What Do XL Events and XL Spikes on an EEG Mean?

XL events and XL spikes on an EEG report refer to “extra-large” epileptiform discharges flagged by automated spike-detection software, most commonly Persyst, one of the leading platforms used by hospitals and epilepsy centers to analyze brainwave recordings. The “XL” label is a size category the software assigns based on how tall, sharp, and electrically prominent a spike is compared to the surrounding brain activity. Seeing that label on a report can be alarming, but it is a software classification rather than a standalone diagnosis, and what it actually means for you depends on the broader clinical picture.

How Automated EEG Software Sorts Spikes by Size

When an EEG is recorded, the raw tracing can stretch across hours or even days of data. No neurologist has time to review every second of that recording by hand, so hospitals rely on software to scan the data and flag patterns that look like epileptiform activity. Persyst and similar programs use algorithms that measure the shape, height, and sharpness of each waveform. When the software identifies something that looks like an epileptic spike, it assigns it a size label. In Persyst’s system, those labels run from small (S) through medium (M), large (L), and extra-large (XL). The XL tag simply means the spike was among the tallest and most morphologically prominent events the algorithm detected in your recording.

These labels are not standardized across all EEG software. A spike labeled “XL” in Persyst might be classified differently by another platform like Encevis or by a custom hospital algorithm. The underlying measurements are similar, though: amplitude (how tall the spike is), duration (how quickly it rises and falls), and sharpness (how pointed the peak is compared to the slower waves around it). Research on spike shape has found that most epileptic spikes show a characteristic asymmetry, with the second half-wave being larger and longer than the first in roughly three-quarters of recorded channels.

What It Means When a Spike Gets the XL Label

A bigger spike does not automatically mean a worse problem. The XL designation tells you that the electrical discharge was particularly strong at the scalp electrode, but a spike’s size on the recording depends on many things beyond how “severe” the underlying brain activity is. The thickness of your skull, the distance between the source and the electrode, whether you were awake or asleep, and even your hydration level can all affect how tall a spike appears on the tracing. Sleep in particular tends to make spikes grow larger and appear more frequently, which is one reason overnight or sleep-deprived EEGs are ordered when a standard recording comes back normal.

That said, neurologists do pay attention to especially prominent spikes. Algorithms designed to detect interictal spikes, the bursts of abnormal electrical activity that occur between seizures, are specifically tuned to find the most conspicuous ones because those tend to be the most reliably epileptiform. One automated detection system showed roughly 89% sensitivity for identifying the electrode contacts that contained the most prominent spikes, compared with about 63% sensitivity for picking up every individual spike in the recording.¹ In other words, the software is better at catching the big, obvious spikes than the small, ambiguous ones, and those big spikes tend to be the ones that matter most for localizing the problem area.

Why Software Labels Exist in the First Place

You might wonder why hospitals let a computer classify brain activity instead of leaving it entirely to human experts. The answer is that human experts disagree with each other far more than most people realize. In one study, three highly trained reviewers were asked to mark individual spikes across an average of 96 electrode channels in ten patients. They collectively flagged nearly 79,000 spike events, but only about 24% of those detections were marked by at least two of the three reviewers. A mere 3% were marked by all three.² That level of disagreement is striking, and it is one of the main reasons automated tools have become so widely adopted.

Persyst’s algorithm, for instance, has been tested against human experts and found to perform comparably. When used to calculate the spike-wave index in a specific type of childhood epilepsy, Persyst 13 showed agreement with human experts that was at least as tight as the agreement between pairs of human experts themselves.³ A separate comparison of two commercial packages found that Persyst achieved about 95% sensitivity when analyzing 30-minute selections, with specificity around 88%, and the agreement between the software and the experts was on par with the agreement between experts themselves.⁴ So when you see an XL spike on your report, it represents a detection the software is fairly confident about, and that confidence generally holds up well against trained human eyes.

Interictal Spikes and What They Tell Your Neurologist

Whether classified as small, large, or extra-large, the spikes on an EEG are collectively called interictal epileptiform discharges, meaning they are the abnormal electrical bursts that happen between seizures rather than during one. Their presence is one of the strongest pieces of evidence a neurologist uses to support an epilepsy diagnosis, but their absence does not rule epilepsy out. Plenty of people with epilepsy have normal EEGs on any given day, which is why the test sometimes needs to be repeated or extended.

When spikes are present, neurologists look at several things beyond size. Location matters: spikes consistently appearing over one part of the brain suggest a focal source, while spikes that appear on both sides simultaneously point toward a generalized process. Research comparing these two patterns has found that high-frequency ripples riding on top of the visible spike can help distinguish focal epilepsy from truly generalized epilepsy, even when the spikes themselves look bilateral and symmetric on the standard EEG.⁵ That kind of detail is invisible to the naked eye and requires specialized analysis, but it illustrates why the raw size category alone does not determine the clinical interpretation.

Frequency also matters. A few scattered spikes across a 30-minute recording carry a different weight than continuous or near-continuous spiking. Certain structural brain abnormalities, such as focal cortical dysplasia, can produce an intrinsic pattern of nearly nonstop spiking that stands out clearly on intracranial recordings.⁶ In those cases, the volume of spike activity, more than any single spike’s size, is what alerts the clinical team that a structural lesion may be involved.

