V_{m} - E_{K}), = (y_{\infty} (V_{m}) - y) / τ_{y} (V_{m}), = (d_{\infty} (V_{m}) - d) / τ_{d} (V_{m}) .

The system is autonomous because no $I_{stim}$ appears. At hyperpolarised $V_{m}$ near $- 60$ mV, $y_{\infty} (V_{m}) \approx 1$ and $I_{f}$ activates, pulling $V_{m}$ inward toward $E_{f} \approx - 25$ mV. The depolarisation activates $d$ , opening L-type calcium channels and driving an upstroke toward $E_{C a}$ . The upstroke deactivates $y$ , terminating $I_{f}$ , and the cell repolarises under $I_{K, b}$ . The cycle is a limit-cycle oscillator on $R \times [0, 1]^{2}$ analogous to the FitzHugh-Nagumo system 02.12.14 but with a single fast gate and one slow gate. The funny current is the load-bearing autorhythmicity element: removing $I_{f}$ (setting $\overset{g}{ˉ}_{f} = 0$ ) collapses the oscillation to a stable equilibrium at $V_{m} = (\overset{g}{ˉ}_{C a} E_{C a} + \overset{g}{ˉ}_{K, b} E_{K}) / (\overset{g}{ˉ}_{C a} + \overset{g}{ˉ}_{K, b})$ near $E_{K}$ (since $\overset{g}{ˉ}_{C a} \approx 0$ at hyperpolarised potentials).

Exercise 5 (medium, numeric).

The autonomic nervous system modulates heart rate. Vagal stimulation (acetylcholine, parasympathetic) shifts the activation curve of $I_{f}$ negatively by about $10$ mV; sympathetic stimulation shifts it positively by about $10$ mV. If the resting diastolic depolarisation has a steady slope $d V_{m} / d t \approx \overset{g}{ˉ}_{f} \cdot y_{\infty} (V_{m}) \cdot (V_{m} - E_{f}) / C_{m}$ during the diastolic interval, estimate the fractional change in heart rate from a $10$ -mV negative shift of $y_{\infty} (V_{m})$ , assuming the firing threshold is fixed at $- 40$ mV and the diastolic minimum is fixed at $- 60$ mV.

Hint

A negative shift of the activation curve reduces $y_{\infty}$ at any fixed $V_{m}$ in the diastolic range. If $y_{\infty}$ at $V_{m} = - 50$ mV drops from (say) $0.5$ to $0.2$ , the slope of phase 4 drops, the time to reach threshold lengthens, and the heart rate drops.

Answer

Negative shifts of $y_{\infty}$ reduce the fraction of $I_{f}$ channels open in the diastolic voltage range. Quantitatively, taking $y_{\infty}$ as a Boltzmann curve with $V_{1/2} = - 65$ mV and slope $k = 8$ mV in baseline, the average $y_{\infty}$ over the diastolic range $[- 60, - 40]$ mV is about $0.30$ ; after a $- 10$ mV shift to $V_{1/2} = - 75$ mV, the average over the same range drops to about $0.10$ , a threefold reduction. The diastolic slope is proportional to $y_{\infty}$ (when $V_{m} - E_{f}$ varies less than $y_{\infty}$ over the range), so the slope drops threefold and the diastolic interval lengthens threefold. For a baseline heart rate of $75$ bpm with diastolic interval $\approx 500$ ms, the new diastolic interval is $\approx 1500$ ms, total cycle $\approx 1750$ ms, new heart rate $\approx 34$ bpm — a substantial slowing. This is consistent with the bradycardia caused by strong vagal tone. Sympathetic shifts in the opposite direction accelerate the heart by the symmetric mechanism (a baseline rate of $75$ becomes about $130$ bpm).

Exercise 6 (medium, symbolic).

The Beeler-Reuter ventricular model has a calcium current $I_{s} = \overset{g}{ˉ}_{s} df (V_{m} - E_{s})$ with $E_{s} \approx + 80$ mV (the apparent reversal includes electrogenic Na+/Ca2+ exchange contributions). Show that during the plateau at $V_{m} \approx + 10$ mV, with both $d$ and $f$ near $0.5$ , the calcium influx delivered per beat ( $\int I_{s} d t$ over a $250$ -ms plateau) is much larger than the influx during a brief sodium spike of the same nominal current amplitude.

Hint

Total charge $Q = \int I d t$ . The plateau lasts $\sim 100 \times$ longer than the upstroke.

Answer

With $\overset{g}{ˉ}_{s} \approx 0.09$ mS/cm² (a representative LR1 value), $df \approx 0.25$ , $V_{m} - E_{s} \approx - 70$ mV, the plateau calcium current is $I_{s} \approx 0.09 \cdot 0.25 \cdot (- 70) \approx - 1.6$ μA/cm². Integrated over 250 ms, $Q_{C a} \approx - 1.6 \cdot 0.25 = - 0.4$ μC/cm² delivered inward. A sodium upstroke with peak current $\sim - 200$ μA/cm² lasting 1 ms delivers $Q_{N a} \approx - 0.2$ μC/cm² — only half the charge. The plateau, despite small current magnitudes, delivers more charge than the upstroke because of its long duration. This delivered calcium is what triggers the ryanodine-receptor-mediated calcium release from the sarcoplasmic reticulum and drives contraction in the calcium-induced-calcium-release cascade quantified by Bers ^{[Bers 2001]}.

Exercise 7 (hard, symbolic).

Consider a one-dimensional ring of cardiac myocytes of total length $L$ . A reentrant excitation wave propagates around the ring with conduction velocity $θ$ and the wavefront is followed by a refractory tail of duration $τ_{R}$ (the action-potential refractory period). Derive the condition on $L$ , $θ$ , and $τ_{R}$ for stable reentry, and identify what happens when the condition is violated.

Hint

The wavefront catches up to its own tail when the ring circumference equals the wavelength $λ_{w} = θ τ_{R}$ .

Answer

Define the wavelength of the reentrant circuit $λ_{w} := θ \cdot τ_{R}$ . Stable reentry requires the wavefront, after traveling all the way around the ring, to encounter recovered (excitable) tissue rather than its own refractory tail. The condition is therefore $$ L ;\geq; \lambda_w ;=; \theta \tau_R. $$ If $L < λ_{w}$ , the wavefront encounters refractory tissue, the wave terminates, and reentry ceases. If $L = λ_{w}$ exactly, the reentrant period is $T = τ_{R}$ . If $L > λ_{w}$ , the wavefront propagates into excitable tissue and the period is $T = L / θ > τ_{R}$ , with an excitable gap between front and tail. Pharmacologic anti-arrhythmics that prolong $τ_{R}$ (class III, e.g., amiodarone, sotalol) increase $λ_{w}$ and force the ring to be longer than $λ_{w}$ to sustain reentry — terminating reentries whose path is too short. Conversely, conditions that shorten $τ_{R}$ (ischaemic potassium accumulation, catecholamine surges) or slow $θ$ (fibrosis, gap-junction uncoupling) allow reentry on shorter circuits, increasing arrhythmia vulnerability. This wavelength formalism (Wiener-Rosenblueth 1946, Allessie 1977) is the foundational organising principle of cardiac reentry.

Exercise 8 (hard, symbolic).

A single Einthoven lead measures the projection $V (t) = H (t) \cdot e$ of the heart vector $H (t)$ onto the lead unit vector $e$ . Suppose $H (t)$ rotates in the frontal plane during the QRS complex from initial direction $\hat{θ}_{0}$ to final direction $\hat{θ}_{1}$ with magnitude varying as $H (t)$ . Predict the shape of $V (t)$ for a lead with $e$ aligned (a) parallel to $\hat{θ}_{0}$ and perpendicular to $\hat{θ}_{1}$ , and (b) at the bisector $(\hat{θ}_{0} + \hat{θ}_{1}) /2$ .

Hint

Use $V (t) = H (t) cos (ϕ (t) - α_{lead})$ with $ϕ (t)$ the instantaneous heart-vector angle and $α_{lead}$ the lead angle.

