Epilepsy and Neuronal Hyperexcitability

A seizure is a transient, abnormal episode of excessive, hypersynchronous neuronal activity, detectable as a distinctive electroencephalographic (EEG) signature and often, though not always, accompanied by a clinical event — motor, sensory, autonomic, or purely cognitive. Epilepsy is the chronic disease state defined by an enduring predisposition to generate such seizures, together with the neurobiological, cognitive, and psychosocial consequences of that predisposition; operationally, a diagnosis requires two unprovoked seizures more than 24 hours apart (or one seizure plus a high recurrence risk). This distinction matters: a single seizure provoked by fever, hypoglycemia, or alcohol withdrawal reflects a transiently perturbed but otherwise normal brain, whereas epilepsy reflects a durably altered circuit. Seizures are classified along two orthogonal axes that map directly onto mechanism. The first axis is onset: a focal seizure begins in a circumscribed region of one hemisphere, while a generalized seizure engages bilateral, distributed thalamocortical networks essentially from onset. The second axis is awareness and semiology: focal seizures can occur with retained or impaired awareness, and generalized seizures include convulsive forms (tonic-clonic) as well as the almost purely electrographic absence seizures. Critically, focal and generalized seizures are generated by different circuit mechanisms — a fact that dictates both which animal models are appropriate and which drugs will work, since a drug that suppresses focal cortical bursting (like a sodium channel blocker) can be ineffective, or even worsen, absence seizures. The unifying mechanistic concept across all seizure types is a breakdown of the normal excitation-inhibition (E/I) balance. In the healthy cortex, fast glutamatergic excitation (via AMPA and NMDA receptors) is held in check by fast GABAergic inhibition (via GABA_A receptors), a balance enforced moment-to-moment by inhibitory interneurons — particularly fast-spiking, parvalbumin-positive basket cells that deliver powerful perisomatic inhibition directly onto the soma and axon initial segment of pyramidal neurons, the site of action potential initiation. Anything that tips this balance — loss of interneurons, reduced GABA_A receptor function, enhanced NMDA receptor signaling, altered intrinsic ion channel properties, or aberrant gap-junction/electrical coupling — can convert a normally-functioning local circuit into one capable of generating synchronized, self-sustaining discharges.

Step 1 — the single-neuron signature: the paroxysmal depolarizing shift. Intracellular recordings from epileptic cortex (originally in animal models, later confirmed in human tissue removed during epilepsy surgery) reveal a stereotyped event in individual pyramidal neurons during interictal spikes and seizure onset: the paroxysmal depolarizing shift (PDS). This is a large, sudden depolarization of 20–40 mV, well beyond what a normal EPSP produces, that plateaus for tens to hundreds of milliseconds and is crowned by a high-frequency burst of Na⁺-dependent action potentials, before the neuron drops into a prolonged afterhyperpolarization. The PDS is not simply a "big EPSP"; it results from a regenerative, largely voltage-gated calcium channel-driven plateau (particularly through dendritic high-threshold Ca²⁺ channels), amplified by synaptically-triggered NMDA receptor currents whose Mg²⁺ block is relieved once the cell depolarizes past roughly −40 mV. This amplification is the same Mg²⁺-unblock nonlinearity central to Excitotoxicity and Glutamate-Mediated Cell Death — during status epilepticus, the same sustained Ca²⁺ influx that drives the PDS can trigger excitotoxic cascades if the seizure is prolonged.

Step 2 — from single neuron to local synchrony. A single neuron's PDS is contained, harmlessly, if surrounding inhibitory interneurons deliver strong, well-timed feedback and feedforward inhibition — visible in surrounding cells as a large IPSP that follows each PDS, called surround inhibition. Epileptiform activity emerges when this surround inhibition fails or is overwhelmed: through loss of GABAergic interneurons (as in hippocampal sclerosis, where somatostatin- and parvalbumin-positive interneurons are selectively vulnerable), through GABA_A receptor subunit changes that reduce chloride conductance, through depolarizing shifts in the chloride reversal potential (via altered KCC2/NKCC1 cotransporter expression, which can make GABA_A currents paradoxically depolarizing in immature or injured tissue), or through recurrent excitatory connections strengthened by axonal sprouting. Once surround inhibition is breached, a cluster of neurons begins firing PDS events synchronously, producing the interictal spike seen on EEG or local field potential recordings — a marker of an epileptic focus that is not itself a seizure, but a sign of the underlying hyperexcitable circuit.

Step 3 — the transition to a seizure: from interictal spike to ictal discharge. Most interictal spikes remain isolated events, self-terminated by strong feedback inhibition and after-hyperpolarizing potassium currents. A seizure begins when this self-limitation fails and a critical mass of synchronized neurons instead recruits its neighbors progressively, through several amplifying mechanisms operating together: (1) extracellular K⁺ accumulates during intense firing, depolarizing neighboring neurons and glia (astrocytic K⁺ buffering can become saturated); (2) repetitive firing depletes presynaptic GABA release and desensitizes GABA_A receptors, causing activity-dependent disinhibition; (3) NMDA receptor Mg²⁺ block is progressively relieved across a widening region as extracellular depolarization spreads; (4) electrical (gap-junction) coupling between neurons and astrocytes can synchronize subthreshold oscillations. The result is a self-sustaining, propagating wave of hypersynchronous firing — the electrographic seizure — which eventually terminates through mechanisms still only partly understood: exhaustion of energy substrates, activation of slow K⁺ conductances, adenosine release (an endogenous anticonvulsant), and recovery of GABAergic inhibition.

