Epileptiform activity on EEG is indicative of cortical hyperexcitability, which carries an increased risk for seizures and the presence of an epileptic network within the brain. There are several types of epileptiform activity, including single discharges (sharps and spikes) and rhythmic and/or periodic activity. It is important to note that the presence of epileptiform activity is not sufficient for a diagnosis of epilepsy in a patient with no history of seizures. In a patient with a history of seizures, epileptiform activity in between seizures (interictal activity) is useful in helping to localize seizure onset.
Generally speaking, the location of an epileptiform (or interictal) discharge suggests cortical irritability of that region. For example, a discharge at T4 suggests right mid-temporal hyperexcitability / epileptogenic potential. However, there are a few fine details or caveats to this rule.
First of all, generalized discharges cannot be localized. Typically found with primary or symptomatic generalized epilepsies, generalized discharges begin in such a widespread fashion that they effectively involve the entire cortex simultaneously. It is possible that deeper structures, such as the thalamus, may be involved with generalized discharges.
Note that on EEG generalized discharges do not have to be completely the same in every single lead; there is often an anterior predominance to them, for example, but as long as the morphology of the discharge remains throughout all the leads (even if some are less well formed or lower amplitude) and the time of onset is the same in all the leads, you should consider it a generalized discharge.
You may also see discharges that appear generalized but are actually focal with rapid bisynchrony, in which they actually arise from a single location but the networks involved propagate the signal too quickly to trace that location, and they appear generalized. Sometimes with rapid bisynchrony, with careful review you can find a small precedent change before the discharge, such as very low amplitude fast activity, to suggest a lateralized onset.
The frontal lobes can be a confusing place for interictal discharges. Deep frontal discharges can be missed on scalp EEG entirely, and mesial frontal discharges can appear on EEG as if they are from the contralateral frontal lobe due to the direction of the discharges’ dipole. This isn’t always the case, but something to keep in mind when viewing frontal discharges.
Furthermore, while anterior temporal interictals are perhaps the most common focal epileptiform discharges, because of the proximity of the F7 and F8 electrodes to the inferior frontal region, what appear to be anterior temporal discharges can, at times, actually be from the inferior frontal region.
The central region is less complicated: discharges here are usually normal in the asleep state (ex. vertex waves from the central region) but basically never normal in the awake state. The one caveat is that very mesial discharges from the frontal or parietal lobes can sometimes be seen on the midline EEG electrodes, so if you see midline discharges in the asleep state that are not as symmetric as vertex waves, or with an atypical formation or field for vertex waves, you should be suspicious.
In the occipital region, remember that the O1 and O2 electrodes, similar to Fp1 and Fp2, suffer from the end of chain issue in bipolar montages. That is to say, because there is no electrode behind them to compare their voltage to, there is no phase reversal to stand out on EEG. So, if you see a positive discharge in the occipital regions that is not consistent with a lambda wave or POST, check a circumferential or referential montage to clarify if you’re seeing a true interictal discharge.
This tracing shows a clear spike and wave discharge at Fp2-F8. However, despite this being a bipolar montage there is no phase reversal. This is a good example of the end of chain issue, where there is nothing anterior to Fp2 to which to compare its voltage. It is possible that the maximal discharge, where the phase reversal would be, is actually anterior to or deep to Fp2, but because there isn't an electrode there to measure it, we can't see the full extend of what would be the phase reversal on scalp EEG. Also note the good field in the right frontocentral region, with some reflection to the left frontal/frontotemporal region.
Generally speaking, periodicity on EEG is almost always suspicious. The background should be a little chaotic looking, because all the neurons are going about their own business within their respective networks. When the background becomes too organized--including via periodicity--this raises suspicion for abnormal synchronous firing of neurons.
Periodicity, as defined by the ACNS criteria, requires the relevant discharges to occur for at least 6 cycles. So, if the discharges are periodic at 1 per second, they must go on for 6 seconds; if they are periodic at 2 discharges per second, they must go on for at least 3 seconds; if they are periodic at 3 discharges per second, they must go on for at least 2 seconds, and so on.
If the interval between each discharge in a periodic set of discharges varies by 25-50%, it is termed quasi-periodic. If the variance between each one is more than 50%, it isn't actually periodic at all. Full details of the ACNS approved documentation of periodic and rhythmic patterns on EEG are available here.
Below is an example of lateralized periodic discharges (LPDs), also termed periodic lateral epileptiform discharges (PLEDs). This example may be borderline between periodic and quasi periodic, as several of the discharges have longer intervals between them bordering on a 25% difference discharge to discharge.
While the above example is lateralized, you may also see generalized periodic discharges (GPDs). These are also concerning but lateralized activity is more likely to be associated with seizure activity. Note that periodic patterns can also be part of a seizure, but seizures require temporal and/or spatial evolution. However, if a periodic pattern arises and reaches 2.5 Hz for at least 10 seconds, that should be considered a seizure and merits treatment.
This tracing shows left central discharges, with a phase reversal at C3. Recall the requirements for periodicity: at least 6 cycles with a no more than 25% difference in time from one discharge to the next. This tracing shows at least 6 discharges at 1Hz (happening around once a second), so it meets the first criteria for periodicity. However, some discharges do have a more than 25%, but less than 50% difference between them, so it is a quasi-periodic pattern. Of note, the last discharge on the page, while in the same area, should not be considered part of this run of quasi-periodicity as there is too great a time between it and the prior discharge.
Recall that some rhythmic activity, such as frontal intermittent rhythmic delta activity (FIRDA) or generalized intermittent rhythmic delta activity (GIRDA), is not necessarily epileptogenic. When you see a consistently lateralized rhythmic pattern (such as theta or delta), however, this should raise suspicion for focal cortical hyperexcitability.
