Proposed classification: Syndromes in Children

Members of the Nosology and Definitions Task Force of the ILAE have developed four position papers on syndrome definitions at various ages. The ILAE guideline process requires obtaining feedback and comments from its members on the proposed paper. These comments from our international community will be reviewed by the working group before finalizing the paper.

The revised version of the manuscript is being reviewed by the ILAE and will then be submitted for publication. Review comments below.

Draft: ILAE Classification and Definition of Epilepsy Syndromes with Onset in Childhood

See all four Proposed papers for Nosology and Definitions


15 July 2021

1. Classification
The childhood-onset syndromes can be broadly divided into three main groups
(1) Focal epilepsy syndromes of unknown cause, most of which are self-limited
(2) Generalized epilepsy syndromes, which are thought to have a genetic basis, and
(3) Developmental and epileptic encephalopathies (DEE) which often have both focal and generalized seizures, including Lennox-Gastaut syndrome (LGS) and Developmental and/or Epileptic Encephalopathy with spike-wave activation in sleep (D/EE-SWAS), or may have generalized seizures alone, such as Myoclonic Atonic Epilepsy (MAE), or just focal/multifocal seizures alone, such as Hemiconvulsion Hemiplegia Epilepsy (HHE) and Febrile Infection-Related Epilepsy Syndrome (FIRES).

2. Terminologies changed
Proposed syndromes are subdivided into
(1) Self-Limited Focal Epilepsies of Childhood (SeLFE)
a) Self-Limited Epilepsy with Centrotemporal Spikes (SeLECTS) (formerly called BECTS)
b) Self-Limited Epilepsy with Autonomic Seizures (SeLEAS) (formerly called Panayiotopoulos syndrome or early-onset benign occipital epilepsy)
c) Childhood Occipital Visual Epilepsy (COVE) (formerly called late-onset benign occipital epilepsy or Gastaut syndrome or Idiopathic childhood occipital epilepsy – Gastaut type)
d) Photosensitive Occipital Lobe Epilepsy (POLE) (formerly called idiopathic photosensitive occipital lobe epilepsy)
(2) Genetic Generalized Epilepsies and
(3) Developmental and/or Epileptic Encephalopathies of Childhood.

1. New terminologies are better descriptor of seizure type and decreases ambiguity related to previous nomenclatures especially the previously termed early and late onset occipital epilepsy.
2. Cases of SeLECTS especially with GRIN2A mutation evolves to ESES so not all cases are self limited. So should “self-limited” be used as prefix for all childhood epilepsy with centro-temporal spikes.
3. FIRES has a stormy acute phase but chronic phase may be variable and mostly a sequelae of the injury to brain during acute phase. So should all of these be categorised under the third category of developmental and/or epileptic encephalopathies of childhood. Same for HHE.

Sheffali Gulati and Sonali Singh

15 July 2021

The ILAE Nosology and Definition Taskforce should be commended on this herculean effort. Clearly your group has spent significant time and thought into creating these series of manuscripts and topics. Due to my area of interest, I wanted to make a few small suggestion regarding the D/EE-SWAS proposal.

1. The high incidence of neuroimaging abnormalities in patients with SWAS should warrant brain MRI in most cases, especially in D-SWAS. I would encourage adding stronger language supporting imaging to that subsection.

2. A common challenge in evaluating patients with either elevated spike-wave indices or concerns for regression is determining response to treatment. A core component of this entails improvement in neurocognitive and behavioral difficulties. Neuropsychological testing should be strongly suggested at baseline and used regularly to evaluate response to therapy. You mention NP testing in regards to follow-up testing, but I would stress it should be performed at baseline as well. While neuropsychological testing is a limited resource, we should encourage its use where available.

3. For Table 10, under “Development at onset” you mention regression or plateauing, but there is no mention of plateau in the discussion – merely regression.

4. I lament that D/EE-SWAS does not exactly roll of the tongue, but I respect its descriptive nature. I tried to come up with something more ‘catchy’ and my front runner is Sleep Potentiated Spiking With Epileptic and/or Developmental Encephalopathy (SPikEDE).

Kevin Chapman

15 July 2021

Thank you for all the work!

A) On self-limited epilepsies:
Some comments:
1) The new syndrome names are heterogeneous, in their wording and meaning:
One is EEG based (CTS), two are seizure/symptom based (autonomic visual), one is anatomic localization based (occipital lobe).

2) The word childhood (versus neonatal-infantile) is not used anymore, only in “childhood onset” visual epilepsy. We lose this important characteristic.

3) The word self-limited is only in the label of 2 childhood syndromes: why?