Answer

Let $ϕ (t)$ interpolate from $\hat{θ}_{0}$ to $\hat{θ}_{1}$ , with the lead angle $α_{lead}$ fixed. Then $V (t) = H (t) cos (ϕ (t) - α_{lead})$ .

(a) With $α_{lead} = \hat{θ}_{0}$ : at $t = 0$ , $cos (0) = 1$ so $V$ is large and positive. As $ϕ (t)$ rotates by $90°$ to $\hat{θ}_{1}$ , $cos (90°) = 0$ so $V \to 0$ . The lead sees a monophasic positive deflection that grows from zero, peaks, and decays — the textbook QRS shape on a lead aligned with the initial activation vector. Many ECG leads in healthy patients show this monophasic morphology.

(b) With $α_{lead}$ at the bisector: $V$ is symmetric in $ϕ (t) - α_{lead}$ , peaking when $ϕ (t)$ crosses $α_{lead}$ (mid-QRS) and approaching zero at both endpoints (where $ϕ (t)$ is offset from $α_{lead}$ by half the rotation angle). The morphology is a symmetric peak — also a common QRS appearance on leads near the electrical axis. Leads at $α_{lead}$ perpendicular to the bisector (rotated by $90°$ ) see biphasic QRS complexes: initial negative deflection as $ϕ (t) - α_{lead}$ crosses $- 90°$ , then a positive deflection as it crosses $+ 90°$ . The textbook electrical-axis determination on the ECG uses exactly this geometry: the lead with the smallest net QRS deflection identifies the perpendicular to the mean QRS vector.

Exercise 9 (hard, symbolic).

Show that a Hopf bifurcation can occur in the minimal SA-node model of Exercise 4 as the maximal funny-current conductance $\overset{g}{ˉ}_{f}$ is decreased from a baseline value, with all other parameters fixed. Specifically: at large $\overset{g}{ˉ}_{f}$ the system has a stable limit cycle (the pacemaker oscillation); reducing $\overset{g}{ˉ}_{f}$ past a critical value the system transitions to a stable equilibrium (quiescent cell). Sketch the bifurcation diagram and discuss its clinical relevance.

Hint

Linearise around the equilibrium found by setting all derivatives to zero. The Jacobian eigenvalues depend on $\overset{g}{ˉ}_{f}$ ; the limit cycle exists when at least one pair of complex-conjugate eigenvalues has positive real part. The Hopf bifurcation occurs when the real part crosses zero.

Answer

At equilibrium, $V_{m}^{*} = (\overset{g}{ˉ}_{f} y_{\infty}^{*} E_{f} + \overset{g}{ˉ}_{C a} d_{\infty}^{*} E_{C a} + \overset{g}{ˉ}_{K, b} E_{K}) / (\overset{g}{ˉ}_{f} y_{\infty}^{*} + \overset{g}{ˉ}_{C a} d_{\infty}^{*} + \overset{g}{ˉ}_{K, b})$ with starred quantities evaluated at $V_{m}^{*}$ . The Jacobian of the three-variable system at $(V_{m}^{*}, y_{\infty} (V_{m}^{*}), d_{\infty} (V_{m}^{*}))$ has eigenvalues whose real parts depend continuously on $\overset{g}{ˉ}_{f}$ . At large $\overset{g}{ˉ}_{f}$ , the equilibrium is unstable (a spiral source) and a stable limit cycle surrounds it (the pacemaker oscillation). As $\overset{g}{ˉ}_{f}$ decreases, the spiral source becomes weaker; at a critical value $\overset{g}{ˉ}_{f, c}$ , the complex-conjugate eigenvalue pair crosses the imaginary axis from right to left (real part going from positive to negative), and the equilibrium becomes a stable spiral. This is a Hopf bifurcation 02.12.17: subcritical or supercritical depending on the cubic-coefficient of the normal form. Clinically, this models sick-sinus syndrome: progressive loss of HCN-channel function (e.g., from cardiomyopathy, ageing, mutations in HCN4) reduces $\overset{g}{ˉ}_{f}$ , the limit-cycle amplitude shrinks, the firing rate slows, and eventually the cell falls off the limit cycle into a quiescent equilibrium — clinical sinus arrest. The same Hopf-bifurcation pattern recurs in 17.09.02 pending for neural cells switched between silent and repetitively firing states, identifying cardiac pacemaking with the same dynamical-systems object that organises neural excitability.

Exercise 10 (hard, symbolic).

In two-dimensional cardiac tissue, a broken wavefront can curl into a stable rotating spiral (the Winfree-Keener rotor) whose tip traces out a small core. The rotor period $T_{rotor}$ is set by the local refractory period $τ_{R}$ and the conduction velocity at the wavefront, but unlike one-dimensional ring reentry, the rotor period can be shorter than the planar action-potential duration. Explain why, and identify what changes between the one-dimensional ring formula $L = θ τ_{R}$ and the two-dimensional spiral case.

Hint

Wavefront curvature changes the local conduction velocity through the eikonal-curvature relation $θ_{curved} = θ_{planar} - D κ$ , where $κ$ is the wavefront curvature and $D$ is the effective diffusion coefficient. Near the spiral tip, curvature is high and conduction velocity drops.

Answer

The eikonal-curvature relation $θ (κ) = θ_{0} - D κ$ (Tyson-Keener 1988) relates the propagation speed of a curved wavefront to the planar speed $θ_{0}$ and the wavefront curvature $κ$ (positive when the front bulges into excitable tissue). Near the spiral tip, $κ$ is large positive, so the local conduction velocity is reduced. The spiral rotates because the slow tip drags the rest of the front around it; the tip itself traces a small core whose size is set by the critical curvature at which $θ (κ) = 0$ (the front cannot propagate; this is the break-up curvature).

The rotor period is approximately $T_{rotor} \approx 2 π r_{core} / θ (κ = 1/ r_{core})$ where $r_{core}$ is the radius of the spiral-tip trajectory. Because the curvature lowers the local conduction velocity, the rotor can complete a revolution faster than a planar wave would traverse a path equal to the rotor's spatial wavelength — the rotor period is set by the curvature-modified dynamics rather than by the planar refractory period alone. This is why ventricular tachycardia driven by a single stable rotor can have a much higher rate than predicted by the planar wavelength formula.

In three dimensions, the spiral becomes a scroll wave with a one-dimensional filament tracing the axis of rotation. Filament instabilities (Winfree-Keener filament dynamics) — alternans-driven breakup, twisting, and re-fragmentation — generate the multiple-rotor turbulent state of ventricular fibrillation, where dozens of short-lived rotors coexist and the activation pattern becomes chaotic. The transition from ventricular tachycardia (one stable rotor) to ventricular fibrillation (multiple unstable rotors) is a fragmentation cascade in the excitable-medium PDE ^{[Winfree 1987]} ^{[Keener-Sneyd]}.

Lean formalization [Intermediate+]

Mathlib does not formalise cardiac action-potential models, the cable equation as a PDE on cardiac geometries, or the reaction-diffusion machinery underlying reentry. The closest existing infrastructure is:

Mathlib.Analysis.ODE.PicardLindelof: existence and uniqueness for ODEs satisfying a Lipschitz condition — the foundation for the cellular kinetic equations.
Mathlib.Analysis.ODE.Gronwall: the Grönwall inequality, relevant for proving sensitivity bounds on conductance-based models.
Emerging measure-theoretic infrastructure for Markov chains (relevant to single-channel stochastic gating).

There is no Mathlib definition of "conductance-based cardiac model", "funny current $I_{f}$ ", "bidomain equations", or "reentrant rotor". The path from existing ODE machinery to a verified cardiac action potential is laid out in the lean_mathlib_gap field of the frontmatter; the load-bearing primary gap is the reaction-diffusion PDE on a domain with anisotropic conductivity, which depends on infrastructure Mathlib has only in fragments. lean_status: none reflects this gap, and the unit ships reviewer-attested pending the cardiovascular-physiologist + clinical-electrophysiologist reviewer recruitment noted in the frontmatter.

