Benign Rolandic Epilepsy: Characteristic EEG Findings

The hallmark EEG finding in benign rolandic epilepsy, now formally called self-limited epilepsy with centrotemporal spikes (SeLECTS), is a distinctive sharp wave over the centrotemporal region of the brain, typically followed by a slow wave. This pattern is so recognizable that experienced neurologists can often identify it within seconds of viewing a tracing. Yet the EEG in this condition carries more information than a simple “spikes present or absent” verdict, and the details of those spikes matter for prognosis, treatment decisions, and distinguishing a straightforward course from one that may evolve into something more complex.

What the Centrotemporal Spike Actually Looks Like

Despite being called a “spike,” the centrotemporal discharge is technically a sharp wave. A quantitative study of 43 children across five different clinical groups found the rolandic discharge had a mean duration of about 88 milliseconds, which is longer than the under-70-millisecond cutoff that defines a true spike in EEG terminology.¹ That distinction sounds pedantic, but it matters when you’re trying to separate rolandic discharges from other types of epileptiform activity. In practice, though, every neurologist still calls them “spikes,” and the name has stuck.

The discharge appears as a high-voltage, sharply contoured negative wave over the central and mid-temporal electrode positions (C3/C4 and T3/T4 in the standard 10-20 system), usually followed by a slower, rounded wave. The morphology is remarkably consistent. That same study of 43 children found the spike-and-wave complex looked identical across children with classic benign rolandic epilepsy, children with rolandic spikes but no seizures, and children with structural brain lesions, suggesting the waveform reflects a specific generator pattern rather than a specific disease.¹

The Dipole and Where It Points

One of the most useful features for identification is the horizontal dipole. A dipole simply means the electrical field has both a negative and a positive pole, like a tiny battery sitting inside the brain. In rolandic epilepsy, topographic EEG mapping shows a strong negativity over centrotemporal electrodes paired with a corresponding positivity over the frontal region.² The spike’s electrical field also spreads to parietal and upper frontal areas, with involvement of the midline.

This dipole pattern is thought to arise from a single generator oriented tangentially to the brain’s surface, sitting in the lower rolandic (sensorimotor) cortex right where the zero-potential zone exists between the frontal positivity and centrotemporal negativity.³ In practical terms, you can think of the spike as being generated in the part of the brain’s motor strip that controls the face, tongue, and throat, which lines up neatly with the typical seizure symptoms of drooling, facial twitching, and difficulty speaking. Functional MRI studies have confirmed this: when researchers triggered the brain’s imaging signal using the timing of the spikes, activation appeared in the somatosensory cortex near the face area.⁴

In a Brazilian clinical study, the horizontal dipole was present in about 75% of EEGs from children with rolandic epilepsy.⁵ Its presence is considered one of several electrographic markers of a benign course. When reading a child’s EEG, finding this dipole formation provides added confidence you’re looking at rolandic epilepsy rather than something else.

Background Activity and Markers of a Benign Course

Beyond the spikes themselves, the rest of the EEG matters. In typical rolandic epilepsy, the background brain rhythms should look normal. The same Brazilian study found that normal background activity was present in about 70% of EEGs overall, but among children who were later confirmed to have clinically benign courses, normal background predominated. Children whose epilepsy eventually took a less benign path were more likely to show abnormal background activity.⁵

Other EEG features associated with a benign course include stereotyped (repetitive, uniform) spike morphology, which was seen in about 60% of recordings, and bilateral discharges in roughly 42% of EEGs.⁵ Spikes can shift sides between recordings or appear independently on both hemispheres; this bilateral bouncing does not mean a worse prognosis and is actually part of the expected pattern.

Why Sleep Lights Up the EEG

One of the most striking features of rolandic epilepsy is how dramatically spikes increase during sleep, particularly during non-rapid-eye-movement (NREM) sleep. This is important practically because a routine daytime EEG in an awake, cooperative child may show only a handful of discharges, or sometimes none at all. A sleep recording captures the full picture.

A combined EEG-fMRI study demonstrated this vividly in a single patient: while awake, about 6.9 spikes per minute were recorded, but once the patient drifted into stage-two NREM sleep, the rate climbed to 12.5 spikes per minute, nearly doubling.⁶ The study also revealed that the thalamus showed significant activation linked to the spikes during sleep, suggesting that the brain’s natural sleep-promoting circuits actively facilitate these discharges. This thalamic involvement helps explain why seizures in rolandic epilepsy so often occur as children fall asleep or wake up.

A follow-up study tracking children over three years confirmed that the average number of spikes during sleep was significantly higher in younger children (ages three to eight) and then dropped substantially in the nine-to-eleven age range, declining further after age twelve.⁷ Sleep EEGs performed at different ages can therefore serve as a rough gauge of where a child stands on the trajectory toward resolution.

