To begin the exercise, it would be best to start with the more obvious and then progress to the less typical and subtler forms of electrographic seizure activity. We will begin by reviewing the graphic features that constitute seizures and understand how they might be put together in a way that reflects an ictal pattern.

The term epileptiform pattern is used for “distinctive waves or complexes, distinguished from background activity, and resembling those recorded in a proportion of human subjects suffering from epileptic disorders.”^1,2 In effect, to find a forest,
we would first have to be able to identify a tree. These waveforms with epileptiform morphology are the spikes and sharp waves and may be recognized as epileptiform with a combination of form and field. Spikes are defined as lasting 20-70 ms, with a voltage of at least 70 µV. Sharp waves last between 70 and 200 ms. The EEGer must be able to distinguish these phenomena from sharply contoured but normal physiologic waves or paroxysmal activity with other pathologic significance. Efforts have been made to define the surround (background) of the epileptiform discharge, but it is most useful to identify the spike or sharp component itself (form). Mathematical modeling using differential equations has enabled computerized recognition of these phenomena with some success, though false positives and negatives are frequent.

Once the sharpness of the discharge has been spotted, the EEGer comes to realize that just spotting the spike implies recognition that the event clearly stood out from its background. In a general sense, it is not necessary to examine each page in second-to-second detail for a possible spike; the process is more like looking for a black pebble on a white-pebbled driveway. Each white pebble does not have to be examined to spot the black one among the whites.

The next step reinforcing the pathologic (rather than artifactual) nature of the discharge is to look for the EEG evidence of a spatial field involving one or more components of the discharge (spike, sharp, after-going slow wave). Recall that the electrode pairs are fed into a differential amplifier that amplifies the difference between the two electrodes and generates a single line of EEG tracing (known as a derivation). The spike/sharp component should be examined for evidence of a region of maximal electronegativity, marked by the derivation with greatest amplitude of sharp activity on a referential montage, or a “reversal” of phase within a chain on a bipolar montage. This “field” of negative potential may be visualized on EEG by its appearance at adjacent electrode positions, with a smooth “falloff” (decreasing amplitude) at electrodes increasingly distant from the area of maximal electronegativity.

In adults, the vast majority of spikes possess such a locus of maximum electronegativity. As a result, the spike in surrounding (but decreasing) areas of negativity will appear on the adjacent derivations as a phase reversal—in effect, the spikes point to each other indicating the locus of maximum electronegativity (Fig. 11.1). As the electrodes in bipolar montages are further removed from the focus, they will continue to point toward the focus whether longitudinally or transversally arrayed. It should be remembered that with bipolar montages, such phase reversals only appear if the focus lies within the reach of the bipolar chain. If it lies at the end of the chain, then the spike will be oriented in the same direction in all derivations of that chain. For spikes, the extent of spatial spread may be relatively limited, because spikes are near-field phenomena, which do not necessarily radiate widely.

The next step is to round up the usual suspects. The earliest ictal components in a complex waveform are commonly high-frequency spikes, polyspikes, or sharp waves, which will evolve with time to a lower frequency and often broader discharge morphology—the slow wave. Such slow waves (which last 150-300 ms) can appear as an after-going component of the spike or sharp wave or in isolation as rhythmic slower waves. These should be understood in two ways. First, ictal slow waves project with the same spatial field properties as the spikes at seizure onset and will often also phase-reverse around the focus of maximum spike negativity. However, the slow component represents a broad inhibitory postsynaptic potential (IPSP) in the cortex surrounding the spike and thus projects a bigger field, which can often be seen in derivations lying further away. The cause of such slow waves is believed to be the far-field projection of this lower-frequency synaptic activity. Second, even when slow waves
appear without a spike component, they may be recognized by their phase reversal and/or field projection (see the slow wave in Fig. 11.1). In summary, we can identify the epileptic discharge by its initial sharp morphology (form), distinguish it from artifact by its field characteristics, and incriminate it as having epileptic significance by its shady association with similarly behaving slow-wave components (Table 11.1).

