Voltage Clamp & Current Clamp

Think of cruise control in a car: you pick a speed and the car automatically presses or eases the gas to hold that speed up every hill. A voltage clamp does the same for a neuron's "charge level" — its membrane voltage. You choose a voltage, and an electronic helper continuously adds or removes electric current to keep the membrane frozen at exactly that value. Because it works so hard to hold the voltage still, the amount of current it has to supply tells you precisely how much current the cell's ion channels are passing. A current clamp is the opposite: you set the current you inject and let the voltage move freely — that is how you watch a real spike happen.

Two quantities do the work. The command voltage $V_{cmd}$ is the value you dial in. The clamp current $I_{clamp}$ is what the amplifier supplies. The rule is simply that the amplifier injects just enough current to cancel whatever the channels do, so $I_{clamp}=-I_{ionic}(V_{cmd})$. Hold a cell at $V_{cmd}=-40$ mV and suppose its sodium channels pour in $2$ nA of inward current; the clamp must pull $2$ nA back out to stop the voltage from moving — and that $2$ nA reading is your measurement of the channel current. No electrode is perfect, though: a small series (access) resistance $R_s$ sits between the wire and the cell, so any current $I$ creates a voltage error $V_{err}=I\,R_s$. The bigger the current, the more the true membrane voltage drifts away from the value you set.

A neuron is not a single point. Its dendrites stretch far from the soma where the electrode sits, and voltage leaks as it spreads: $V(x)\approx V_{soma}\,e^{-x/\lambda}$, where the space constant $\lambda$ is the distance over which the voltage falls to about $37\%$. So a somatic "voltage clamp" really only clamps the soma — distant dendrites settle to their own voltages and contaminate the measurement. This is the space-clamp problem, and it is why electrotonically compact cells (electrotonic length $L<0.3$) give the cleanest recordings. Modern patch clamp (Neher & Sakmann, Nobel 1991) presses a glass pipette against the membrane to form a gigaohm "gigaseal," so almost no current leaks around the electrode and both clamp modes become far more accurate. On the controls here, read Input I as the command level you impose, Parameter 1 as the series resistance $R_s$, and T sim as how long you record after a voltage step.

Use the sliders as the experiment's knobs and reason through each move. Raise Input I (the command level): the ion channels respond more strongly, so the clamp current must grow to cancel them — a larger reading. Now raise Parameter 1 ($R_s$): the same current now drops more voltage across the electrode, $V_{err}=I\,R_s$ climbs, and the cell is no longer truly held at $V_{cmd}$ — exactly the error a gigaseal minimises. Picture moving the recording electrode out along a dendrite and the $e^{-x/\lambda}$ decay means a step you command at the soma barely reaches the tip — that is space clamp failing. Lengthen T sim to give the clamp time to settle after each step.

Feedback amplifier: compare V_membrane to V_command → inject I to minimise difference. I_clamp = -I_ionic at V_cmd. Hodgkin & Huxley used axial wire voltage clamp in giant squid axon. Modern: patch clamp (Neher & Sakmann, Nobel 1991). Whole-cell: G_seal > 1 GΩ required.

Cell-attached → whole-cell (rupture) → inside-out (rip). Or: whole-cell → outside-out (pull). Each exposes different membrane face to controlled solutions. Single-channel: cell-attached or excised patch. Whole-cell: action potentials, synaptic currents. Outside-out: ligand-gated channels with fast solution exchange.

Somatic VC controls only soma V. Dendrites: V_dend(x) ≈ V_soma × e^{-x/λ}. Distal dendrites remain poorly clamped → contaminating currents. Solution: electrotonically compact cells (L<0.3), outside-out patches, or compartmental corrections (Rall 1967). Ignored space clamp → systematic errors in conductance estimates.