Rhythmic patterns carry similar requirements to periodic patterns, needing at least 6 cycles to be considered a rhythmic pattern. While periodic patterns, however, have discrete intervals between each discharge, rhythmic patterns do not--one waveform flows directly into the next.
Temporal intermittent rhythmic delta activity (TIRDA), in particular, is an epileptogenic pattern often associated with an underlying focal temporal lesion such as a tumor or cortical dysplasia.
Occipital intermittent rhythmic delta activity (OIRDA) is more common in children and usually also considered epileptogenic, although perhaps not as universally as TIRDA.
This tracing shows intermittent runs of approximately 3 Hz delta activity over the left temporal region, more prominent in the anterior than posterior temporal region. Because these intervals of 3Hz delta last at least 2 seconds, they complete the 6 cycle requirement for rhythmic delta activity. Note that the delta is very apparent in the lateral temporal leads below the traditional parasagittal and temporal chains--a reminder to always keep an eye on all of your available chains on EEG.
When you see diffuse fast activity, its often in the setting of excess beta activity (usually a benzodiazepine effect) and is benign, while very fast activity over the frontal regions is usually muscle artifact. However, paroxysmal runs of fast activity can be epileptiform, and are most commonly seen in patients with generalized epilepsy, Lennox-Gastaut Syndrome, and tonic seizures.
Paroxysmal fast activity can be localized, as in the example below, or generalized. When you see burst of generalized paroxysmal fast activity, ensure it is not actually a tonic seizure, of which fast activity is classically a key part.
This is a complicated tracing. First note the state of the patient, who appears to be in slow wave sleep (relatively synchronized high amplitude delta activity diffusely, no eye blinks or other evidence of being awake). There are multiple epileptiform discharges on the page, including a run of right frontal paroxysmal fast activity, right frontal spike and waves (Fp1 max), right mid-temporal sharps (T6 max, although poorly formed), and left temporal spike and waves. Note the field of the right temporal sharps into the right parasagittal chain. This kind of tracing--disorganized with multifocal discharges--is typical for Lennox Gastaut Syndrome.
If you look up in the sky you may see birds, but you can also see them on EEG tracings. Brief potentially ictal rhythmic discharges, or B(I)RDs, are ictal appearing rhythms that evolve but do not meet the minimum 10 second criteria to be considered a seizure. They can be seen in a broad range of patients, but most commonly in critically ill patients (especially those with newfound tumors and strokes), neonates, and sometimes those with NMDA receptor encephalitis.
While B(I)RDs are not technically seizures, they are associated with a similarly poor outcome and thus merit treatment; in fact, over 90% of patients with B(I)RDs on EEG will go on to have seizures. In the example below, we see an ~8 second run of evolving left temporal spike wave activity consistent with B(I)RDs, and note the field into the left parasagittal and subtemporal regions.
Is the discharge below most consistent with an epileptiform discharge, artifact, or nonepileptiform abnormality?
Here we see a three-phased generalized discharge with slight right hemispheric and bifrontal predominance, with a slight anterior to posterior lag in onset and in which each each phase is slightly longer than the preceding one. These are all hallmarks of triphasic waves, which are not epileptic in nature and are most often seen with toxic-metabolic encephalopathy.
The photoparoxysmal response, while perhaps one of the most well known facets of epilepsy, is in actuality quite rare. It describes the emergence of interictal activity as a result of photic stimulation. Typically, this interictal activity is generalized or focal occipital, arises during but outlasts the photic stimulation itself, and is always reproducible at the same frequency of light flashes. Note that photic driving does not confer an increased risk for a photoparoxysmal response or photosensitive seizures.
Here we have multiple, very well formed phase reversing spike and slow wave discharges that point toward T4 and C4, with slightly better formed morphology over the temporal than central region. These would thus be right centrotemporal spikes.
Note the field into the left central region, and also that in the right parasagittal chain, the frequent negative phase reversals can appear somewhat similar to electrode artifact, but you don't see evidence of such artifact elsewhere on the page.
While this discharge is best seen on the left, there is clearly a fairly symmetric spike and wave on the right side within the myogenic artifact.
This interictal discharge is maximal at T4 per the point of phase reversal, with a clear field throughout the left hemisphere.
On close look you can see that each slow wave on the left has a clear "notch" of a spike prior, while the right side has these less commonly, suggesting a left sided predominance.
PLEDS, or periodic lateralized epileptiform discharges (the new preferred term is lateralized periodic discharges, LDPs), are a descriptive term classically affiliated with HSV encephalitis. Here, note their periodicity is ~1 Hz. PLEDs merit urgent workup when seen.
Note that most of these very high amplitude slow waves has a pair of spikes immediately preceding it, consistent with polyspikes
These discharges are fairly bisynchronous, although the amplitudes are highest over P3. On the right side, C4, seems to be the focus. Note the field to the bitemporal regions.
These generalized discharges have a frontal predominance, which is fairly common in generalized epilepsy. Their symmetry and presence of clear spikes before the wave in each lead support these being generalized.
At times, interictal discharges will come in bursts or runs like this; as long as they do not continue to evolve, it is not a seizure. Note the small field to the right occipital region, seen best in the temporal chain.
Note that while generalized discharges can have bifrontal predominance, these discharges are truly just bifrontal--not generalized--as they are only well formed with clear spikes and waves n the frontal regions.
Temporal intermittent rhythmic delta activity (TIRDA) is a descriptive term--it is a slow rhythmic delta over a temporal region. While most IRDAs (ex. FIRDA) are not epileptogenic, TIRDA is like a spike or sharp: an interictal finding suggestive of cortical hyperexcitability from the affected temporal region.