4) Self-limited does not translate well in many languages; has this been considered? Age-limited? Time-limited?
Suggestion: Childhood self-limited epilepsy with visual seizures/with autonomic seizures/with rolandic seizures/with photosensitive seizures seems more straightforward.

"SeLECTS may be seen in children with a history of prior neurological injury or intellectual disability, however, these features are considered coincidental and not causal": this is not clear.

Fragile X in the epilepsy DD list: 1) this is not the same level as "other epilepsies" 2) other genetic abnormalities than Fra X can lead to developmental delay and EEG/clinical features of SeLECTS (for intsance del16p11.2...) so that it would be better to comment in the paragraph of genetics that when there is a developmental delay prior to epilepsy onset, genetic etiologies such as FraX and others have to be looked for.

In COVE: the DD is focal visual seizures due to structural abnormalities, and occipital epilepsy with calcifications and celiac disease and MELAS are very rare examples of a subcategory of these (immune structural/metabolic structural)

B) Childhood D/EEs
The new label DEE is not clear here: the word developmental is meant to be added to epileptic encephalopathy when there is a difficulty to know what is due to the etiology and the epilepsy, when seizures are numerous, refractory and start early. This applies best to genetic diseases affecting brain development. It may also apply when development is affected early by an external agent, for instance to an infant with a history hypoxic ischemic encephalopathy and epileptic spasms, but it would, in my view, not apply to the same infant with HIEE or a 4 year-old child witha prenatal stroke, episodic focal seizures and and “quiet” EEG, since the EE component is lacking.
This becomes more difficult in older children who have shown a window of development before epilepsy onset - be it normal or abnormal. The temptation is to mix both terms and ask no more questions. All the efforts we made to try to disentangle the role of the underlying pathology and the epilepsy are at risk to disappear. An example I have seen many times: for the non-specialist, every child who has a delay and seizures is now a DEE (bad brain + bad development + epilepsy), and this causes confusion and misinterpretations.

The syndrome D/EE-SWAS is in this aspect unsatisfactory:
1) Developmental: the acquired part, the regression which is cardinal in this syndrome disappears in the word “developmental".

2) Epileptic encephalopathy: this obscures the more specific nature of the language/cognitive/motor/behavioral deficits, which is often quite different from the more global impairments seen in LGS and other syndromes. The acquired aphasia (which is not limited to auditory agnosia) and its spectrum, the opercular, the frontal or the motor manifestations (gait disorder of the previously “atypical CECTS") disappear. All the richness of the semiology of these syndrome continuum which is unique in epileptology is “drowned” by two vague terms.
This continuum - clinical, neuropsychological and electrophysiological - with self-limited epilepsies - is not acknowledged enough in the text.
When there is a lesion, the prognosis of the epilepsy after puberty is more reserved.

Suggestion: “self-limited (childhood) focal epilepsy with spike wave activation by sleep” and specify with (acquired) aphasia or opercular or gait or behavioral disorder or whatever is the main or global impairment. When there is a lesion, omit “self-limited” and specify "structural".

C: On genetic generalized epilepsies in childhhood
"Among the more severe DEE, syndromes often have rare genetic etiologies (such as Angelman syndrome, 15q inversion-duplication) typically arising de novo in the patient": This is NOT clear at all in this context. On LGS: "Rare metabolic disorders may lead to a LGS phenotype..." NCL is rather a genetic neurodegenerative disorder...

D:The inclusion of FIRES and HHE in childhood epilepsy syndromes is a bit puzzling without more explanations: these are a particular situation of acute refractory SE followed by sequelar epilepsy. Shouldn't these be rather classified in SEs – ie. a sub-category of “new onset refractory status epilepticus”?

Eliane Roulet-Perez

14 July 2021

As Dr Camfield, I completely disagree with the extended use of abbreviations such as POLE, SELECT and so on.

Charlotte Dravet

14 July 2021

Thank you ILAE Nosology and Definitions Task Force for such a astounding summary on Childhood-Onset Epilepsy Syndrome.

In Self-Limited Focal Epilepsies of Childhood (SeLFE) syndromes,
The terminology POLE and COVE syndrome are nearby in position. As this two syndrome have good prognostic values, to exclude them from focal structural Occipital Lobe Epilepsy (OLE), the investigation of MRI and other test if possible should be done for presurgical evaluation in childhood.

In Genetic Generalized Epilepy Syndrome of Childhood,
Clinical course of Epilepy with Eyelid Myoclonia and Epilepsy with Myoclonic Absence vary from normal to borderline functional & intellectual disability. EEG background is normal in the discussion of each of this two syndrome. But the findings EEG background wasn’t in Table 6 and 7. The correlation between normal EEG background and developmental & cognitive disability should be addressed.