Cardiac action-potential phases 0-4 and the ion-channel basis [Master]

The defining feature of the ventricular cardiac action potential is the plateau — a sustained depolarised phase of about 200-300 ms during which the membrane sits near $0$ mV, supplied by a dynamic balance of inward L-type calcium current $I_{Ca, L}$ and outward delayed-rectifier potassium currents $I_{Kr}$ and $I_{Ks}$ . This plateau is what distinguishes cardiac excitability from the neural Hodgkin-Huxley spike of 17.09.02 pending: where the squid axon spends less than a millisecond above threshold and snaps back, the ventricular myocyte commits to a roughly 250-ms refractory period during which it cannot re-excite. The plateau exists for two intertwined reasons. First, it allows time for calcium to enter the cell in bulk through L-type channels, triggering calcium-induced calcium release from the sarcoplasmic reticulum and driving the contractile machinery 18.04.02 pending. Second, it enforces an absolute refractory period long enough that the ventricle cannot tetanise — a critical safety feature for a pump.

Phase 0 (the upstroke) is dominated by the fast sodium current $I_{Na}$ in working myocardium (ventricular and atrial muscle, Purkinje fibres), with kinetics qualitatively similar to the squid-axon HH sodium current — fast activation gate $m$ , fast inactivation gate $h$ , plus a slow inactivation gate $j$ that gives ventricular sodium channels their characteristic two-component inactivation. The upstroke slope reaches $200$ - $300$ V/s, comparable to neural spike upstrokes. In nodal cells (SA, AV), $I_{Na}$ is sparse and $I_{Ca, L}$ is the principal upstroke current; the resulting upstroke is an order of magnitude slower (about $5$ - $10$ V/s), which is one ingredient of the slow conduction velocity through the AV node and of the AV delay that separates atrial from ventricular activation.

Phase 1 (the early-repolarisation notch) is the activation of the transient outward potassium current $I_{to}$ — Ito1, a 4-aminopyridine-sensitive Kv4 channel current — that briefly hyperpolarises the cell after the upstroke before the plateau settles in. The depth of the notch varies regionally: epicardial cells have a deep notch (prominent $I_{to}$ ), endocardial cells a shallow notch (less $I_{to}$ ); this regional difference contributes to the J-wave (Osborn wave) seen on the ECG in hypothermia and Brugada syndrome. Phase 2 (the plateau proper) is the long quasi-equilibrium between $I_{Ca, L}$ inward and $I_{Kr} / I_{Ks}$ outward. The plateau voltage is set self-consistently: at the equilibrium plateau voltage $V_{pl}$ , the net current is zero, $I_{Ca, L} (V_{pl}, d, f) + I_{Kr} (V_{pl}, x_{r}) + I_{Ks} (V_{pl}, x_{s}) = 0$ , with the gating variables themselves slowly evolving. Phase 3 (final repolarisation) begins as $I_{Ca, L}$ inactivates (voltage-driven $f$ closes; calcium-dependent inactivation $f_{C a}$ closes faster as intracellular calcium rises) and $I_{Kr} / I_{Ks}$ continue to activate; the balance shifts outward, the voltage falls, and as it falls $I_{K1}$ opens (rectifier with high conductance at hyperpolarised voltages) to clamp the cell back at the resting potential. Phase 4 (electrical diastole) is the stable resting voltage near $E_{K}$ in ventricular cells, with $I_{K1}$ the dominant resting conductance; in nodal cells, phase 4 is the diastolic depolarisation rather than a quiescent rest, as analysed below.

The refractory period of a cardiac myocyte is approximately the action-potential duration: during the plateau the sodium channels are inactivated ( $h$ and $j$ closed at depolarised voltages), so the cell cannot re-fire. The refractory period bounds the maximum firing rate at about $4$ Hz ( $240$ bpm) for normal ventricular cells, which is why ventricular tachycardia at much higher rates implies an underlying conduction abnormality (typically reentry). The plateau also gates the wavefront of any propagating wave: the wave can advance into cells in phase 4 (resting, excitable) but is blocked by cells in phases 1-3 (plateau, refractory). This refractory tail is the central ingredient of reentry, analysed in the conduction sub-section below.

The Beeler-Reuter 1977 model ^{[Beeler-Reuter 1977]} was the first to capture the plateau quantitatively in a Hodgkin-Huxley-style framework, with a slow inward current $i_{s}$ (calcium-dominated) and a time-dependent outward potassium current $i_{K}$ generating the plateau-and-repolarisation pattern. The Luo-Rudy 1991 model (LR1) ^{[Luo-Rudy 1994]} separated $i_{s}$ into L-type and T-type calcium components and the outward $i_{K}$ into rapid ( $I_{K r}$ ) and slow ( $I_{K s}$ ) delayed rectifiers, matching the experimentally resolved kinetics from voltage-clamp recordings of isolated ventricular myocytes (the late 1980s revolution in single-cell patch-clamp electrophysiology). LR2 (1994) added explicit intracellular calcium dynamics with the SERCA pump, the ryanodine-receptor release flux, and calcium buffers — making the model self-consistent on the calcium-transient timescale rather than treating calcium as an external parameter. Modern human-ventricular models (ten Tusscher-Noble-Noble-Panfilov 2004, O'Hara-Rudy 2011) extend this template with species-specific parameter fits and additional currents (the late sodium current $I_{N a L}$ , the small-conductance calcium-activated potassium current $I_{S K}$ ).

Pacemaker cells: the funny current and the coupled-clock model [Master]

SA-node and AV-node cells are autorhythmic: they fire repetitively without external input. The mechanism is the diastolic depolarisation (phase 4 depolarisation) — a steady upward drift of the membrane voltage from the maximum diastolic potential of about $- 60$ mV up to the firing threshold near $- 40$ mV, taking about $500$ ms at a baseline heart rate of $75$ bpm. The drift is supplied by the cooperative action of two coupled oscillators: a membrane clock of voltage-gated ionic currents and a calcium clock of intracellular calcium-release events.

The funny current $I_{f}$ is the defining ionic current of the membrane clock. Discovered by DiFrancesco in 1980-81 ^{[DiFrancesco-Noble 1985]} in Purkinje fibres and subsequently identified in SA-node cells, $I_{f}$ is an inward mixed-cation current (carrying both Na+ and K+ with relative permeability about $4 : 1$ ) flowing through HCN channels (HCN1, HCN2, HCN4 isoforms, with HCN4 dominant in SA-node). Its activation gate $y$ opens on hyperpolarisation — the opposite voltage dependence of the standard $m$ -gate of $I_{Na}$ — with half-activation $V_{1/2} \approx - 65$ mV and a slow time constant of order $100$ - $1000$ ms. The "funny" name reflects this unusual hyperpolarisation-activated behaviour. The reversal potential $E_{f} \approx - 25$ mV lies between $E_{K}$ and $E_{N a}$ because both cations flow through the channel, so opening $I_{f}$ at hyperpolarised diastolic voltages drives a net inward current that depolarises the cell. The current was simultaneously discovered as $I_{h}$ in neurons (Brown-DiFrancesco-Noble 1979 in cardiac fibres; Halliwell-Adams 1982 in hippocampal pyramidal neurons), and the molecular identification as HCN channels by the Yanagihara-Brown-Hofmann groups in the late 1990s confirmed the unified channel family.

The calcium clock, formalised by Maltsev and Lakatta in their 2009 coupled-clock model ^{[Maltsev-Lakatta 2009]}, is a parallel oscillator on the intracellular calcium dynamics: ryanodine receptors on the sarcoplasmic reticulum spontaneously release small calcium puffs during diastole (sub-sarcolemmal local calcium release events, LCRs), the calcium activates the Na+/Ca2+ exchanger which extrudes one Ca2+ in exchange for three Na+ (a net inward current, $I_{N a C a}$ , that further depolarises the cell), and the depolarisation accelerates voltage-dependent calcium-channel reactivation. The two clocks are coupled because the membrane-clock-driven depolarisation triggers L-type calcium influx (the upstroke), which refills the sarcoplasmic reticulum and primes the next round of calcium-clock release events; symmetrically, calcium-clock release events feed depolarising current into the membrane clock. The coupled system is more robust to perturbation than either clock alone — pharmacologic block of one clock (e.g., ivabradine blocks $I_{f}$ ; ryanodine blocks the calcium clock) degrades but does not abolish pacemaking, whereas block of both completely silences the SA node.

