Ictal EEG Patterns

An interictal discharge is a snapshot; an electrographic seizure is a film. The defining property of an ictal pattern is not the appearance of any particular waveform but its dynamic evolution over seconds to minutes. A pattern that begins, builds, transforms, spreads, and then terminates with a recognizable aftermath is an electrographic seizure; a pattern that simply persists unchanged, however rhythmic or sharp, is something else and belongs on the continuum of patterns that are neither clearly ictal nor clearly benign. Mastering ictal EEG therefore means learning to read change across multiple dimensions simultaneously, a fundamentally different cognitive task from spotting a single spike. This module formalizes that skill as the law of evolution, grounds it in the network mechanisms that generate focal and generalized seizures, follows the seizure through propagation with both visual and quantitative tools, and ends at the post-ictal state and its prognostic weight.

The law of ictal evolution

An electrographic seizure characteristically displays evolution along four coupled axes. The first is frequency: most focal seizures begin with a discrete rhythmic discharge and then change frequency in an orderly way, classically slowing as the seizure matures, though many begin with low-amplitude fast activity that first accelerates before decelerating. The second is amplitude: the discharge typically grows from a low-voltage onset to a higher-voltage, more organized rhythm as more cortex is recruited into the synchronized population. The third is morphology: waveforms transform from sinusoidal to spiky, from monomorphic to polymorphic, or from rhythmic spike-and-wave to slower complexes as the run proceeds. The fourth is spatial field: the discharge recruits adjacent and then distant electrodes, the involved region enlarging and sometimes migrating across the scalp. A pattern that evolves convincingly across several of these axes meets the operational definition of an electrographic seizure; the axes are coupled because they all reflect the same underlying process of a self-organizing, spreading, and eventually exhausting population discharge.

The temporal cadence of evolution matters as much as its presence. A typical focal seizure unfolds over tens of seconds to a few minutes, with a clear onset, build-up, and decrescendo. The onset may take several stereotyped forms: low-voltage fast activity in the beta or gamma range, a focal rhythmic theta or alpha discharge, a run of repetitive spiking, an electrodecremental flattening in which the background voltage suddenly drops, or a sentinel spike followed by suppression. Recognizing these onset patterns is clinically valuable because low-voltage fast activity and the electrodecremental onset are particularly associated with a well-circumscribed seizure-onset zone and, in intracranial series, with favorable surgical outcomes, whereas a poorly localized rhythmic onset implies a more distributed generator and a harder surgical problem. The onset pattern is therefore not merely descriptive but prognostic, and it is one of the features the presurgical team weighs when deciding whether and where to resect.

It is essential to separate the electrographic seizure from the electroclinical seizure. Many electrographic seizures, especially in the critically ill, have no visible clinical correlate, while some clinical events, such as certain psychogenic nonepileptic seizures, have a dramatic semiology and a completely normal ictal EEG. The reader must hold both channels in mind: a convincing evolving discharge confirms an electrographic seizure regardless of behavior, and a normal ictal EEG during a convulsive-appearing event with preserved alpha rhythm and no post-ictal slowing is strong evidence against an epileptic mechanism. Video-EEG, which time-locks behavior to the tracing, exists precisely to adjudicate this correspondence and is the gold standard for the diagnosis of paroxysmal events of uncertain nature.

Focal onset versus generalized onset

The 2017 ILAE operational classification, which remains the working framework through the mid-2020s, classifies seizures first by onset: focal, generalized, or unknown, and then by motor versus non-motor features and, for focal seizures, by retained or impaired awareness. The electrographic distinction between focal and generalized onset is mechanistic. A focal-onset seizure begins in a circumscribed network within one hemisphere, so its ictal rhythm appears first over a restricted electrode field and then evolves and propagates outward. A generalized-onset seizure engages bilaterally distributed cortical and thalamocortical networks from the outset, so its discharge is bisynchronous and widespread within the first instants, without a localizable starting focus. This dichotomy is not merely descriptive; it dictates drug selection, prognosis, and surgical candidacy, which is why the EEG distinction carries such weight and why misclassification has real clinical cost.