In LGS ,
In EEG diagnostic criteria, generalized slow spike wave (

I strongly agree with the terminology about the syndrome “Developmental and Epileptic encephalopathy with Spike-wave Activity in sleep (D/EE-SWAS)” which was previously called LKS, Epileptic Encephalopathy with Continuous SWAS. As it has abnormal developmental regression features due to underlying cause, in addition to an epileptic encephalopathy.

This is all I have to say. Thank you very much for giving me an opportunity to comment on this paper.

Kazi Jannat Ara

15 July 2021

The classification has become more elaborate and appears great.

  1. In the diagnostic criteria of SeLECTS and MAE, mandatory criteria include EEG features. However, in resource poor settings it was stated that the syndromes can be diagnosed even in the absence of EEG and MRI (if they are satisfying mandatory and exclusionary criteria). This statement is contradicting itself as EEG is an integral part of the mandatory criteria. Is this about the ictal EEG? This needs to be clarified.
  2. In SeLEAS, EEG criteria are mentioned as mandatory. However, a substantial number of this subgroup can have normal routine EEG recording. Doesn’t this lead to underdiagnosis of this syndrome? Kindly clarify.
  3. In case of D/EE-SWAS, the phrase ‘continuous slow spike-wave in >50% of non-REM sleep’ has been mentioned. How prudent is it to use the term ‘continuous’ for activation noted in >50% of the record? The term has been traditionally used to indicate >85% of the record.
  4. The electrographic criteria for ‘continuous activation in sleep’ must be defined.
  5. The importance of sleep architecture in the diagnosis of D/EE-SWAS has not been discussed.
  6. Does the presence of treatment refractoriness (after an initial diagnosis of SeLFE) call for a syndromic re-classification?

Sanjeev V. Thomas

15 July 2021

This is an excellently written article.

* Rasmussen encephalitis can be included under DEE in this age group.

A. R. M. Sakhawat Hossain Khan

15 July 2021

This submission pertains only to the entry regarding Developmental and/or Epileptic Encephalopathy with Spike-Wave Activation in Sleep (D/EE-SWAS). I am writing on behalf of several researchers and clinicians with specific interest in this condition.

Comment#1: Proposed syndrome name
An attempt to subsume the varied nomenclature of all conditions associated with ESES is admirable and very challenging. We applaud the attempts to subsume the veritable alphabet soup of abbreviations into a unifying syndrome that requires a temporally-related developmental regression. The section of epilepsy working to understand the entity of abundant epileptiform discharges in the sleep state has been challenged with an identity crisis since the incipient descriptions of nearly continuous spike-wave in sleep in children with severe neurodevelopmental phenotypes by Patry et al in 1971. The ILAE committee makes the separate distinctions that the EEG pattern is defined as continuous spike-wave in sleep (CSWS) and that the clinical entities in association with such an EEG pattern include ESES, Landau-Kleffner syndrome (LKS), epileptic encephalopathy with CSWS (EE-CSWS), and atypical benign partial epilepsy (pseudo-Lennox syndrome). The verbiage acquired epileptic aphasia/agnosia and the proposed Penelope Syndrome eponym (Tassinari, 2009) are not mentioned but may be appropriate to further include in the list of other nomenclature.

The ILAE committee does not address the widespread discordance of definitions when approaching ESES as the EEG phenotype and CSWS and LKS as clinical phenotypes (Nickels and Wirrell, 2008; Loddenkemper et al 2011; Sánchez Fernández et al 2013; Eeg-Olofsson et al 2013). This confusion and lack of unifying nomenclature has been a significant limiting factor to substantial scientific progress, collaboration, and understanding.

Lumping epileptic and developmental encephalopathies together may be problematic. The epileptologists’ focus should pertain to the epileptic encephalopathy wherein treating activated spike-waves spikes may make a difference, whereas aggressive antiseizure treatment is unlikely to be helpful in developmental encephalopathies (Kalser and Cross, 2018). The importance of this last point reflects when and how a physician will attempt to treat the SWAS pattern, which is presumptively different between developmental and epileptic encephalopathies. The greatest (or only) confidence we have as practitioners is when there is a defined timeline of developmental and/or behavioral regression associated with an EEG demonstrating new findings of significant spike-wave activation in sleep.