Autonomic modulation of heart rate works through both clocks. Sympathetic stimulation (noradrenaline → β1-adrenergic receptors → Gs → adenylyl cyclase → cAMP) directly binds cAMP to HCN channels, shifting the $y_{\infty}$ activation curve positively by about $10$ - $15$ mV; this increases $I_{f}$ at any given diastolic voltage, steepens the diastolic depolarisation, and accelerates heart rate (positive chronotropy). cAMP also accelerates the calcium clock by phosphorylating phospholamban (relieving inhibition of SERCA) and the ryanodine receptor (increasing its open probability). Parasympathetic stimulation (acetylcholine → muscarinic M2 receptors → Gi → reduced cAMP, plus activation of the inwardly rectifying potassium channel $I_{K, A C h}$ ) shifts $I_{f}$ activation negatively, slows the calcium clock, and opens $I_{K, A C h}$ to hyperpolarise the cell — three mechanisms together producing the bradycardia of vagal tone. The hyperpolarisation from $I_{K, A C h}$ also drives the maximum diastolic potential more negative, lengthening the climb to threshold.

The pharmacology of $I_{f}$ has clinical reach. Ivabradine, an HCN-channel blocker approved for chronic stable angina and heart failure with reduced ejection fraction, slows heart rate by selectively inhibiting $I_{f}$ without affecting blood pressure, contractility, or AV conduction — a selectivity profile no other rate-control drug matches. Mutations in HCN4 produce inherited sinus bradycardia, sick-sinus syndrome, and (rarely) atrial fibrillation, providing the human genetic evidence that $I_{f}$ is necessary for normal pacemaking. The slow, complex, redundant pacemaking mechanism — two clocks, multiple modulation pathways, cell-to-cell variability across the SA-node tissue — is the foundational reason the heart can change rate over a fivefold range in response to demand while never falling silent under physiological conditions.

Conduction, gap junctions, and reentry as a limit-cycle phenomenon [Master]

The heart is a coupled excitable medium: each myocyte is connected to its neighbours by gap junctions — protein pores formed by connexin proteins (Cx43 in working ventricular myocardium, Cx40 in atria, Cx45 in nodal tissue) that allow ions and small molecules to pass directly between cytoplasms. From the cable-equation perspective, each myocyte is a short capsule of active membrane ( $\sim 100$ μm long, $\sim 10$ μm wide) connected to its neighbours by an axial resistance $R_{g j}$ set by the gap-junction conductance. In the continuum limit (each myocyte short compared to the membrane length constant), the discrete cell chain is well approximated by the active cable equation of the cable theorem above; in the discrete limit (poor coupling, small gap-junction conductance), the wave propagates as a sequence of capsule-by-capsule firings whose timing is set by the local axial-current load.

The empirical conduction velocities in mammalian cardiac tissue are: ventricular myocardium $50$ - $80$ cm/s longitudinally, $17$ - $25$ cm/s transversely; atrial myocardium $80$ - $100$ cm/s; Purkinje fibres $2$ - $4$ m/s (the fastest in the heart, ensuring near-simultaneous ventricular activation); AV-node $5$ - $10$ cm/s (the slowest, producing the AV delay). The longitudinal-to-transverse anisotropy ratio of $\sim 3$ in ventricular myocardium reflects gap-junction concentration at intercalated discs (end-to-end contacts along the fibre axis); the slow AV-node conduction is a consequence of low connexin expression (Cx40 and Cx45 in nodal cells, with sparser gap junctions overall) and small-amplitude calcium-driven upstrokes. The Purkinje system is a fibre-network specialisation evolved to deliver the activation wave to the ventricular endocardium quickly enough that the wall depolarises within $\sim 80$ ms — fast enough to act as a near-synchronous unit but slow enough that the wave still has the locally cable-like character that determines QRS morphology.

Reentry — the central mechanism of most clinical tachyarrhythmias — is the closed-circuit propagation of an excitation wave around an inexcitable obstacle or through a region of slow conduction. Mines 1913 ^{[Mines 1913]} gave the original demonstration in ring preparations of fish atrium, identifying the three conditions: (i) a unidirectional block (the wave goes one way around the ring but not the other), (ii) slow conduction in the available pathway (so the wave returns to its starting point after the cells have recovered from refractoriness), and (iii) a critical relationship between the wavelength $λ_{w} = θ \cdot τ_{R}$ and the available pathway length $L$ . The condition for sustained reentry is $L \geq λ_{w}$ . Fibrosis, gap-junction uncoupling, and ischaemia all slow $θ$ , shortening $λ_{w}$ and allowing reentry on smaller circuits. Class-III antiarrhythmics (amiodarone, sotalol) extend $τ_{R}$ , increasing $λ_{w}$ and forcing the reentry circuit to lengthen or terminate.

In two-dimensional cardiac tissue, the reentrant wave is a spiral wave or rotor — a free-ended wavefront that curls around its own tip, the tip tracing a small core whose size is set by the critical curvature at which the eikonal-curvature relation $θ (κ) = θ_{0} - D κ$ predicts wavefront failure (Tyson-Keener 1988). The rotor period is determined by the local curvature dynamics near the tip, not by the planar refractory period alone, so rotor-driven tachycardia can run faster than the planar wavelength formula predicts. Winfree 1987 ^{[Winfree 1987]} gave the systematic theory of spirals and their three-dimensional generalisation (scroll waves with one-dimensional rotation-axis filaments). The transition from a single stable rotor (monomorphic ventricular tachycardia) to multiple unstable rotors (polymorphic ventricular tachycardia, then ventricular fibrillation) is a fragmentation cascade driven by alternans-induced wavefront break-up — when the action-potential duration starts to alternate beat-to-beat (a Hopf bifurcation in the restitution dynamics, Nolasco-Dahlen 1968), the spiral becomes unstable and breaks into daughter rotors. The Fenton-Karma 1998 reduction (a four-variable phenomenological model with adjustable restitution slope) is the canonical setting in which to study this fragmentation cleanly.

Defibrillation works by depolarising enough of the myocardium simultaneously to extinguish all reentrant wavefronts within a single global refractory period; an electric shock delivered through the chest creates a transmembrane voltage distribution that drives the entire ventricle into refractoriness at once, and the SA node — whose pacemaking is intrinsic and not dependent on a propagating wavefront — resumes normal rhythm a beat later. The bidomain model (Tung 1978, Henriquez 1993) is the standard mathematical framework for analysing defibrillation thresholds and electrode geometry; it tracks both intracellular and extracellular potentials with separate (anisotropic) conductivity tensors and couples them through the membrane current. The success of defibrillation is one of the clearest existence-and-stability results in clinical excitable-medium theory: an external boundary perturbation reliably resets the entire excitable-medium PDE to its homogeneous quiescent state, from which the autorhythmic SA node — operating on its own limit-cycle oscillation — can re-initiate normal rhythm.

ECG genesis: from cellular dipoles to body-surface vectorcardiography [Master]

The electrocardiogram measures voltage differences between electrodes on the body surface as a function of time. To connect this surface signal to the cellular action potentials, the standard framework is the equivalent-dipole approximation: at any instant $t$ , the entire electrical activity of the heart is represented by a single dipole (the heart vector $H (t)$ ) located at the centroid of activation, with magnitude and direction determined by the spatial distribution of transmembrane potentials. The voltage at a body-surface electrode is then approximately $V (t) = H (t) \cdot e_{lead} /∣ r ∣^{3}$ in a homogeneous-conductor approximation, with the actual torso geometry introducing corrections from the heterogeneous conductivity (lungs, ribs, fat).