Step 4 — a distinct circuit mechanism for generalized absence seizures. Absence seizures illustrate that hyperexcitability need not mean "too little inhibition" in the simple sense — it can instead mean an intrinsically oscillatory circuit tuned into a pathological resonance. The relevant circuit is the thalamocortical loop: cortical pyramidal neurons excite thalamic relay neurons and the GABAergic reticular thalamic nucleus (RTN); RTN neurons in turn inhibit thalamic relay neurons via GABA_A and, importantly, GABA_B receptors, and inhibit each other. This loop normally generates the sleep spindle, a brief 7–14 Hz oscillation, through a well-characterized cellular mechanism: GABA_B-mediated IPSPs in relay neurons are slow and hyperpolarize the cell enough to de-inactivate low-threshold T-type Ca²⁺ channels; when the IPSP decays, the T-current triggers a rebound low-threshold Ca²⁺ spike crowned by a Na⁺ burst, which re-excites RTN and cortex, restarting the cycle. In absence epilepsy, this normal spindle-generating mechanism becomes pathologically amplified and slowed to the characteristic 3 Hz spike-and-wave discharge, through exaggerated T-type current (from genetic channel variants or altered subunit expression), enhanced GABA_B signaling in relay neurons, and reduced fast GABA_A-mediated inhibition within RTN itself — producing a self-sustaining 3 Hz oscillation engaging the whole cortex bilaterally and synchronously, consistent with the sudden, generalized "absence" of behavior with no local seizure focus.

Griffith and Penfield's clinical mapping — establishing that seizures originate from focal cortical regions. Long before molecular mechanisms were known, neurosurgeon Wilder Penfield's intraoperative cortical stimulation studies (1930s–1950s) at the Montreal Neurological Institute showed that electrically stimulating specific cortical points in awake patients could reproduce a patient's habitual seizure aura or automatism, and that resecting the corresponding tissue often abolished seizures. This established, empirically, that many epilepsies arise from a discrete, localizable region of abnormal cortex — the epileptogenic focus — a foundational clinical fact that later cellular work would explain mechanistically.

The penicillin focus model — establishing disinhibition as a seizure mechanism. In a series of experiments beginning in the 1970s, researchers including H. Prince, D. Prince, and colleagues applied topical penicillin to the exposed cortex of animals and recorded intracellularly from underlying neurons. Penicillin, it was discovered, is a GABA_A receptor antagonist (blocking the chloride channel), and its local application reliably converts normal cortex into an epileptic focus generating spontaneous interictal spikes, each corresponding intracellularly to a PDS. This was decisive evidence, by direct pharmacological manipulation rather than correlation, that removing GABAergic inhibition is sufficient to convert normal cortical circuitry into a hyperexcitable, spike-generating focus — directly supporting the disinhibition mechanism described above. Complementary experiments using other convulsants acting at different targets (bicuculline, another GABA_A antagonist; picrotoxin, a channel blocker; 4-aminopyridine, a K⁺ channel blocker enhancing glutamate release) converged on the same conclusion from independent pharmacological routes.

Goddard's kindling model — establishing that a normal brain can be made permanently hyperexcitable through repeated stimulation. In 1969, Graham Goddard and colleagues discovered that daily brief, subconvulsive electrical stimulations of the amygdala or hippocampus in rats produced no seizure on the first several days, but progressively lowered the seizure threshold until, after repeated stimulations, the identical stimulus produced a full generalized tonic-clonic seizure — and this heightened sensitivity persisted permanently, even after months without stimulation. This phenomenon, called kindling, proved two crucial points: first, that epilepsy can be an acquired, activity-dependent property of an initially normal circuit (implicating synaptic plasticity mechanisms overlapping with LTP, including NMDA receptor-dependent strengthening of excitatory synapses and loss of inhibitory interneurons); second, that epileptogenesis — the process of becoming epileptic — is dissociable from the seizures themselves, since kindling represents a slow, cumulative circuit remodeling triggered by, but outlasting, the initiating stimulations. Kindling remains one of the most widely used models for testing new anticonvulsant drugs and for studying epileptogenesis after brain injury.

Genetic channelopathy mice — establishing a causal, single-gene basis for generalized absence epilepsy. The tottering, lethargic, and stargazer mouse mutants, and later engineered knock-in mice carrying human epilepsy mutations in T-type calcium channel genes (CACNA1H) or in the GABA_A receptor gene GABRG2, spontaneously display 3 Hz spike-and-wave discharges on EEG closely resembling human absence seizures, together with behavioral arrest. Because these are single-gene mutations with a defined biophysical consequence — for example, gain-of-function mutations that increase T-type Ca²⁺ current density and speed channel recovery from inactivation — these models provided direct causal proof, rather than correlation, that exaggerated T-type current in the thalamocortical circuit is sufficient to generate spike-and-wave absence seizures, closing the causal loop from single-channel biophysics to whole-circuit oscillation to behavior.