A definite concern that arises in subsuming all conditions of LKS and CSWS/ESES under D/EE-SWAS is the potential loss of distinction between potentially pathophysiologically distinct conditions with a shared/similar EEG phenotype and perhaps shared phenomenon of thalamocortical network dysfunction. Patients with LKS and (EE-)CSWS can be rather highly divergent in regards to clinical presentations, from the specific neurodevelopmental regression phenotypes to a broader clinical seizure phenotype. There are reports supporting an immunologic/auto-inflammatory etiology (van den Munckhof, et al 2016), which may help to support some suggestions of hormonal therapy and IVIG efficacy in these patients. It is unclear if this is isolated to patients with LKS. Furthermore, there are prior definitions of LKS with EEGs negative for SWAS (or not reaching >50%) but deficient frequency modulated auditory evoked responses (Duffy et al, 2013). There are far too many letters in the current abbreviation to append prior clinical correlate (e.g. D/EE-SWAS-LKS) but it may be important for clinicians and researchers to delineate those with predominant auditory aphasia/agnosia phenotypes.

Comment#2: EEG Pattern Definition
The group offers a replacement terminology of “spike wave activation in sleep” (SWAS) to define the EEG phenotype. This “softening” of terminology like ESES (when it is used to define an EEG only) is important; families and practitioners struggle to understand the clinical significance of “status epilepticus of sleep” when it has no defined clinical manifestations.

The ILAE committee remarks that the literature often defines the threshold as >85% of slow-wave sleep, however further mention lower percentages may produce clinical phenotype and offer “>50% of slow sleep” as another common feature. Guidelines for EEG related to ESES/CSWS in children were provided that supported a broader phenotype than the classic >85% threshold of slow-wave sleep (Scheltens de Boer, 2009). This proposed guideline suggested a spike-wave index of at least 50% in non-rapid eye movement (NREM) sleep with varied localizing (e.g. focal, symmetric bilateral, multifocal, etc.) and temporal patterns (e.g. [sub]continuous, fragmented, periodic). While the hypothesis of impaired slow-wave sleep downscaling (Bolsterli et al, 2011; Cantalupo et al, 2011; Issa 2014; Tanritanir et al 2020) may be crucial to some of the clinical phenotypes, many publications support spike-wave quantification in earlier stages of NREM sleep (Sánchez Fernández, et al 2012; Weber et al, 2017; Reus et al, 2020). Furthermore, some patients demonstrate such significantly activated spike waves in sleep which obliterate any normal features of NREM sleep (e.g. no discernible slow-wave sleep), thereby limiting confidence in scoring a spike-wave index based on the classic definitions.

It is suggested that the ILAE committee adjust the definition of the SWAS pattern to “>50% in NREM sleep, which may predominate early NREM, slow-wave sleep, or both.” The ILAE committee acknowledges that “epileptiform discharges during wakefulness are not continuous” but perhaps should amend this to state that epileptiform discharges during wakefulness should not be >50%. There is no established degree of activation from wakefulness to sleep that has been agreed upon to constitute a clinically important EEG sleep pattern, however (e.g. does a spike-wave index increase from 10% in wakefulness to 60% in sleep have more clinical significance than an increase from 35% in wakefulness to 60% in sleep?).

Comment#3: Clinical Context and Table 10
The ILAE committee defines D/EE-SWAS with onset between 2 and 12 years of age. While this is the most common age group for the development of the SWAS pattern in patients, there are descriptions of younger patients with this EEG pattern with developmental regression prior to the age of 2 years (Gong et al, 2021); these patients with defined ESES pattern prior to age 2 years all had de novo pathogenic variants in KCNQ2. In patients with known developmental and epileptic encephalopathies, it is important to still consider the contribution of activated spike-waves in sleep and how those may manifest in negative developmental or neurobehavioral phenotypes. It may be more appropriate to expand the Table 10 age of onset definition to to include ages 1-2 years as an “Alert.”

It may be further worthwhile to place “(nocturnal) tonic seizures” as at least an “Alert” for the type of seizures which should prompt consideration of another diagnosis such as Lennox-Gastaut syndrome. Some authors would argue that tonic seizures as exclusionary to a diagnosis of CSWS (Tassinari et al 2000; Sánchez Fernández et al, 2012).

Comment#4: Differential Diagnosis
The ILAE committee does mention that several other conditions may demonstrate the SWAS EEG phenotype. One of the greatest challenges in this field is parsing out the significance of SWAS in children with autism or in whom a known genetic developmental or epileptic encephalopathy exists - the ILAE committee wisely leaves this distinction to the clinician and so it shall remain a clinical question attempting to delineate the significance of SWAS when encountered, which at least will no longer have a connotation of seizure activity with “ESES.”