When Spike Patterns Point to a Structural Cause

Certain EEG patterns are associated with specific types of brain malformations. A systematic review of interictal and seizure-onset patterns in malformations of cortical development found that while some patterns appeared across multiple types of malformations, others were seen exclusively in focal cortical dysplasia. Repetitive bursting spikes on intracranial EEG and certain rhythmic epileptiform discharge patterns on scalp EEG showed up only in that particular structural abnormality and not in other types of cortical malformations.⁷

This does not mean that seeing XL spikes on your report signals a structural problem. Most people with epilepsy and interictal spikes do not have a visible brain lesion. But when spikes are unusually frequent, unusually focal, or have specific morphological features, the neurologist may recommend an MRI or other imaging to look for an underlying structural cause, especially if surgery is being considered as a treatment option.

How Frequent Spikes Can Affect Thinking and Memory

One aspect of interictal spikes that often surprises patients is that they can affect cognition even though they are not seizures. Each spike represents a brief disruption in normal neural processing, and when those disruptions happen often enough, they can interfere with learning, attention, and memory consolidation. This effect has been documented across multiple studies, with evidence suggesting that frequent and widespread interictal spikes impair cognitive abilities by disrupting both waking learning and the memory consolidation that normally happens during sleep.⁸

This finding has practical implications. If you or your child has epilepsy and is struggling with memory, attention, or school performance, part of the problem may not be the seizures themselves but the constant background static of interictal spikes. The spike-wave index, which represents how much of a sleep recording is occupied by spike-and-wave discharges, is one way clinicians quantify this burden. A high spike-wave index during sleep is associated with more significant cognitive effects, which is why reducing spike frequency is sometimes a treatment goal even when seizures are already controlled.

Do Seizure Medications Reduce Spike Activity?

The relationship between anti-seizure medications and interictal spikes is more complicated than you might expect. Broad-spectrum agents like valproate and lamotrigine have shown consistent ability to suppress interictal discharges, and that suppression appears to translate into cognitive benefit, especially in children. Sodium channel blockers such as lamotrigine and oxcarbazepine have been linked to cognitive improvements as well. Newer therapies including cannabidiol and perampanel also show promising results for reducing spike activity and improving behavioral outcomes.⁹

However, a recent study complicated the picture by showing that when you carefully control for sleep and recent seizure activity, the link between medication levels and spike rates largely disappears. Spike rates increased after seizures and during sleep, and once those two factors were accounted for, there was no significant relationship between overall medication load and how often spikes occurred. The one exception was levetiracetam, which did show a measurable effect on spike rates during medication taper.¹⁰ This suggests that spike frequency may be driven more by brain state, such as whether you are asleep or have recently had a seizure, than by how much medication is in your system at any given moment. For clinicians, it means spike counts on an EEG should not be used as a simple yardstick for whether the medication dose is right.

Reading Your Own EEG Report

If you have received a report that lists XL events or XL spikes, here are a few things worth keeping in mind. First, the XL label comes from the software, not your neurologist. The neurologist reviews the flagged events and decides which ones are genuinely epileptiform and which are artifacts, muscle twitches, or other non-epileptic sharp transients that the algorithm misidentified. Software sensitivity is good but not perfect, and false positives do occur. The final interpretation in the written report matters far more than the raw software output.

Second, a handful of XL spikes does not necessarily mean your epilepsy is “worse” than someone else’s whose report shows only medium-sized spikes. Spike amplitude varies enormously between people for reasons that have nothing to do with disease severity. What your neurologist cares about is the overall pattern: where the spikes are, how often they appear, whether they change during sleep, and how they fit with your seizure history and imaging results.

Third, if your report mentions a large number of events across the recording, that spike burden is worth discussing with your doctor regardless of the size labels attached. As covered above, frequent spiking can affect cognition and may warrant treatment adjustments even when your seizures are well controlled.

The Expanding Role of AI in Spike Detection

The algorithms behind labels like “XL” are evolving rapidly. Traditional spike-detection software relies on hand-engineered rules about what a spike should look like: certain amplitude thresholds, certain rise-and-fall times, certain ratios of sharpness to background activity. Newer systems are moving toward deep learning, where neural networks learn to recognize epileptiform patterns from large datasets of expert-labeled recordings rather than following pre-programmed rules. One such system, SpikeNet2, uses a residual network architecture enhanced with a technique called hard-negative mining, which specifically trains the model to avoid flagging waveforms that look like spikes but are not.¹¹

These AI-based approaches are designed to reduce the false-positive rate, which is one of the main frustrations with current software. A neurologist who opens a report and sees hundreds of flagged “spikes” that are mostly eye blinks or muscle artifact wastes valuable time sorting through them. Better algorithms mean cleaner reports, which in turn means the size classifications attached to genuine spikes become more trustworthy. As these tools mature, the labels on your EEG report, including that XL tag, will increasingly reflect what trained human experts would have said if they had the time to review every second of the recording themselves.

Why the Same Brain Can Produce Different Spike Counts on Different Days

One thing that catches many patients off guard is that two EEGs done a week apart can look quite different in terms of spike frequency and prominence. This is not because your epilepsy changed in a week. Spike activity is heavily influenced by your sleep state, how recently you had a seizure, stress levels, and even caffeine intake. The study mentioned earlier on medication taper found that seizure occurrence and sleep were the dominant drivers of spike rate changes, easily overshadowing the effect of medication adjustments.¹⁰ So if your first EEG was recorded during a 20-minute awake session and your second was an overnight study that captured deep sleep, the second one may show far more (and larger) spikes without anything having changed about your condition.

This variability is also why a single normal EEG does not rule out epilepsy, and why a single EEG full of XL spikes does not by itself confirm a dire prognosis. The recording is a snapshot. Your neurologist interprets it in context: your clinical history, your seizure descriptions, your imaging, and sometimes multiple EEGs over time. The XL label on any one spike is a useful piece of that puzzle, but just one piece.