The network mechanisms differ as fundamentally as the patterns. Focal seizures arise from local imbalances of synaptic excitation and inhibition, dysfunctional potassium and chloride homeostasis, and pathological recurrent excitation within a circumscribed cortical patch, where the same paroxysmal depolarizing shift that generates interictal spikes recruits its neighbors once surround inhibition fails decisively rather than transiently. Generalized spike-wave, by contrast, exploits the intact, normally functioning thalamocortical synchronizing machinery and runs it at the wrong gain; the cortex and the thalamus, gated by the reticular nucleus and its T-type calcium currents, oscillate together across both hemispheres in a circuit that is anatomically normal but functionally dysregulated. This is why genetic generalized epilepsy typically arises on a normal background and a normal MRI, whereas focal epilepsy so often has a structural or microstructural substrate.

Generalized-onset patterns include the generalized spike-and-wave and polyspike-and-wave of absence and myoclonic seizures, and the dramatic sequence of the generalized tonic-clonic seizure: an epileptic recruiting rhythm of generalized low-voltage fast activity during the tonic phase that increases in amplitude and decreases in frequency, becoming interrupted by slow waves that fragment the discharge into the clonic phase, followed by post-ictal suppression. Focal-onset patterns, by contrast, are defined by their restricted, evolving, propagating field. A critical caveat is focal to bilateral tonic-clonic evolution, the term that in the 2017 framework replaced the older secondarily generalized seizure, in which a focal seizure spreads to engage both hemispheres; the late generalized appearance can masquerade as primary generalized onset unless the brief focal onset is captured, a distinction with major surgical implications because the former may be surgically remediable while the latter is not.

The 2017 framework also refines the language of consciousness for focal seizures, replacing the older terms simple and complex partial with focal aware and focal impaired-awareness seizures, a change that matters electrographically because the spread of the discharge into bilateral networks is what typically extinguishes awareness. A focal aware seizure often shows a discharge confined to a restricted field, whereas the loss of awareness in a focal impaired-awareness seizure correlates with propagation into and disruption of the bilateral thalamocortical and default-mode networks that subserve consciousness. The reader should therefore expect the moment of behavioral arrest to coincide with electrographic spread beyond the onset zone, and should use that coupling as another check that the electrographic and clinical accounts of the seizure agree. The same logic explains why the semiology of a focal seizure unfolds in stages that mirror its anatomical march, and why a careful description of the first subjective symptom, the aura, frequently localizes the onset to the symptomatogenic cortex even before the EEG declares itself.

Reading a focal seizure: onset, evolution, propagation

The recording below is a focal seizure. Track it deliberately across the four axes of evolution. Identify the electrode field where rhythmic activity first emerges, note the onset morphology, then follow how the frequency and amplitude change second by second and how the involved field expands to neighboring and contralateral contacts. The seizure-onset zone is inferred from the earliest unequivocal ictal change, not from where the seizure looks most dramatic at its peak, because by mid-seizure widespread propagation has obscured the source. This is one of the most common errors of the inexperienced reader, who is drawn to the high-amplitude climax and reports its location rather than patiently rewinding to the first electrode to declare itself.

Localizing the onset is the same geometric problem as localizing an interictal spike, governed by the field, the gradient, and the phase reversal across montages. The double-banana longitudinal bipolar montage is the workhorse for this task, and the figure below lets you rehearse the chain logic by which a focal maximum produces a phase reversal that points back at the source. The same caveats apply that apply to interictal localization: an onset at a chain terminus produces no true reversal, a tangential generator can displace the apparent maximum, and the reader must cross-check montages rather than trust a single derivation.