Comment#5: Figure 8
Figure 8 is somewhat misleading in that it still demonstrates this discrepant EEG definition of “~85% of non-REM sleep” as the far end of the spectrum around which Encephalopathy with CSWS and Landau-Kleffner bubbles are located. The image suggests that this 85% number has a pathophysiological basis of significance greater than that of ~50%; that has not been proven. It may be more appropriate to have on the color-graded arrow to state “>50% of non-REM sleep” and at the far right end of the spectrum include “bilateral synchronous” and “continuous” whereas the left may show “focal or independent bilateral spikes” and “fragmented” and “periodic.”


  • Sánchez Fernández I, et al. Clinical staging and electroencephalographic evolution of continuous spikes and waves during sleep. Epilepsia 2012;53(7):1185-95.
  • Sánchez Fernández I, et al. Patients with electrical status epilepticus in sleep share similar clinical features regardless of their focal or generalized sleep potentiation of epileptiform activity. J Child Neurol 2013;28(1):83-9.
  • Sánchez Fernández I, et al. Electrical status epilepticus in sleep: clinical presentation and pathophysiology. Pediatr Neurol 2012;47:390-410.
    Eeg-Olofsson O, et al. The way out of Babel. Epilepsia 2013;54(4):767-8.
  • Patry G, et al. Subclinical “electrical status epilepticus” induced by sleep in children. A clinical and electroencephalographic study of six cases. Arch Neurol 1971; 24:242–252.
  • Tassinari CA, et al. Encephalopathy with status epilepticus during slow sleep: “the
    Penelope syndrome”. Epilepsia 2009; 50(Suppl. 7):4–8.
  • Cantalupo G, et al. Night-time unravelling of the brain web: impaired synaptic downscaling in ESES – the Penelope syndrome. Clin Neurophysiol 2011;122:1691–2.
  • Tassinari CA et al. Encephalopathy with electrical status epilepticus during slow sleep or ESES syndrome including the acquired aphasia. Clin Neurophysiol 2000;(Suppl. 2):94-102.
  • Sánchez Fernández I, et al. The tower of Babel: survey on concepts and terminology in electrical status epilepticus in sleep and continuous spikes and waves during sleep in North America. Epilepsia 2013;54(4):741-50.
  • Loddenkemper T, et al. Continuous spike and waves during sleep and electrical status epilepticus in sleep. J Clin Neurophysiol 2011;28(2):154-64.
  • Nickels K and Wirrell E. Electrical status epilepticus in sleep. Semin Pediatr Neurol 2008;15:50-60.
    Scheltens-de Boer M. Guidelines for EEG in encephalopathy related to ESES/CSWS in children. Epilepsia 2009;50(Suppl. 7):13-7.
  • Tanritanir A et al. Slow wave activity during NREM sleep in patients with electrical status epilepticus in sleep. Seizure 2020:80;257-8.
  • Reus EE, et al. Determining the spike-wave index using automated detection software. J Clin Neurophysiol 2021;38(3):198-201.
  • Gong P, et al. Genetic etiologies in developmental and/or epileptic encephalopathy with electrical status epilepticus during sleep: cohort study. Front Genet 2021;12:607965.
  • van den Munckhof B, et al. Serum inflammatory mediators correlate with disease activity in electrical status epilepticus in sleep (ESES) syndrome. Epilepsia 2016;57(2):45-50.
  • Kalser H and Cross JH. The epileptic encephalopathy jungle - from Dr West to the concepts of aetiology-related and developmental encephalopathies. Curr Opin Neurol 2018;2:216-22.

Robert Stowe, Sonal Bhatia, Sarah Aminoff Kelley, Elia Pestana Knight and Anthony Fine

14 July 2021

What about epilepsy with perioral myoclonia only?

How to classify a child with a brain lesion, previous some neurological abnormalities and then he stops to make progress, even there is some regression when you record CSWS and again improvement when you can control this EEG pattern?

van Rijckevorsel Kenou

14 July 2021

Much good can be said about this proposal. My comment concerns the “Self-Limited Focal Epilepsies of Childhood (SeLFE) syndromes” where a definition of focal epilepsy is missing. The term is not self-explanatory. The only available ILAE definition is that of the 1989 Classification of Epilepsies and Epileptic Syndromes which reads:

“Localization-related (focal, local, partial) epilepsies and syndromes Localization-related epilepsies and syndromes are epileptic disorders in which seizure semiology or findings at investigation disclose a localized origin of the seizures. This includes not only patients with small circumscribed constant epileptogenic lesions (anatomic or functional), i.e., true focal epilepsies, but also patients with less well-defined lesions, whose seizures may originate from variable loci. In most symptomatic localization-related epilepsies, the epileptogenic lesions can be traced to one part of one cerebral hemisphere, but in idiopathic age-related epilepsies with focal seizures, corresponding regions of both hemispheres may be functionally involved.”

