Excitatory and Inhibitory Postsynaptic Potentials

When a neurotransmitter binds its postsynaptic receptor, an ion channel opens (directly, for ionotropic receptors, or indirectly via a G-protein cascade for metabotropic receptors). Ions then flow through that channel down their electrochemical gradients, and this ionic current changes the membrane potential — the resulting deflection is the postsynaptic potential. If the potential moves the membrane toward the threshold for action potential firing, it is called an excitatory postsynaptic potential, or EPSP. If it moves the membrane away from threshold, or otherwise makes firing less likely, it is an inhibitory postsynaptic potential, or IPSP. The crucial conceptual move is to stop thinking of "excitatory transmitter" and "inhibitory transmitter" as fixed labels and instead think of reversal potential, sometimes written Erev. The reversal potential of a synaptic conductance is the membrane voltage at which net current through that channel is zero — the voltage at which the electrical force pushing ions one way is exactly balanced by the chemical (concentration-gradient) force pushing them the other way. For a channel selective for a single ion species, the reversal potential equals that ion's Nernst equilibrium potential. For a channel permeable to more than one ion (like most excitatory ionotropic receptors, which pass both Na⁺ and K⁺), the reversal potential is a weighted average of the individual ionic equilibrium potentials, set by their relative permeabilities — essentially a synaptic-channel version of the Goldman–Hodgkin–Katz relationship. The sign and magnitude of a postsynaptic potential at any given moment is fixed by one simple relationship:

I = g(V_m − E_rev) where I is the synaptic current, g is the conductance (how many channels are open, scaled by single-channel conductance), V_m is the current membrane potential, and E_rev is the reversal potential of that conductance. Whenever V_m is more negative than E_rev, current flows inward (depolarizing, EPSP-like); whenever V_m is more positive than E_rev, current flows outward (hyperpolarizing, IPSP-like). This single equation is why the same transmitter, acting through the same receptor, can even reverse its apparent "sign" if the recording electrode artificially clamps the membrane potential past E_rev — a fact exploited directly in the defining experiments below.

Step 1 — Transmitter release and receptor binding. An action potential invading the presynaptic terminal opens voltage-gated Ca²⁺ channels; Ca²⁺ influx triggers fusion of synaptic vesicles and release of neurotransmitter into the synaptic cleft. The transmitter diffuses across the ~20 nm cleft and binds postsynaptic receptors.

Step 2 — Channel opening and ion selectivity. For fast ionotropic transmission, transmitter binding directly opens a channel that is part of the same protein complex as the receptor. The identity of that channel's selectivity filter determines which ions can pass:

Step 3 — Local current flow and the driving force. The instant the channel opens, the net direction of current is set by the driving force, (V_m − E_rev), exactly as in the equation above. At a typical neuronal resting potential of about −65 mV:

• An AMPA-receptor channel (E_rev ≈ 0 mV) has a driving force of about +65 mV pulling it toward depolarization — Na⁺ rushes in, K⁺ trickles out, net inward current, EPSP.
• A GABA_A channel (E_Cl ≈ −70 mV in a mature neuron) sits close to resting potential; the driving force is small (~ −5 mV), so Cl⁻ moves only slightly, but the conductance increase still "shunts" the membrane and opposes depolarizing currents — this is an IPSP even though the voltage change may be modest.
• A GABA_B–activated K⁺ channel (E_K ≈ −90 mV) has a driving force of about −25 mV at rest, driving K⁺ out and hyperpolarizing the cell — a slower, deeper IPSP.

Step 4 — Spread and summation. The local current spreads passively along the dendrite and soma according to the neuron's cable properties (length constant, time constant), decaying as it goes, and combines algebraically with all other simultaneous postsynaptic potentials at the axon hillock — the topic developed further in Synaptic Integration. An EPSP that reaches the hillock while still above threshold voltage will help trigger an action potential; an IPSP arriving simultaneously subtracts from that depolarization or "shunts" it by increasing local membrane conductance, making the membrane harder to move at all (divisive rather than purely subtractive inhibition).

Step 5 — Termination. The potential decays as transmitter is cleared from the cleft (diffusion, reuptake transporters, or enzymatic breakdown such as acetylcholinesterase) and channels close, and the membrane potential relaxes back toward rest through the neuron's passive properties and voltage-gated leak conductances.

The idea that a postsynaptic potential's sign is dictated by a reversal potential — rather than by the transmitter itself — was established through intracellular recording combined with imposed changes in membrane potential, most famously in the work of John Eccles, John Coombs, and Paul Fatt on cat spinal motoneurons in the mid-1950s.