When the Spikes Disappear

The natural history of rolandic epilepsy is one of resolution. Seizures and EEG abnormalities almost always disappear by puberty. This age-dependency is so consistent that researchers have long attributed the condition to a hereditary variation in brain maturation rather than to structural damage.⁸ In other words, the developing brain goes through a window of cortical excitability in the rolandic region, and it grows out of it.

The serial spike-count data bear this out: paroxysmal discharges during sleep decrease significantly between early childhood and the preteen years.⁷ For parents tracking their child’s EEGs over time, this declining spike count is generally reassuring. Clinicians sometimes use serial sleep EEGs to decide when it’s appropriate to taper medication.

Ripples on Spikes and What They Predict

Standard clinical EEG filters out very fast brain activity, but research-grade recordings can detect “ripples,” which are brief bursts of oscillation in the 80-to-250 Hz range superimposed on top of the rolandic spikes. These high-frequency oscillations have attracted attention because they seem to carry prognostic information that the spikes alone do not.

A study comparing children with rolandic spikes who had epilepsy against those who had spikes but no seizures found a clear split. Children without epilepsy showed essentially no ripples. More strikingly, the number of ripples correlated strongly with the number of seizures, whereas the count of conventional spikes had only a borderline relationship. Having more than two ripples predicted whether a child would actually have seizures with high accuracy, while the spike count alone could not distinguish between children who would and would not seize. When the threshold was raised to more than five ripples, this measure could differentiate a benign course from an atypical or symptomatic epilepsy with perfect specificity.⁹

This finding is clinically meaningful because many children have rolandic spikes on EEG and never develop seizures. Ripple analysis could eventually help clinicians decide which children actually need treatment and which can be watched. It’s not yet part of routine clinical practice, but the research is promising enough that high-frequency oscillation detection is being explored as a new biomarker for epilepsy broadly.¹⁰

Atypical EEG Evolution and Continuous Spike-Waves During Sleep

The word “benign” in the condition’s old name sometimes misleads families into thinking nothing can go wrong. In a small minority of cases, rolandic epilepsy takes an atypical course. The EEG is often the first signal that something has shifted. Warning signs include spreading of the spike focus from centrotemporal to fronto-centrotemporal areas, increasing spike frequency during both wakefulness and sleep, and eventually the appearance of nearly continuous spike-and-wave discharges during slow-wave sleep, a pattern known as CSWS.¹¹

Atypical evolutions tend to be accompanied by new or worsening neuropsychological problems, including language regression and cognitive decline.¹² In a retrospective case series, children whose rolandic epilepsy evolved toward CSWS had early seizure onset (average age five) and showed clinical and EEG deterioration roughly a year and a half later. The features flagging risk were early onset of seizures, new seizure types with increasing frequency, and the EEG changes mentioned above.¹¹ This underscores why follow-up EEGs matter even in a “benign” epilepsy. A child whose spike distribution is widening or whose sleep EEG is becoming more dense with discharges deserves closer monitoring.

What the Spikes Do to Thinking and Language

Even when seizures are infrequent and the overall course is benign, the spikes themselves appear to transiently disrupt brain function. Children with rolandic epilepsy have been found to show deficits in articulation, auditory and visual memory, language comprehension, and visual-constructive performance compared to healthy controls.¹³ These are not catastrophic impairments, but they can affect school performance, particularly in reading and spoken language.

A combined EEG-fMRI study demonstrated that individual centrotemporal spikes directly disrupt functional brain networks involved in language, behavior, and cognition in real time. The researchers concluded that suppressing the discharges could reduce the risk of neuropsychological difficulties.¹⁴ This creates a clinical dilemma: many children with rolandic epilepsy have very few seizures and might not need antiseizure medication on seizure grounds alone, yet their abundant interictal spikes may still be affecting cognition. Whether to treat in that situation is an evolving discussion that the EEG findings feed directly into.

Medication Effects That Show Up on EEG

Carbamazepine, one of the older antiseizure medications, deserves special mention because it can paradoxically worsen rolandic epilepsy in some children. In a study of paradoxical reactions to carbamazepine, children with benign rolandic epilepsy who had diffuse interictal sharp-and-slow-wave discharges were among those more likely to experience seizure exacerbation on the drug. In all cases, the clinical and electrical changes reversed when carbamazepine was withdrawn.¹⁵

A case report illustrated this scenario in detail: a child diagnosed with rolandic epilepsy was started on carbamazepine and initially improved, but when the dose was increased, seizure frequency rose and a repeat EEG showed electrical status epilepticus in slow-wave sleep, the very CSWS pattern associated with atypical evolution. Once carbamazepine was stopped, the EEG improved.¹⁶ The takeaway for parents and clinicians is that worsening EEG findings in a child on carbamazepine should prompt consideration of the drug itself as the culprit.