Propagation dynamics follow anatomical highways. Mesial temporal seizures characteristically spread first to the ipsilateral lateral temporal neocortex, then via the hippocampal commissure and corpus callosum to the contralateral temporal lobe, and through the limbic and frontal connections to produce the autonomic and behavioral semiology of a focal impaired-awareness seizure. The measurable lag between the onset region and the regions recruited later is informative: a long, orderly propagation suggests a discrete origin amenable to localization, whereas near-instantaneous bilateral involvement suggests either a generalized network or very rapid transcallosal spread. The semiology evolves in lockstep with the electrographic propagation, which is why a careful clinician reads the seizure as a coupled electroclinical trajectory rather than a tracing alone, using lateralizing signs such as unilateral dystonic posturing or post-ictal aphasia to corroborate the electrographic onset.

Seizures also stop, and the mechanisms of seizure termination are as important as those of initiation. A focal seizure does not simply run out of fuel; its offset is an active, often abrupt transition in which the discharge becomes increasingly synchronized and rhythmic in its final seconds, the inter-burst intervals lengthen, and the whole network shifts into the profound inhibition of the post-ictal state. The accumulation of extracellular potassium, depletion of synaptic resources, activation of adenosine and other endogenous anticonvulsant systems, and a surge of inhibitory conductances together force the population out of the seizure state. This matters at the bedside because a seizure that fails to engage these terminating mechanisms is precisely what becomes status epilepticus, the subject of the next module, and because the abruptness and synchrony of the pre-termination phase are part of what the reader recognizes as a seizure that is ending rather than merely pausing. Understanding offset also disciplines interpretation of the ictal-interictal continuum, where a pattern that neither clearly evolves toward a seizure nor clearly terminates occupies the ambiguous middle ground.

Quantitative analysis increasingly supplements the visual read of propagation, particularly in presurgical evaluation. Time-frequency analysis of the onset characterizes the low-voltage fast activity that the eye struggles to resolve, directed-connectivity estimates attempt to infer which region drives which, and the concept of the epileptogenicity index quantifies the rapid appearance of high-frequency activity together with its timing to rank electrode contacts by their likely role at onset. These tools are most powerful on intracranial recordings, where the signal-to-noise ratio supports analysis into the high-frequency bands, and they should be understood as formalizations of the visual reader's intuition rather than oracles; they inherit the assumptions of their models and can be misled by the same volume conduction and limited sampling that challenge the human eye. A complementary modality, simultaneous EEG and functional MRI, can map the hemodynamic correlates of interictal discharges and, less often, of seizures, and contributes to noninvasive localization in selected presurgical cases, though its temporal resolution and the indirectness of the blood-oxygen signal keep it an adjunct rather than a replacement for the electrographic read.

The post-ictal state

Seizure termination is followed by the post-ictal state, the electrographic and clinical aftermath of intense synchronized discharge. The EEG shows post-ictal suppression or attenuation, often with regional or generalized slowing that gradually recovers over minutes to hours, reflecting the metabolic exhaustion and active inhibition that bring the seizure to an end. The post-ictal record is doubly valuable. First, focal post-ictal slowing or attenuation lateralizes and localizes the seizure to the hemisphere and region of the slowing, a robust confirmatory sign that complements the onset localization and is especially useful when the ictal onset itself was obscured by muscle artifact or rapid propagation. Second, the depth and duration of post-ictal suppression index the severity of the event; generalized post-ictal suppression after a convulsion reflects profound cortical inhibition.

Postictal generalized EEG suppression (PGES) after a generalized tonic-clonic seizure has drawn particular attention because it has been studied as a possible marker of risk for sudden unexpected death in epilepsy (SUDEP). The literature is genuinely controversial: some studies have associated prolonged PGES with later SUDEP risk while others have not replicated the association, and the field has not converged on PGES as a validated biomarker. The honest position for the trainee is that prolonged post-ictal suppression reflects the profound physiologic disruption a convulsion produces and is an active area of SUDEP research, without overstating it as an established predictor. The orderly return of normal background frequencies marks recovery, whereas failure to recover, or transition back into ongoing rhythmic discharges, signals evolution toward status epilepticus and is itself an indication for continued monitoring.