The experiment. Eccles and colleagues impaled individual spinal motoneurons with fine glass microelectrodes filled with conducting salt solution — a major technical feat at the time, since motoneurons are only tens of micrometers across. They stimulated presynaptic afferent fibers (for example, group Ia sensory afferents from muscle spindles, which make monosynaptic excitatory connections onto motoneurons, or inhibitory interneurons activating disynaptic inhibitory pathways) while recording the resulting postsynaptic potential intracellularly. Critically, they then used the same electrode (or a second one) to inject steady current and artificially hold the neuron's resting membrane potential at different fixed values — depolarizing it or hyperpolarizing it away from its natural resting level — before repeating the synaptic stimulation.

The observation. For the excitatory (Ia-afferent) input, the EPSP amplitude shrank as the membrane was depolarized toward roughly 0 mV, disappeared entirely near 0 mV, and reversed in sign (became a downward, hyperpolarizing deflection) when the membrane was artificially held positive of that point. This behavior matched almost exactly what a cation channel permeable to both Na⁺ and K⁺ should do, since a mixed-cation conductance has a Nernst-weighted reversal potential near 0 mV. For the inhibitory pathway, Coombs, Eccles, and Fatt (1955) showed the IPSP reversed sign near −70 to −80 mV, and — crucially — that this reversal potential shifted in a predictable way when they altered the intracellular chloride concentration by injecting Cl⁻ or other anions directly into the cell through the recording electrode. Injecting chloride made the IPSP reversal potential more depolarized (consistent with raising [Cl⁻]_in and thus raising E_Cl by the Nernst equation); injecting impermeant anions or extruding chloride shifted it the other way.

What it proved. This was decisive on two counts. First, it showed directly that the same postsynaptic membrane, at different holding potentials, produces currents of opposite sign from the identical synaptic input — proving the potential's sign is a property of the driving force (V_m − E_rev), not an intrinsic property of the transmitter or the synapse. Second, by manipulating intracellular chloride and watching the reversal potential move exactly as the Nernst equation predicted, they identified chloride as the ion carrier of spinal inhibitory synaptic current, confirming that IPSPs are Nernst-governed chloride events rather than some separate "inhibitory substance" acting by a different logic. This work, alongside the contemporaneous voltage-clamp studies of the action potential by Hodgkin and Huxley, helped cement ionic-driving-force reasoning as the basic explanatory currency of cellular neurophysiology, and it earned Eccles a share of the 1963 Nobel Prize in Physiology or Medicine (with Hodgkin and Huxley) for these discoveries about the ionic mechanisms of excitation and inhibition. A complementary line of evidence came from pharmacological dissection: applying the GABA_A antagonist bicuculline or the glycine antagonist strychnine abolished IPSPs at central synapses, while substituting extracellular Cl⁻ with an impermeant anion abolished the IPSP's reversal near E_Cl and shifted it toward the substitute anion's own equilibrium potential — directly linking the specific channel protein and the specific permeant ion to the inhibitory potential, exactly as receptor pharmacology predicts.

Example 1 — Calculating a GABA_A IPSP reversal potential and predicting its sign. A cortical pyramidal neuron has:

Using the Nernst equation at 37 °C (RT/F ≈ 26.7 mV, and remembering Cl⁻ is a monovalent anion so the sign flips relative to a cation): E_Cl = −(26.7 mV) × ln([Cl⁻]_in / [Cl⁻]_out) E_Cl = −(26.7 mV) × ln(8/120) E_Cl = −(26.7 mV) × ln(0.0667) E_Cl = −(26.7 mV) × (−2.708) E_Cl ≈ +72.3 mV... Careful with the anion sign convention: for an anion X⁻, the Nernst equation is E_X = −(RT/F) ln([X]_in/[X]_out), which is algebraically identical to +(RT/F) ln([X]_out/[X]_in). Recomputing explicitly: E_Cl = (26.7 mV) × ln(120/8) = (26.7 mV) × ln(15) = (26.7 mV) × (2.708) ≈ +72.3 mV That sign looks wrong for a normal mature neuron — and it is deliberately so, to make a teaching point: immature neurons have high intracellular Cl⁻ (maintained by the importer NKCC1 rather than the exporter KCC2), so E_Cl actually sits positive of rest, and GABA_A activation is depolarizing (though usually still functionally inhibitory via shunting, since it does not reach action-potential threshold). Now compare the mature case, where KCC2 expression drives [Cl⁻]_in down to about 5 mM against a 120 mM extracellular concentration: E_Cl = (26.7 mV) × ln(120/5) = (26.7 mV) × ln(24) = (26.7 mV) × (3.178) ≈ −85 mV At a resting potential of −65 mV, the driving force is (−65) − (−85) = +20 mV, meaning Cl⁻ flows out of the cell when channels open (net outward negative-charge movement = hyperpolarizing current), producing a classical hyperpolarizing IPSP. This example demonstrates why the same GABA_A receptor produces opposite physiological effects across development — the transmitter and channel are unchanged, but the transmembrane chloride gradient, and hence E_Cl, has reversed.

