Channel Currents & I–V Curves

How voltage and channel gating shape membrane currents

Click a current trace or I–V point to select a sweep. All plots share the channel colours; the voltage protocol is shared by A and B. Hover over controls and channel headings for explanations and model references; hover over graphs for numerical values.

How to use this simulation

Choose channels and a voltage-step family. Click an I–V point to highlight its current trace. A is teal and B is orange; stronger traces mark the selected voltage.

Currents are densities (µA/cm²); conductances are densities (mS/cm²). Positive current is outward. These are isolated channel currents, not their sum. No leak or capacitive current is included, and holding current is not subtracted.

Changing a channel restores its default conductance and reversal potential. All other parameter edits update automatically. The voltage increment may leave the final step below the specified upper limit.

Receptor populations begin closed; their corresponding agonist pulses begin at voltage-step onset. Conductance defaults outside HH are illustrative amplitude scales, not measured channel densities. The fixed kinetic rates are not temperature-matched across models. Desensitisation is omitted.

AMPAR: Ca²⁺-impermeable AMPARs use a reduced two-state model, ds/dt = α[Glu](1 − s) − βs, with α = 1.1 /mM/ms and β = 0.19 /ms. Current is ohmic at a fixed open fraction. Longer pulses recruit more receptors; End can be nearly zero after a brief pulse.

NMDAR: The same framework uses α = 0.072 /mM/ms and β = 0.0066 /ms, with instantaneous external Mg²⁺ block and sufficient co-agonist assumed. Try a 100 ms step to see slower decay; setting Mg²⁺ to zero makes its I–V curve linear.

CP-AMPAR: AMPAR kinetics are multiplied by the equilibrium polyamine-block factor 1/[1 + (PA/20 µM)exp(V/15 mV)]. The 20 µM and 15 mV parameters were chosen for teaching, not fitted to experimental data. The cited paper supports the block mechanism, not these numerical choices. This curve illustrates rectification but does not predict its precise strength in a particular receptor subtype. Outside −100 to +50 mV the curve is an extrapolation; high-voltage relief and use dependence are omitted. Zero polyamine reproduces AMPAR at matching settings. Current is total cation current, not Ca²⁺ influx.

nAChR: Muscle-type C ↔ AC ↔ A₂C ↔ O populations follow a reduced adaptation of Land et al. (1981), with their own ACh pulse. Kinetics are fixed across voltages; diffusion, acetylcholinesterase and ACh block are omitted. This is not a generic neuronal nAChR model.

GABAₐ: A separate GABA pulse drives the two-state model with α = 5 /mM/ms and β = 0.18 /ms (Destexhe et al., 1998). Its linear I–V curve crosses the effective EGABA (default −70 mV), set using Erev. Chloride and bicarbonate fluxes are not separated. Move the reversal potential to explore current direction; this voltage-clamp model does not demonstrate shunting inhibition or determine whether GABA increases or decreases firing.

Kir2.1: Uses the published two-mode steady-state spermine-block fit of Ishihara & Ehara (2004), with parameters confirmed by Ishihara & Yan (2007). High- and low-affinity fractions are 0.9 and 0.1; their dissociation constants are 0.7 exp(−u/4.8) and 40 exp(−u/9.1) µM, where u = V − EK in mV. Spermine is fixed at 5 µM. These block parameters come from experimental fits.

Kir2.1 limits: The original fit used mouse channels in inside-out HEK293T patches at 24–26 °C with symmetric high K⁺ and no Mg²⁺ block. We shift its voltage dependence with EK, following the parallel shifts reported in 2007; this does not reproduce all effects of changing ion concentrations. Block is applied instantaneously at equilibrium, so Peak and End match: these traces do not predict block/unblock transients or the approach to steady state during short pulses. Try Start = −120 mV to see inward current below EK (default −90 mV).

HCN1: The ten-state population model uses Altomare et al. (2001), Table 2 parameters. Hyperpolarisation activates the channel; holding-voltage equilibrium sets its initial state. Use HCN protocol applies a 1000 ms step family from −35 mV. Step durations up to 5000 ms are available. Peak can select current through channels already open before the step; End reports the final sampled current, which need not have reached equilibrium.

HCN1 limits: These are room-temperature mouse HCN1 kinetics from an expression system, not all native Iₕ. cAMP modulation is not varied. Current uses a linear mixed-cation driving force with adjustable Erev (−30 mV teaching default); Na⁺/K⁺ concentrations are not calculated. Tail display is only 5 ms, so slow deactivation is truncated. This voltage-clamp model does not generate voltage sag or pacemaking.

For long sweeps, samples remain 0.01 ms apart through the first 200 ms and up to 1 ms apart afterwards. Downloads contain the actual, nonuniform sample times. Zoom in on early times when comparing slow HCN1 with fast HH currents.

HH Na⁺ and HH K⁺: Classical squid-axon kinetics at 6.3 °C: INa = gNam³h(V − ENa) and IK = gKn⁴(V − EK). Each sweep starts with gates at holding-voltage equilibrium. These are not generic mammalian channels. Reversal potentials are adjustable; ion concentrations are not calculated.

References