The microscopic origin of the dipole is the moving boundary between depolarised and resting tissue. Each cell at the wavefront contributes a small contribution to the field; integrated over the activation surface, the contributions sum to a single net dipole oriented along the local activation gradient. When the entire ventricle is at the plateau (all cells depolarised together), there is no moving boundary and the heart vector vanishes — this is why the ST segment is electrically silent on the surface ECG even though the cellular action potentials are at their peak. When repolarisation begins, the moving boundary between still-depolarised cells (plateau) and already-repolarised cells (phase 4) generates a new dipole pointing in the direction of repolarisation propagation, which produces the T wave. Because epicardial cells repolarise earlier than endocardial cells (their plateaus are shorter), the T-wave dipole points from endocardium to epicardium — the same direction as the QRS dipole — and the T wave has the same polarity as the QRS in a normal ECG.

Einthoven's 1903 invention of the string galvanometer ^{[Einthoven 1903]} gave clinical medicine its first non-invasive recording of cardiac electrical activity, sensitive enough to capture the millivolt-scale potentials on the body surface and fast enough to resolve the P-QRS-T pattern at heart rates up to several hundred per minute. Einthoven 1924 received the Nobel Prize for the development. The three standard limb leads — Lead I (left arm − right arm, $0°$ ), Lead II (left leg − right arm, $60°$ ), Lead III (left leg − left arm, $120°$ ) — form Einthoven's triangle, a roughly equilateral triangle in the frontal plane with the heart at its centre. The voltage on each lead is the projection of the heart vector onto the corresponding lead axis; by Kirchhoff's voltage law applied to the triangle, Lead I + Lead III = Lead II (Einthoven's law), a constraint that all noise-free ECG recordings satisfy and that provides a real-time check for lead-misplacement artefacts.

The augmented leads (aVR, aVL, aVF) and the precordial leads (V1-V6) extend the geometry: aVR, aVL, aVF point from the limbs toward the heart at $- 150°$ , $- 30°$ , $+ 90°$ ; V1-V6 are arrayed across the anterior chest from the right of the sternum to the left axilla, with V1-V2 looking at the right ventricle and septum, V3-V4 at the anterior left ventricle, V5-V6 at the lateral left ventricle. Together the twelve leads sample the heart vector from many directions, allowing the mean QRS axis (the average direction of $H (t)$ during depolarisation) to be reconstructed in the frontal plane. Axis deviation — leftward (mean axis $< - 30°$ ) or rightward ( $> + 90°$ ) — corresponds to ventricular hypertrophy, bundle-branch blocks, or anatomical variants, each of which redistributes the activation centroid and thus the mean dipole direction.

Specific ECG morphologies trace back to specific cellular and tissue events: the QT interval (start of QRS to end of T) is the population-averaged action-potential duration of the ventricular myocardium and prolongs in long-QT syndromes (mutations of $I_{K r}$ , $I_{K s}$ , $I_{N a}$ that slow repolarisation); the PR interval is the time from atrial activation to ventricular activation and prolongs in AV-node disease (first-degree AV block); the width of the QRS reflects the synchrony of ventricular activation and widens when conduction through the bundle branches is blocked (left or right bundle-branch block, intraventricular conduction delay); the ST elevation of myocardial infarction reflects an injury current between infarcted and surrounding tissue, where the resting potential of injured cells is depolarised and creates a steady dipole during the otherwise-silent ST segment. The link from cellular pathology through the dipole representation to the surface ECG is the foundational diagnostic logic of clinical electrocardiography.

The macroscopic ECG can also be inverted to recover information about the local cellular state: body-surface mapping uses dozens of electrodes (typically 64-256) arranged in a grid over the torso and solves the inverse problem of the dipole field (with a torso-geometry model and a regularisation prior) to estimate the activation sequence on the epicardial surface. The inverse problem is ill-posed (the forward map from epicardium to torso surface is many-to-one), but with regularisation and prior structure (e.g., the activation surface is a connected wave moving at known speeds) the reconstruction is useful for non-invasive identification of arrhythmia foci. The technology is the modern descendant of Einthoven's string galvanometer, and the underlying mathematics is the same volume-conductor integral connecting cellular dipoles to body-surface voltages.

Synthesis. The cardiac action potential is the foundational atomic unit out of which heartbeat, conduction, and the ECG are constructed; the central insight is that three distinct cellular shapes (pacemaker, atrial, ventricular) coupled by gap junctions form a single excitable medium whose limit-cycle behaviour (the SA-node pacemaker) and traveling-wave behaviour (the conducted action potential) together generate the macroscopic electrical sequence. Putting these together with the dipole representation, the surface ECG identifies cellular electrophysiology with body-surface measurement through the volume-conductor integral, and the bridge is the moving boundary between depolarised and resting tissue. The plateau-driven calcium influx links cellular electrical excitation to mechanical contraction through the calcium-induced calcium release of 18.04.02 pending, and the bifurcation taxonomy generalises from neural Type-I/Type-II distinctions 17.09.02 pending to cardiac sinus arrest (Hopf in the SA-node oscillator), early afterdepolarisations (Hopf in the plateau dynamics), and reentry instability (alternans-driven Hopf at the tissue scale). The same dynamical-systems language that organises neural excitability appears again in 02.12.14 as the Liénard/Van der Pol limit-cycle theory and in 02.12.17 as the bifurcation hierarchy, identifying cardiac electrophysiology with the rest of mathematical excitability theory. The pattern recurs in every excitable medium that biology has invented — neural circuits, calcium waves in oocytes, cAMP waves in slime moulds — but the cardiac specialisation (the plateau, the funny current, the gap-junction-coupled syncytium) is the version most consequential for clinical medicine and the version with the deepest century-long pharmacology and electrophysiology tradition.

Full proof set [Master]

Proposition (anisotropy ratio for ventricular conduction velocity). For a cardiac monodomain model with bidirectional axial conductivity tensor $D = diag (D_{L}, D_{T})$ in coordinates aligned with local fibre orientation, propagating planar waves along the fibre and across the fibre have velocities $θ_{L} = D_{L} \cdot f (g)$ and $θ_{T} = D_{T} \cdot f (g)$ for a kinetic-dependent factor $f$ that is the same in both directions. The anisotropy ratio is therefore $θ_{L} / θ_{T} = D_{L} / D_{T}$ , independent of the ionic kinetics.

Proof. The monodomain reaction-diffusion PDE $$ C_m \partial_t V_m = \nabla \cdot (\mathbf{D} \nabla V_m) - I_{\rm ion}(V_m, \mathbf{g}) $$ admits planar traveling waves in any direction $\overset{n}{^}$ , with velocity $θ (\overset{n}{^})$ depending on the projection of $D$ onto $\overset{n}{^}$ . Substituting $V_{m} (x, t) = U (\overset{n}{^} \cdot x - θ t)$ and exploiting that the Laplacian of a planar wave reduces to $(\overset{n}{^}^{T} D \overset{n}{^}) \cdot U^{''} (ξ)$ in the rotated coordinate $ξ = \overset{n}{^} \cdot x - θ t$ , the wave equation becomes $$ (\hat n^T \mathbf{D} \hat n) U''(\xi) + \theta C_m U'(\xi) - I_{\rm ion}(U, \mathbf{g}) = 0. $$ Non-dimensionalising with $\hat{ξ} = ξ / \overset{n}{^}^{T} D \overset{n}{^} \cdot R_{m}$ and $\hat{t} = t / (R_{m} C_{m})$ removes the direction-dependence from the equation; the dimensionless eigenvalue $\hat{θ}$ depends only on the ionic kinetics. Dimensional restoration gives $θ (\overset{n}{^}) = (\overset{n}{^}^{T} D \overset{n}{^}) / R_{m} C_{m}^{2} \cdot \hat{θ}$ , i.e. $θ (\overset{n}{^}) = \overset{n}{^}^{T} D \overset{n}{^} \cdot f (g)$ with $f (g) = \hat{θ} / (C_{m} R_{m})$ . Along the fibre, $\overset{n}{^}^{T} D \overset{n}{^} = D_{L}$ ; across, $\overset{n}{^}^{T} D \overset{n}{^} = D_{T}$ . The ratio is $θ_{L} / θ_{T} = D_{L} / D_{T}$ , independent of $g$ . $□$

This anisotropy result is the foundational reason fibre-orientation reconstruction from diffusion-tensor MRI is useful for predicting reentry pathways: knowing $D$ on the patient-specific geometry determines $θ (\overset{n}{^})$ everywhere up to a single kinetic-dependent prefactor.