Example 2 — Estimating the reversal potential of a mixed-cation EPSP channel. An AMPA receptor channel is permeable to both Na⁺ and K⁺, with a relative permeability ratio P_Na : P_K of about 1 : 1 (roughly true for AMPA receptors, unlike the highly K⁺-selective leak channels that set resting potential). Using representative mammalian neuron concentrations:

Apply a simplified GHK-style weighted average (since P_Na ≈ P_K, weights are ≈ 0.5 each): E_rev(AMPA) ≈ (26.7 mV) × ln[ (P_Na[Na]_out + P_K[K]_out) / (P_Na[Na]_in + P_K[K]_in) ] E_rev(AMPA) ≈ (26.7 mV) × ln[ (145 + 5) / (15 + 140) ] E_rev(AMPA) ≈ (26.7 mV) × ln(150/155) E_rev(AMPA) ≈ (26.7 mV) × (−0.0328) E_rev(AMPA) ≈ −0.9 mV, i.e., essentially 0 mV This matches experimental measurements of AMPA/nicotinic reversal potentials (close to 0 mV) and explains why, at a resting potential of −65 mV, the driving force is roughly +65 mV — a large depolarizing force, which is why single AMPA-mediated EPSPs, despite being only ~0.2–0.4 mV each at a single synapse, can sum rapidly toward threshold when many occur together, a phenomenon developed further in Synaptic Integration. If a clinical mutation altered the channel's Na⁺:K⁺ permeability ratio to, say, 3:1 (favoring Na⁺ entry more strongly), recompute: E_rev ≈ (26.7) × ln[(3×145 + 1×5)/(3×15 + 1×140)] = (26.7) × ln(440/185) = (26.7)(0.867) ≈ +23 mV A more Na⁺-selective channel would shift E_rev positive, increasing driving force at rest and producing larger EPSPs for the same conductance — illustrating how receptor stoichiometry, not just channel number, tunes synaptic strength.

The reversal-potential logic of EPSPs and IPSPs is the mechanistic foundation for a large fraction of modern neuropharmacology. Benzodiazepines and barbiturates enhance GABA_A receptor Cl⁻ conductance (increasing channel opening frequency or duration), deepening IPSPs and treating anxiety, seizures, and insomnia — their action makes sense only in terms of increasing g_Cl in the I = g(V_m − E_rev) relationship. Local anesthetics and some antiepileptics target excitatory Na⁺-dependent EPSPs indirectly by blocking voltage-gated Na⁺ channels needed to propagate the depolarization the EPSP initiates. Clinically, disruption of the Cl⁻ gradient that sets E_Cl is now recognized as a mechanism of disease: in some models of neuropathic pain and temporal lobe epilepsy, downregulation of the chloride exporter KCC2 in spinal or hippocampal neurons raises intracellular Cl⁻, shifting E_Cl positive and converting GABAergic input from hyperpolarizing to depolarizing or weakly effective inhibition — directly reproducing, pathologically, the immature-neuron condition described in Worked Example 1. Drugs that restore KCC2 function are an active area of therapeutic development for both conditions. In general anesthesia, propofol and etomidate potentiate GABA_A currents; in neonatal and pediatric anesthesiology, awareness that GABA can be excitatory in immature circuits has changed how some sedatives are used and studied. In basic neuroscience research, engineered optogenetic and chemogenetic tools are explicitly built around this logic — channelrhodopsin variants selected for cation permeability produce optically triggered EPSPs, while anion-conducting channelrhodopsins are used to produce optically triggered IPSPs, letting researchers impose excitation or inhibition on demand in behaving animals.

Ask the tutor: How do EPSPs and IPSPs actually combine at the axon hillock to decide whether a neuron fires?