Rolandic Spikes in Children Without Epilepsy

A fact that often surprises parents: centrotemporal spikes can appear on the EEGs of children who have never had a seizure. These incidental rolandic discharges are found in roughly 2-3% of healthy school-age children. The spikes look identical to those in children with rolandic epilepsy. As the morphology study noted, the waveform is the same regardless of whether the child has seizures, has a structural brain abnormality, or has no clinical symptoms at all.¹ This means the EEG pattern alone does not equal a diagnosis of epilepsy. Clinical context, particularly the history of seizures, is essential.

This overlap also has implications for differential diagnosis. A study of 1,340 children with focal seizures identified 24 who had both emetic (vomiting) episodes during seizures and centrotemporal spikes on EEG. Of those, 83% had seizure patterns more consistent with Panayiotopoulos syndrome, a different childhood epilepsy that features autonomic symptoms and typically involves occipital or multi-focal spikes. Some of these children later developed pure rolandic seizures, and a minority had typical rolandic seizures with concurrent vomiting.¹⁷ The presence of centrotemporal spikes, in other words, does not automatically mean rolandic epilepsy. The seizure semiology has to match.

Structural brain lesions can also produce similar-looking spikes. A case report described an eight-year-old with facial motor seizures whose EEG showed left-hemisphere centrotemporal spikes fully consistent with benign rolandic epilepsy, but MRI revealed focal cortical dysplasia. Dipole analysis of the spikes confirmed they originated from the rolandic fissure rather than the dysplasia, but the structural abnormality would not have been discovered without imaging.¹⁸ Cases like this are uncommon but explain why many clinicians order an MRI at least once, even when the EEG looks textbook.

Genetics Behind the Spikes

The tendency to produce rolandic spikes runs in families, and recent genetic work has identified a concrete contributor. Mutations in the gene GRIN2A, which encodes a subunit of the NMDA receptor involved in excitatory brain signaling, were found in about 7.5% of individuals across two large cohorts with idiopathic focal epilepsy and rolandic spikes. The mutation rate was higher in more severe phenotypes: roughly 5% in children with classic benign rolandic epilepsy, rising to about 18% in those with CSWS.¹⁹ GRIN2A alterations are now recognized as a major genetic risk factor for this spectrum of conditions. The gradient in mutation frequency from mild to severe fits the clinical observation that rolandic epilepsy, atypical rolandic epilepsy, and CSWS may represent a continuum rather than wholly separate disorders.

How Neuroimaging Complements the EEG

Standard clinical EEG remains the primary diagnostic tool, but combined EEG-fMRI studies have added a layer of understanding. When researchers time-locked brain imaging to the centrotemporal spikes, they found that the shape of the blood-flow response in rolandic epilepsy did not follow the standard template used for most brain activation studies. In a group of eight well-characterized children with the condition, the average blood-flow response to spikes differed from the expected pattern, and using a tailored model improved both the sensitivity and specificity of localization.²⁰

This is mostly a research finding at the moment, not something that changes bedside care. But it reinforces the idea that rolandic epilepsy involves a distinct and somewhat unusual pattern of cortical excitability, not just a generic epileptic discharge. It also explains why standard fMRI analysis protocols can underestimate the extent of brain involvement in this condition.

Automated Spike Detection and Where It Stands

Reading EEGs for rolandic spikes still depends heavily on trained human eyes, but automated detection is advancing. A systematic review of deep-learning approaches to epileptiform discharge detection on scalp EEG found a median area-under-the-curve of 0.94 across studies, indicating strong overall accuracy.²¹ Most of these models used convolutional neural networks, and only a minority were trained on data from multiple clinical centers, which limits how well they generalize to new settings. For rolandic spikes specifically, the stereotyped morphology and predictable location make them a relatively friendly target for automation. Still, the field is not yet at a point where automated detection replaces expert review; it’s more of a screening tool that flags suspicious segments for a neurologist to confirm.

Cortical Excitability Beyond the EEG Tracing

Transcranial magnetic stimulation (TMS) has been used to probe the excitability of the motor cortex directly in children with rolandic epilepsy. In one study, motor-evoked potentials were triggered in hand muscles using TMS and then conditioned by electrical stimulation that produced evoked spikes, at varying time intervals.²² This work helps clarify the role of the motor cortex in generating or amplifying the discharges. A pilot TMS-EEG-EMG study further tested the hypothesis that net cortical excitability decreases with age in these children, aligning with the clinical observation that spikes fade around puberty, and that patterns of cortical plasticity predict learning outcomes.²³ If these findings hold up in larger samples, they could eventually provide a physiological measure, beyond spike counting, that tells clinicians how mature and stable the motor cortex is becoming in an individual child.