This terminology and definition were the result of long and detailed deliberations by the 1981-1989 ILAE Commission on Classification and Terminology. We took great care to distinguish “true focal epilepsies” with circumscribed constant lesions as they had been described by Jackson and Horsley from other conditions with focal seizures. We were particularly concerned about the self-limited Rolandic epilepsy of childhood where no lesion is found and both seizures and EEG discharge often alternate between sides (this highly significant feature underscoring the non-focality of the condition is missing in your description!). On the other hand, the seizures in this syndrome were not generated at random but displayed a specific relation to the primary motor cortex, bilaterally. To provide a place for this unique situation we introduced “localization-related epilepsy” as a preliminary term until new research would provide a better understanding of the pathophysiology and eventually suggest a more suitable name. This has not yet happened.

The ill-suited classification as focal of the self-limited syndromes in this position paper has two main negative consequences:

1. It counteracts the important message of the ILAE that, fundamentally, all patients with focal epilepsies should be considered as potential candidates for surgery so this option is not overlooked when it becomes the therapy of choice.

2. It confounds the nosology of the epilepsies as it throws the self-limited localization-related epilepsies into one category with lesional focal epilepsies although, apart from the focal seizures, they share many more features with the “generalized” epilepsies (e.g. age-relation, lack of gross morphological changes, bilaterality, genetic background, favourable treatment response, uniform syndrome-specific rather than individual semiology).

Peter Wolf

12 July 2021

I would like to give a huge thank you to the nosology task force chairs and members for the incredible amount of time and effort they have put in to get these massive documents to this stage. They have done an incredible job and should be congratulated. I hope they are feeling appreciated. These documents are well written and the format will be very useful for clinicians with all levels of epilepsy expertise.

The task of delineating all of these epilepsy syndromes was enormous. Defining the core features and edges of these syndromes is difficult. Experts will vary on their opinions as to where these boundaries exactly lie. The aim presumably is to provide a framework that will include the majority of cases that “have” the syndrome and exclude the majority of cases that “do not have”. I am using quotation marks because clearly there is no truth here! Having reviewed the manuscripts, I think the task force has achieved this aim very well.

I do have a few comments regarding DE/EE-SWAS. While the boundaries provided for this syndrome will capture all of the cases that “have” the disorder I am concerned that they may be interpreted in an unintended way and capture cases who do not have this syndrome (or at least cases who it was not intended to capture in the boundaries). If you step back and look at what the mandatory and exclusionary features are (ie slow (<2Hz) spike-wave discharges in >50% of non-Rem sleep that are activated by sleep temporally related to cognitive behavioural or motor regression or plateauing without epileptic spasms) it seems to me that definition is so wide that it will include children with DEE who have a bad EEG which gets worse in sleep – that is reasonable number of children with DEE. A lot of kids with DEE will have activated EEGs in sleep and >50% of the recording isn’t that much – and by definition all DEEs have developmental plateauing or regression. There is note specifically regarding LGS not being part of DE/EE-SWAS but there are other children with DEE (that do not fit into specific other DEE syndromes) who I think most epileptologists would not previously have considered to be part of the epileptic aphasia spectrum that might now be classified in this new group. I understand the rationale for making the boundaries wide here but it is possible with this approach the very unique syndromes of LKS and Epileptic Encephalopathy with Continuous Spike and Wave during Sleep will be drowned within a bigger group of non-specific DEEs whose EEGs get worse in sleep. This is something for the group to consider.

With regard to the name of the syndrome I am a bit lost as to why there is a backslash in the title – I think it might be to say it can be a DEE or a EE? I didn’t think this was clear and it doesn’t appear to be consistent with the way DEE is used in other disorders (for example Infantile Spasms Syndrome) or within the document (for example it says “and/or” in the syndrome section in the text but just “and” in Table 10.

Lynette Sadleir

12 July 2021

Thank you for the well-constructed paper.

The tables are clear and easy to read. However, in many of the syndromes the normality/abnormality of the EEG background is an important feature, but is not described in most of the Tables (Table 2-10).

SeLECTS and SeLEAS: In description of both of these syndromes, the spikes are described to be “high amplitude”. This should be clarified: what constitutes a high amplitude spike?

The EEG examples are very nice in this kind of paper; however, the quality of the figures was poor and the figures could not be read. The sensitivity and time scale markers were not correct in many figures.