Example 1 — reasoning through a channelopathy pedigree (Dravet syndrome). A child presents at 6 months with prolonged febrile seizures, and by age 2 has developed multiple seizure types (myoclonic, atonic, focal) refractory to standard sodium-channel-blocking drugs, alongside developmental regression. Genetic testing reveals a de novo heterozygous loss-of-function mutation in SCN1A, encoding the α-subunit of the voltage-gated Na⁺ channel Na_v1.1. Reasoning through the mechanism: Na_v1.1 is expressed predominantly in GABAergic interneurons, particularly fast-spiking parvalbumin-positive cells, rather than in excitatory pyramidal neurons. A loss-of-function mutation therefore selectively impairs the ability of inhibitory interneurons to fire high-frequency trains, producing a net disinhibition of cortical circuits — paradoxically, a sodium channel loss-of-function mutation causes hyperexcitability, because the channel loss disproportionately cripples the inhibitory cells that normally restrain the network. This explains two clinically important facts: first, why standard Na⁺ channel blockers (carbamazepine, phenytoin, lamotrigine) often worsen Dravet syndrome — they further impair the already-compromised residual firing of interneurons, tipping the balance further toward disinhibition — and second, why drugs that enhance GABAergic transmission directly (such as valproate, clobazam, or the newer fenfluramine and stiripentol, which respectively enhance serotonergic modulation and directly potentiate GABA_A receptors) are preferred. This case illustrates why matching drug mechanism to the specific circuit lesion, not just to "seizures" as a generic category, is essential.

Example 2 — quantifying a dose-response decision for a Na⁺ channel blocker. A drug that blocks voltage-gated Na⁺ channels preferentially in their inactivated state — the defining property of most first-line anticonvulsants (phenytoin, carbamazepine, lamotrigine) — will block channels more effectively in neurons firing at high frequency, since inactivated channels accumulate during rapid depolarization; this is use-dependent block. Suppose in a voltage-clamp experiment a candidate drug produces 10% blockade of Na⁺ current when neurons fire at a normal physiological rate of 1 Hz, but 70% blockade when the same neurons are driven to fire at a pathological seizure-like rate of 50 Hz. Using the simple logic of state-dependent binding: at 1 Hz, channels spend most of their time in the resting, closed state to which the drug binds weakly, so minimal blockade accumulates between spikes as channels recover; at 50 Hz, the inter-spike interval (20 ms) is too short for full recovery from inactivation between action potentials, so channels accumulate in the drug-preferred inactivated state, and blockade builds up over successive spikes. The therapeutic implication follows directly from this biophysics: such a drug will preferentially suppress pathologically high-frequency, hypersynchronous firing (a seizure focus) while sparing the low-frequency firing that carries normal information — explaining why phenytoin can abolish seizures at doses that leave ordinary neurological function largely intact, and why efficacy is fundamentally a property of firing-rate-dependent pharmacology, not simple mass-action blockade.

Anticonvulsant (more precisely, anti-seizure) drugs remain the mainstay of epilepsy treatment for roughly two-thirds of patients whose seizures are controlled by medication, and each major drug class traces directly to the mechanisms above. Use-dependent Na⁺ channel blockers (phenytoin, carbamazepine, lamotrigine, oxcarbazepine) suppress high-frequency focal firing while sparing normal activity. GABA_A receptor potentiators (benzodiazepines, barbiturates) directly restore or boost chloride-mediated inhibition, making them first-line for acute status epilepticus. T-type Ca²⁺ channel blockers (ethosuximide) are specifically effective — and specifically only effective — against absence seizures, a direct pharmacological confirmation of the thalamocortical T-current mechanism. SV2A ligands (levetiracetam, brivaracetam) modulate synaptic vesicle release machinery, dampening excessive glutamate release during high-frequency firing through a mechanism distinct from classical channel block. Newer agents targeting the disinhibition mechanism directly, such as fenfluramine for Dravet syndrome, exemplify precision matching of drug mechanism to circuit lesion. Beyond pharmacology, the E/I-imbalance framework guides epilepsy surgery (resecting or disconnecting a well-localized focus identified by scalp and intracranial EEG, guided by the imaging methods that extend naturally from this material — see Brain Imaging Methods: Structural and Functional), responsive neurostimulation devices that detect an incipient seizure electrographically and deliver interrupting stimulation before clinical onset, and the ketogenic diet, whose anticonvulsant effect is thought to involve altered neuronal metabolism and potassium channel (K_ATP) activation that stabilizes membrane potential. The kindling model continues to underpin the search for genuinely antiepileptogenic therapies — drugs that would prevent epilepsy from developing after a brain insult such as traumatic injury or prolonged febrile seizures, rather than merely suppressing seizures once the disease is established, which remains one of the field's central unmet goals.

Ask the tutor: How does status epilepticus transition from a self-limited seizure into a sustained, treatment-resistant state, and why does prolonged benzodiazepine exposure lose efficacy the longer a seizure continues?