If this was not enough to clinch the conviction, there is also the offender’s rap sheet and its tendency to return to the scene of the crime. Discharges rarely occur in total isolation. When they appear, they often do so repeatedly and in the same area (Fig. 11.1). Thus, the EEGer must look elsewhere in the EEG record for this same ictal constellation leaving its fingerprints again and again, like the telltale glove of The Pink Panther. However, always be wary and look for the possibility of false circumstantial evidence, that is, repeated events produced in the electrodes due to artifacts. Spiky discharges induced by cardiac activity, IV drips, or other extracerebral causes can sometimes appear ictal but are the equivalent of innocent bystanders!

So far, this diagnostic process has identified the footprints of the culprit, but the art of seizure detection is to capture the offender in flagrante delicto—the seizure itself, as it occurs. This brings us to the concept of progression of the ictal focus over time, space, and morphology, that is, the seizure evolution.

An easy way to conceptualize interictal foci and seizures is to think of earthquake or volcanic rumblings vs the eruption itself. From a spatial perspective, observers (electrodes) in the areas around the miniquakes can each point to (triangulate) the area of maximal activity even before the full event occurs. The pattern of buildup, intensity, repetition, and progression of waveforms, frequencies, amplitudes, and spatial distributions defines the active process—the eruption of the seizure (Figs. 11.2 and 11.3). The buildup and temporal evolution of these events in the seizure eruption (and not necessarily the number of such miniquakes) will distinguish a seizure from other epileptic discharges that do not represent seizures.

To explain some of this ambiguity, we can best return for a moment to a previous analogy—the forest. Imagine if you will, traveling along a straight country road through a farmer’s fields. The isolated trees encountered along the way represent the isolated epileptic discharges. Great distances can be crossed in which no trees are seen, or perhaps the occasional tree is spotted, each of which can be readily identified. The forest represents the seizure, and it too may be easily identified as a chaotic and often sudden appearance of a profusion of trees. However, if we were to define a forest (and hence a seizure) as a certain density or frequency of trees per se, our definition would force us to misdiagnose a row of trees planted by the side of the road at frequent but regular intervals as a forest—and clearly it is not. Such groupings of quasiregular ictal discharges are referred to as periodic discharges (PDs). PDs are pseudometronomic (nearly rhythmic) events that are often lateralized over one hemisphere (lateralized periodic discharges or LPDs)^3,4,5,6 (Fig. 11.4); as bilateral independent periodic discharges (BIPDs)^6,7; or as generalized (synchronous bilateral) periodic discharges (GPDs).^6,8,9 To return to our metaphor: (1) LPDs are trees planted on one side of the road; (2) BIPDs are trees planted on both sides at different but regular intervals; and (3) GPDs are trees planted in the same place on
both sides of the road and at the same intervals. It can be hard to differentiate rapid LPDs (discharges at one per second or faster) from ongoing focal status epilepticus, or whether fast generalized periodic sharp discharges represent generalized NCSE (GNSE, a seizure pattern) or GPDs (an interictal pattern). One way to distinguish between these possibilities is that there is little frequency variation from beat-tobeat (discharge to discharge) with LPDs and GPDs (hence the name periodic), but seizures often have waxing and waning intervals and migration of activity from one area to another. But the challenge comes in differentiating between PDs and status epilepticus,¹⁰ since both patterns are relatively invariant and fast. In such cases, circumstantial clinical information must play a more important role. For example, rapid or slow GPDs with a suppressed background are typically observed after cardiorespiratory arrest. Certainly, seizures have been described in this setting, but in the absence of myoclonus or other behavioral correlate, this pattern is usually interpreted as interictal.

One strategy to distinguish between PDs and seizures is a therapeutic challenge with intravenous (IV) benzodiazepines. If benzodiazepine administration produces prompt resolution of EEG and clinical seizure activity, or increased mental alertness, this confirms a diagnosis of seizures, since patients with encephalopathy will not improve clinically (although the EEG may improve). When discharges on EEG abate and the patient does not wake up, it is not possible to determine whether one was treating LPDs or seizures. When there is rapid EEG regression of epileptiform activity with IV lorazepam, but a more gradual clinical improvement over days because of an intercurrent encephalopathy, concluding that a patient has been in electrographic status epilepticus is less certain. The diagnosis becomes a tug of war between two electrophilosophic camps and will depend more on supportive and compelling clinical details than specific EEG characteristics.^{11,12,13,14,15}