Päivi Nevalainen, Maria Peltola, Leena Lauronen, Jukka Vanhanen and Hanna-Reetta Lajunen

12 July 2021

In this Epileptic syndromes of childhood onset in which secondly addition of post infection, trauma, malignancy, etc. associated causes for the epileptic events can be mentioned. This may tell the Dual causes for epilepsy in the childhood onset of epileptic syndromes.


12 July 2021

The terms used in this paper are self-explanatory in nature. Clinically you many times get an overlap syndrome between SeLECTS and D/EE-SWAS. Can we mention this overlap syndrome in the classification? Most of these patients were classified as atypical BRE initially.

Sandeep Patil

9 July 2021

You must have thought very hard as whether to rename LGS. In a time when the L-L lecture at the AES has been renamed - is it still appropriate that this epilepsy syndrome carries his name?

The LGS definition is excellent, however it does favour making the diagnosis in childhood and we need a way to diagnose LGS in adult clinics too. The characteristic EEG patterns are not so readily discernable (even if I can get a young adult with ID and ASD to the EEG lab for a sleep study) when on polytherapy and in adulthood. Are EEG features mandatory for adult diagnosis, as the paper states slow spike and wave is rare in adulthood.

Rhys Thomas

4 July 2021

The Task force members deserve huge applause for such comprehensive and laborious effort for classification of the epilepsy syndromes in childhood. I welcome most of the changes (like the use of SeLECTS and SeLEAS)

However, I have the following suggestions:

  1. It Is confusing to use MAE (Doose syndrome) and E-EM (Jeavons) as both have overlapping letters, hence it will be preferable if some other terminology can be used.
  2. I do not agree with the term D/EE-SWAS for Developmental and/or Epileptic Encephalopathy with spike-wave activation in sleep (D/EE-SWAS), (previously called CSWS/LKS) as often there is no developmental encephalopathy (i.e. prior cognitive or speech delay) so clubbing these entities with other DEEs like channelopathies will give a very wrong impression. Also, often there is a good scope of improvement with steroids, sleep EEG is important to diagnose and behavioural abnormalities, learning disability, hyperactivity are predominant.
  3. The figures depicting the three categories of syndromes are very good, but the EEGs are difficult to read. It will be better if we have colored pages of EEG as well (e.g. right sided depicted in red, left in blue).

Kavita Srivastava

2 July 2021

Under Lennox Gastaut Syndrome (LGS) you state:

Pathogenic variants in many genes have been associated with the etiologies that causes LGS and
are usually de novo in the child 146,147. A range of chromosomal abnormalities and copy number
variants have been associated with LGS, so chromosomal microarray is essential. A range of next
generation sequencing approaches can be taken, ideally with whole exome sequencing, or an
epilepsy gene panel, particularly if no etiology is found after clinical examination and MRI.

It is important to note that ring chromosomes cannot be observed using chromosomal microarray, nor NGS, exome sequencing or epilepsy gene panel testing. In the majority of cases of ring chromosome 20 syndrome, there are no losses or gains of DNA to be observed using any of these techniques, therefore it is also important to run a karyotype for 50-100 cells (in case of low level mosaicism) to diagnose r(20) syndrome where suspected and other genetic tests have given a negative result.

r(20) syndrome should be listed under the Differential Diagnosis section for LGS.

Allison Watson 

1 July 2021

Trés bonne revue des syndromes de l'enfant. Les tableaux sont d'une grande valeur avec des shémas clairs et bien expliqués. Néanmoins, quelques remarques:

  • Ne pas inclure les Epilepsies Absence de l'enfant et de l' adolescent dans le groupe des épilepsies auto-limitées, peut prêter à confusion que sont deux groupes nosologiquement différents.
  • Traduction du terme du Self Limited en français autolimité peut faire croire que le patient peut guérir sans consulter le médecin.

Wafa Bouchaala

12 June 2021

Comprehensive and clinically relevant. The introduction of the term self limiting in the nomenclature would combat a lot of stigma against these conditions.

Arun Swaminathan MD

9 June 2021

I feel that the 4 syndromes classified under self limited epilepsy can be named in a uniform manner. COVE and POLE seems to stand apart. Why can't they be renamed as 1. Selflimited epilepsy with visual paroxysm and Selflimited epilepsy with photosensitivity respectively?

Pradeep Pankajakshan Nair

7 June 2021

I was not able to find Hypothalamic Hamartoma Syndrome mentioned in any of the papers. While it might be argued that it could be included in the Neonate and Infant definition as it is a structural DEE, it may be better described in the Onset in Childhood as that is when comorbidities become advanced enough and gelastic seizures may have progressed (and may be better recognized), therefore it is most often diagnosed.