Proposition (wavelength criterion for sustained reentry on a ring). On a one-dimensional ring of myocardium of circumference $L$ , a single propagating excitation wave with conduction velocity $θ$ and refractory tail of duration $τ_{R}$ sustains stable reentry if and only if $L \geq λ_{w} := θ τ_{R}$ .

Proof. Necessity ( $L \geq λ_{w}$ required for sustained reentry). Suppose a wavefront at position $x_{0}$ at time $t_{0}$ propagates around the ring at velocity $θ$ . At time $t_{0} + L / θ$ , the wavefront has traveled distance $L$ and returned to its starting point. Behind the wavefront, the refractory tail extends a distance $θ τ_{R} = λ_{w}$ — cells within $λ_{w}$ of the trailing edge of the wavefront are still in refractoriness (phases 1-3) and inexcitable. For the returning wavefront at time $t_{0} + L / θ$ to find excitable tissue, the cells at $x_{0}$ must have completed their refractory period by then. This requires $L / θ \geq τ_{R}$ , i.e. $L \geq θ τ_{R} = λ_{w}$ .

Sufficiency ( $L \geq λ_{w}$ allows sustained reentry). Suppose $L > λ_{w}$ . The wavefront propagates at velocity $θ$ into tissue that has recovered (excitable), the action potential fires normally, the wave continues around the ring; the period of the reentrant rhythm is $T = L / θ > τ_{R}$ with an excitable gap of duration $T - τ_{R}$ separating the trailing refractory tail from the leading wavefront. The dynamics is stable to small perturbations because the excitable gap accommodates the perturbation without forcing the wavefront into refractory tissue. (Linearisation: define $ϕ (t) =$ phase of wavefront around the ring, $\dot{ϕ} = θ / L$ at steady state. Perturbations $δ ϕ$ decay because conduction velocity weakly increases with the time since previous excitation — the standard restitution argument.) At $L = λ_{w}$ exactly the gap closes; for $L < λ_{w}$ the wavefront enters refractory tissue, conduction fails, and reentry terminates. $□$

The wavelength criterion is the canonical reentry-vulnerability metric of clinical cardiac electrophysiology, and it connects directly to the pharmacology: class-I antiarrhythmics slow $θ$ (Na-channel blockade), class-III antiarrhythmics extend $τ_{R}$ (K-channel blockade), and class-IV antiarrhythmics slow AV-node conduction (Ca-channel blockade). Each class modifies $λ_{w}$ in a different direction, with corresponding clinical indications for the arrhythmias whose pathway lengths are matched by the resulting wavelength change.

Connections [Master]

The action potential — ionic basis 17.09.02 pending. The cellular peer unit for neuronal excitability: the squid-axon HH analysis of $I_{Na} / I_{K}$ , the cable equation, and the limit-cycle classification in $R^{4}$ . The cardiac unit builds above the neural framework rather than re-deriving it. Cardiac contributions distinct from the neural parent: the plateau (L-type calcium current and delayed-rectifier potassium), the funny current $I_{f}$ for autorhythmicity, the gap-junction-coupled tissue syncytium, and the body-surface ECG as a volume-conductor signature.
Cardiovascular physiology — the heart 18.02.01. The parent unit at the organ-and-haemodynamics level (cardiac cycle, pressure-volume loop, Frank-Starling, Windkessel). The current unit deepens the same chapter at the cellular-electrophysiology level: phases 0-4, pacemaker mechanism, conduction, ECG genesis. Cross-reference flows both ways — 18.02.01 cites the ECG events; 18.02.02 cites the cardiac cycle that the cellular events drive.
Muscle contraction — actin-myosin 18.04.02 pending. Cardiac excitation-contraction coupling links the cellular electrical events analysed here to the actin-myosin force generation analysed there. The plateau-driven calcium influx through L-type channels triggers calcium-induced calcium release from the sarcoplasmic reticulum, calcium binds troponin C, and the cross-bridge cycle proceeds. Cardiac muscle uses the same molecular contractile machinery as skeletal but with longer plateau-driven calcium transients and tighter coupling between electrical and mechanical events.
Resting membrane potential and ion channels 17.09.01. The cellular peer for resting-state biophysics: Nernst equation, Goldman-Hodgkin-Katz multi-ion resting potential, inward-rectifier $K^{+}$ channels. Cardiac myocytes inherit the resting-potential framework directly; the cardiac specialisation (L-type calcium, HCN funny-current channels, regional connexin distribution) lives on top of this foundation.
Limit cycles and Liénard / Van der Pol systems 02.12.14. The pacemaker oscillation is a limit cycle on the cellular state space; the relaxation-oscillation analysis of Liénard-type systems is the mathematical setting for the SA-node and AV-node dynamics. The coupled-clock pacemaker reduces to a two-time-scale Liénard system in the singular limit of fast membrane gating against slow calcium dynamics.
Bifurcation theory pointer 02.12.17. The transitions in cardiac excitability — sinus arrest from $I_{f}$ reduction, early afterdepolarisations from $I_{K r}$ block, alternans-driven wavefront break-up at the tissue scale — are organised by the bifurcation taxonomy. Hopf bifurcation in the cellular dynamics and at the tissue restitution level recurs as the load-bearing mechanism for clinically significant arrhythmia transitions.
Phase space, vector field, integral curve 02.12.01. The cardiac conductance-based model is a multi-dimensional autonomous flow ( $R^{N}$ for $N$ gating variables plus voltage plus intracellular concentrations) on which the action potential is a trajectory and the pacemaker oscillation is a closed orbit; the framework of integral curves and phase-space geometry applies directly.
Skeletal muscle physiology 18.04.01. Cardiac and skeletal muscle share the actin-myosin contractile apparatus but differ sharply in their excitation-contraction coupling architecture: the plateau-driven, calcium-induced calcium release analysed in this unit contrasts with the direct DHPR-RyR1 voltage-sensor mechanical coupling that organises skeletal-muscle ECC, and the skeletal peer catalogues the alternative coupling mode that the cardiac plateau replaces along with the differences in twitch versus tetanic mechanics that the long cardiac plateau enforces.

Historical & philosophical context [Master]

The mechanistic understanding of the cardiac action potential built on three foundational lineages over the twentieth century. The first is the neural Hodgkin-Huxley framework of 1952 ^{[Hodgkin & Huxley 1952]} from the squid giant axon, which gave the world the conductance-based modelling paradigm — voltage-dependent gating variables obeying first-order kinetics, ionic currents written as conductance times driving force, the full action potential predicted from voltage-clamp data alone. Hodgkin and Huxley received the 1963 Nobel Prize for physiology or medicine. The second lineage is the Oxford group around Denis Noble, whose 1962 paper ^{[Noble 1962]} adapted the HH framework to cardiac Purkinje fibres — the first quantitative cardiac action-potential model. Noble identified that a delayed-rectifier-like potassium current together with a slow inward current could generate the long plateau and the Purkinje pacemaker depolarisation, founding the field of computational cardiac electrophysiology that has run continuously from 1962 through the modern detailed cell-specific models. The DiFrancesco-Noble 1985 model ^{[DiFrancesco-Noble 1985]} introduced the funny current $I_{f}$ and the dynamic intracellular ion concentrations that all modern models inherit.

The third lineage is the clinical electrocardiography tradition founded by Willem Einthoven in Leiden. Einthoven's 1903 string galvanometer ^{[Einthoven 1903]} gave clinical medicine a non-invasive method for recording the body-surface ECG, sensitive to millivolt potentials and fast enough to resolve the QRS complex at heart rates up to several hundred per minute. Einthoven 1924 received the Nobel Prize for the discovery, and his nomenclature — the P, Q, R, S, T waves; the Einthoven triangle; Einthoven's law on lead voltages — remains canonical. The intervening century saw the development of the twelve-lead ECG (Wilson 1934, Goldberger 1942), exercise testing, Holter monitoring (Norman Holter 1949), and the modern body-surface mapping and electrocardiographic imaging technologies that close the loop between cellular electrophysiology and clinical diagnostics.

The mid-century convergence of these three lineages was the work of George Mines, Carl Wiggers, and the British and American electrophysiology schools. Mines 1913 ^{[Mines 1913]} showed that a circulating wave on a ring of cardiac tissue could sustain itself indefinitely, identifying reentry as a mechanism of arrhythmia distinct from automaticity. The reentry framework was extended to two-dimensional spiral waves by Wiener and Rosenblueth 1946 (in formal models) and by Allessie and colleagues 1973-1977 (in experimental atrial preparations), with the modern spiral-and-scroll-wave dynamics formalised by Arthur Winfree from the 1970s ^{[Winfree 1987]}. The 1980s and 1990s patch-clamp revolution — Neher and Sakmann's technique applied to isolated cardiac myocytes — separated the cardiac action-potential current ensemble into the L-type calcium, delayed-rectifier potassium, transient-outward, and inward-rectifier components that the modern Luo-Rudy and Ten-Tusscher-Panfilov models codify. The pharmacology and clinical electrophysiology of arrhythmia mapping, catheter ablation, and modern device therapies (pacemakers, implantable cardioverter-defibrillators, cardiac resynchronisation therapy) all rest on this cellular-tissue-organ framework whose three foundational layers were laid by Hodgkin-Huxley, by Noble's cardiac extension, and by Einthoven's clinical ECG.

Modern computational cardiology — patient-specific simulation of arrhythmia, in-silico drug testing for proarrhythmic risk (the CiPA initiative), atrial-fibrillation rotor mapping for catheter ablation — uses detailed conductance-based models embedded in patient-specific tissue geometries reconstructed from MRI. The framework is the foundational reason cardiac electrophysiology is the cellular-systems biology subfield with the deepest century-long mathematical-modelling tradition; the gap between cellular biophysics and clinical practice is shorter here than in any other organ system, and the mathematical-physics machinery (reaction-diffusion PDEs, limit cycles, bifurcation theory, spiral-wave dynamics) is the same machinery analysed for excitable media in chemistry, ecology, and neural circuits.

Bibliography [Master]

Primary literature.

Hodgkin, A. L. & Huxley, A. F., "A quantitative description of membrane current and its application to conduction and excitation in nerve", J. Physiol. 117 (1952), 500–544.

Noble, D., "A modification of the Hodgkin-Huxley equations applicable to Purkinje fibre action and pace-maker potentials", J. Physiol. 160 (1962), 317–352.

Beeler, G. W. & Reuter, H., "Reconstruction of the action potential of ventricular myocardial fibres", J. Physiol. 268 (1977), 177–210.

DiFrancesco, D. & Noble, D., "A model of cardiac electrical activity incorporating ionic pumps and concentration changes", Phil. Trans. R. Soc. B 307 (1985), 353–398.

Luo, C.-H. & Rudy, Y., "A model of the ventricular cardiac action potential. Depolarization, repolarization, and their interaction", Circ. Res. 68 (1991), 1501–1526.

Luo, C.-H. & Rudy, Y., "A dynamic model of the cardiac ventricular action potential. I. Simulations of ionic currents and concentration changes", Circ. Res. 74 (1994), 1071–1096.

Luo, C.-H. & Rudy, Y., "A dynamic model of the cardiac ventricular action potential. II. Afterdepolarizations, triggered activity, and potentiation", Circ. Res. 74 (1994), 1097–1113.

Mines, G. R., "On dynamic equilibrium in the heart", J. Physiol. 46 (1913), 349–383.

Einthoven, W., "Die galvanometrische Registrierung des menschlichen Elektrokardiogramms, zugleich eine Beurteilung der Anwendung des Capillarelektrometers in der Physiologie", Pflügers Arch. 99 (1903), 472–480.

Einthoven, W., Fahr, G. & de Waart, A., "Über die Richtung und die manifeste Größe der Potentialschwankungen im menschlichen Herzen und über den Einfluss der Herzlage auf die Form des Elektrokardiogramms", Pflügers Arch. 150 (1913), 275–315.

Maltsev, V. A. & Lakatta, E. G., "Synergism of coupled subsarcolemmal Ca2+ clocks and sarcolemmal voltage clocks confers robust and flexible pacemaker function in a novel pacemaker cell model", Am. J. Physiol. Heart Circ. Physiol. 296 (2009), H594–H615.

ten Tusscher, K. H. W. J., Noble, D., Noble, P. J. & Panfilov, A. V., "A model for human ventricular tissue", Am. J. Physiol. Heart Circ. Physiol. 286 (2004), H1573–H1589.

O'Hara, T., Virág, L., Varró, A. & Rudy, Y., "Simulation of the undiseased human cardiac ventricular action potential: model formulation and experimental validation", PLoS Comput. Biol. 7 (2011), e1002061.

Fenton, F. & Karma, A., "Vortex dynamics in three-dimensional continuous myocardium with fiber rotation: filament instability and fibrillation", Chaos 8 (1998), 20–47.

Tyson, J. J. & Keener, J. P., "Singular perturbation theory of traveling waves in excitable media (a review)", Physica D 32 (1988), 327–361.

Wiener, N. & Rosenblueth, A., "The mathematical formulation of the problem of conduction of impulses in a network of connected excitable elements, specifically in cardiac muscle", Arch. Inst. Cardiol. Mex. 16 (1946), 205–265.

Textbook and monograph.

Bers, D. M., Excitation-Contraction Coupling and Cardiac Contractile Force, 2nd ed. (Kluwer, 2001).

Zipes, D. P., Jalife, J. & Stevenson, W. G. (eds), Cardiac Electrophysiology: From Cell to Bedside, 7th ed. (Elsevier, 2017).

Boron, W. F. & Boulpaep, E. L., Medical Physiology, 3rd ed. (Elsevier, 2017).

Klabunde, R. E., Cardiovascular Physiology Concepts, 3rd ed. (Wolters Kluwer, 2021).

Berne, R. M. & Levy, M. N., Cardiovascular Physiology, 10th ed. (Mosby, 2008).

Katz, A. M., Physiology of the Heart, 6th ed. (Lippincott Williams & Wilkins, 2022).

Keener, J. P. & Sneyd, J., Mathematical Physiology, vol. II: Systems Physiology, 2nd ed. (Springer, 2009).

Winfree, A. T., When Time Breaks Down: The Three-Dimensional Dynamics of Electrochemical Waves and Cardiac Arrhythmias (Princeton University Press, 1987).

Goldberger, A. L., Clinical Electrocardiography: A Simplified Approach, 9th ed. (Elsevier, 2017).

Cycle 5 Track C — new production at math-style depth from line 1. Lane: chem-bio-phys; cardiac specialisation deepens the §18.02 chapter at the cellular electrophysiology level above the organ-and-haemodynamics peer 18.02.01. Cross-domain prereqs to the neural HH unit 17.09.02 pending, the cellular contractile unit 18.04.02 pending, and the ODE bifurcation pointer 02.12.17 route through Connections rather than prerequisites: because the bio peers are still status: draft in the chapter pending Tyler's review.

Prerequisites

02.12.01
02.12.14
02.12.17

Tier anchors

beginner: Crash Course Anatomy & Physiology (heart and cardiac conduction episodes); Khan Academy cardiovascular electrophysiology modules; Goldberger, *Clinical Electrocardiography: A Simplified Approach* (9th ed., Elsevier 2017) — beginner-accessible ECG primer
intermediate: Boron & Boulpaep, *Medical Physiology* (3rd ed., Elsevier 2017), Ch. 20 *The Heart as a Pump* and Ch. 21 *Cardiac Electrophysiology and the Electrocardiogram*; Klabunde, *Cardiovascular Physiology Concepts* (3rd ed., Wolters Kluwer 2021), Chs. 2-4; Berne & Levy, *Cardiovascular Physiology* (10th ed., Mosby 2008), Chs. 2-3
master: Bers, *Excitation-Contraction Coupling and Cardiac Contractile Force* (2nd ed., Kluwer 2001) — the canonical reference on cardiac calcium handling; Zipes, Jalife & Stevenson, *Cardiac Electrophysiology: From Cell to Bedside* (7th ed., Elsevier 2017); Noble 1962 (*J. Physiol.* 160, 317-352) — the first Purkinje action-potential model; DiFrancesco & Noble 1985 (*Phil. Trans. R. Soc. B* 307, 353-398) — the modern pacemaker model; Luo & Rudy 1991 (*Circ. Res.* 68, 1501-1526) and 1994 (*Circ. Res.* 74, 1071-1096; 1097-1113) — the dynamic ventricular model; Beeler & Reuter 1977 (*J. Physiol.* 268, 177-210); Maltsev & Lakatta 2009 (*Cardiovasc. Res.* 83, 367-374) — the coupled-clock pacemaker model

References

TODO_REF pending
Boron, W. F. & Boulpaep, E. L. — Medical Physiology, 3rd ed. (Elsevier, 2017) · Ch. 20 The Heart as a Pump; Ch. 21 Cardiac Electrophysiology and the Electrocardiogram — the integrated-physiology reference for cardiac action potentials, conduction, and ECG · see docs/catalogs/NEED_TO_SOURCE.md#bio-boron-boulpaep
TODO_REF pending
Klabunde, R. E. — Cardiovascular Physiology Concepts, 3rd ed. (Wolters Kluwer, 2021) · Chs. 2-4 on cardiac electrophysiology, pacemaker mechanisms, and the ECG; the standard concise pedagogical reference · see docs/catalogs/NEED_TO_SOURCE.md#bio-klabunde-cv
TODO_REF pending
Hodgkin, A. L. & Huxley, A. F. — A quantitative description of membrane current and its application to conduction and excitation in nerve · J. Physiol. 117 (1952) 500-544; the parent neural-AP model that all cardiac models extend · see docs/catalogs/NEED_TO_SOURCE.md#bio-wave1-hh1952
TODO_REF pending
Noble, D. — A modification of the Hodgkin-Huxley equations applicable to Purkinje fibre action and pace-maker potentials · J. Physiol. 160 (1962) 317-352; first cardiac Hodgkin-Huxley-style model — Purkinje fibres with a slow inward current and a delayed-rectifier-like potassium current capturing the plateau and pacemaker depolarisation · see docs/catalogs/NEED_TO_SOURCE.md#bio-noble-1962
TODO_REF pending
DiFrancesco, D. & Noble, D. — A model of cardiac electrical activity incorporating ionic pumps and concentration changes · Phil. Trans. R. Soc. B 307 (1985) 353-398; the modern Purkinje model — introduces the funny current $I_f$ (then $i_f$), the Na+/K+ pump, Na+/Ca2+ exchanger, and dynamic intracellular ion concentrations as state variables · see docs/catalogs/NEED_TO_SOURCE.md#bio-difrancesco-noble-1985
TODO_REF pending
Beeler, G. W. & Reuter, H. — Reconstruction of the action potential of ventricular myocardial fibres · J. Physiol. 268 (1977) 177-210; the first quantitative ventricular AP model with the slow inward calcium current driving the plateau · see docs/catalogs/NEED_TO_SOURCE.md#bio-beeler-reuter-1977
TODO_REF pending
Luo, C.-H. & Rudy, Y. — A model of the ventricular cardiac action potential · Circ. Res. 68 (1991) 1501-1526 (LR1 — depolarisation and early repolarisation); Circ. Res. 74 (1994) 1071-1096 and 1097-1113 (LR2 dynamic — full ionic concentrations, calcium-induced calcium release, calcium buffering); the canonical modern ventricular ionic model · see docs/catalogs/NEED_TO_SOURCE.md#bio-luo-rudy
TODO_REF pending
Einthoven, W. — The galvanometric registration of the human electrocardiogram, likewise a review of the use of the capillary-electrometer in physiology · Pflügers Archiv 99 (1903) 472-480; the string-galvanometer-based ECG. Einthoven 1924 Nobel Lecture (Nobel Foundation, Stockholm). See also Einthoven, Fahr & de Waart 1913 *Pflügers Arch.* 150, 275-315 for the triangle and lead derivations · see docs/catalogs/NEED_TO_SOURCE.md#bio-einthoven
TODO_REF pending
Maltsev, V. A. & Lakatta, E. G. — Synergism of coupled subsarcolemmal Ca2+ clocks and sarcolemmal voltage clocks confers robust and flexible pacemaker function in a novel pacemaker cell model · Am. J. Physiol. Heart Circ. Physiol. 296 (2009) H594-H615; companion review Cardiovasc. Res. 83 (2009) 367-374; the coupled-clock pacemaker formalisation · see docs/catalogs/NEED_TO_SOURCE.md#bio-maltsev-lakatta-2009
TODO_REF pending
Bers, D. M. — Excitation-Contraction Coupling and Cardiac Contractile Force, 2nd ed. (Kluwer, 2001) · Comprehensive reference on cardiac ion handling: L-type Ca2+ current, ryanodine-receptor calcium-induced calcium release, SERCA, Na+/Ca2+ exchanger, troponin C activation, and the link to mechanics · see docs/catalogs/NEED_TO_SOURCE.md#bio-bers-2001
TODO_REF pending
Zipes, D. P., Jalife, J. & Stevenson, W. G. (eds) — Cardiac Electrophysiology: From Cell to Bedside, 7th ed. (Elsevier, 2017) · Modern reference monograph spanning cellular electrophysiology, conduction, arrhythmia mechanisms, mapping, and clinical translation · see docs/catalogs/NEED_TO_SOURCE.md#bio-zipes-jalife
TODO_REF pending
Mines, G. R. — On dynamic equilibrium in the heart · J. Physiol. 46 (1913) 349-383; the originator paper for reentry as a mechanism of cardiac arrhythmia — a circulating excitation wave around an inexcitable obstacle · see docs/catalogs/NEED_TO_SOURCE.md#bio-mines-1913
TODO_REF pending
Winfree, A. T. — When Time Breaks Down: The Three-Dimensional Dynamics of Electrochemical Waves and Cardiac Arrhythmias (Princeton University Press, 1987) · Spiral and scroll-wave reentry; topological classification of cardiac arrhythmias as excitable-medium phenomena · see docs/catalogs/NEED_TO_SOURCE.md#bio-winfree-1987
TODO_REF pending
Keener, J. P. & Sneyd, J. — Mathematical Physiology, vol. II: Systems Physiology (2nd ed., Springer 2009) · Chs. 12-14 on cardiac electrophysiology, bidomain modelling, and reentry; the canonical mathematical-biology treatment with explicit conduction-velocity calculations and reentry analysis · see docs/catalogs/NEED_TO_SOURCE.md#bio-keener-sneyd
tong
raw/pdfs/mathbio/mathbio.pdf · Tong's *Mathematical Biology* lecture notes — background reference for the reaction-diffusion machinery underlying cardiac excitable-medium dynamics, the Liénard / FitzHugh-Nagumo reduction, and the limit-cycle / traveling-wave analyses used at master tier; readers focused on the physiology may skip

Reviewer

Tyler (pending external cardiovascular physiologist + clinical electrophysiologist reviewers per BIOLOGY_PLAN §7 — the cardiac-electrophysiology specialty is the next-priority recruit alongside the cellular-neuroscience reviewer on 17.09.02)

Estimated time

beginner: 18m
intermediate: 40m
master: 75m