Lisa Dunn Soeby

5 June 2021

The term “childhood epilepsy with centrotemporal spikes” in the present implies that this syndrome is limited within childhood. I believe that renaming it to “self-limited epilepsy with centrotemporal spikes” with exclusion of “childhood” should blur its concept as childhood-limited epilepsy.

Katsuhiro Kobayashi

3 June 2021

Thank you for this important work. I have comments about the terms, Epilepsy with Eyelid Myoclonia (E-EM) and Epilepsy with Myoclonic Absence (E-MA).

The clinical difference between a clonic vs myoclonic seizure is fairly well established. As stated under the seizure description of E-EM, the eyelid motor symptom consists of "of brief, repetitive and often rhythmic, 3-6 Hz myoclonic jerks of the eyelids..." It is more typical that the "myoclonus" occurs more than once consecutively, usually multiple times. Epilepsy with Eyelid "Clonus" (E-EC) seems more appropriate. Similarly, E-MA is described as "rhythmic 3 Hz jerks of the upper limbs, superimposed on tonic abduction of the arms during the seizure (giving a ratcheting appearance)." The rhythmic jerks typically last "10-60 seconds" and actually describe clonus. The term, Epilepsy with "Clonic" Absence (E-CA), seems more appropriate. The use of the term Clonus (and not Myoclonus) for both seizure subtypes will prevent confusion and may improve identification of these subtypes.

Jun T. Park

2 June 2021

In SeLEAS, EEG characteristics are only poorly described. If SeLEAS is a clinico-electrial syndrome, the EEG characteristics should be better described in a more detailed fashion as follows because the detailed interictal EEG study of Panayiotopoulos syndrome showed characteristic age-dependent change in localization of epileptiform abnormality (Ohtsu et al. Epilepsia 2003;44:435-442.)

Multifocal, high voltage spike- or sharp-and-slow-waves are typically seen, often over the posterior regions between 2 and 5 years of age and later the predominance moves to either centrotemporal region or frontopolar region with age progress. Generalized discharges may also be seen at the peak of EEG worsening.

Myoclonic-atonic epilepsy

Development before the onset:
MAE itself has long been considered primary generalized or idiopathic generalized epilepsy, similar to absence epilepsy or juvenile myoclonic epilepsy, because patients exhibit normal development and are neurologically normal without imaging abnormality in addition to genetic predisposition. Although Doose described that 16% of 117 patients with MAE demonstrated signs of developmental retardation before the onset of epilepsy or had high-risk factors in their history, the original case series of Doose potentially included heterogeneous epilepsy syndromes because he himself admitted that he refused to subclassify the syndrome by clinicoelectrical details. According to his original concept of this syndrome, the presence of even mild developmental delay should be also placed on the “Alerts” like childhood absence epilepsy.

Myoclonic-flexor seizures or massive myoclonic seizures causing drop attacks should be included in addition to myoclonic-atonic seizures as a main seizure type (Oguni H, et al. Epilepsia 1992;33(5):805-813; Adv Neurol 2005;95:157-174; Epilepsia 1997;38(7):813-818).

Interictal discharges comprised of generalized 3-6 Hz spike-and-slow-wave or polyspike-and-slow wave often occurring in bursts lasting 2-6 seconds are seen (Figure 6A). This expression of 3-6Hz SWC is not relevant for SWC frequency of MAE. Although Doose described typical EEG finding as 2-3Hz SWC, he potentially included patients with myoclonic epilepsy in infancy (presenting with 3Hz SWC) in his series. As MAE is categorized in DEE, the SWC frequency at least includes 2Hz SWC or lesser frequency (Oguni H, et al. Adv Neurol 2005;95:157-174). The 3-6Hz spike-and-wave or polyspike-wave is the same SWC frequency as that of JME, but it is doubtful whether MAE and JME have SWC of the same frequency.

Fig6BC is not typical ictal polygraph of atonic drop attacks because only brief atonia was recorded at bilateral Deltoid muscles.

Hirokazu Oguni

31 May 2021

I would like to make comments to the SeLECTS category.

It is more or less evidenced that the core syndromes (RE and Panayitopoulos sy) may transform to what is in the proposition named as „Developmental and/or Encephalopathy with spike-wave acrivation in sleep (D/S-SWAS) (incorporating previous LKS).

This would make a spectrum of the perisylvian network epilepsies with cognitive (especially causing language , or more widespread cognitive impairements, which is congruent with functions of this network, as a human communication network).The frequency of this devastating transformation is underestimated according my experiences, and the early recognition of it is highly important, threfore to dividing it fom the earlier and also present „benign” forms, hide certan dangers.

The naming the EEG abnormality as „spike-wave”and identiying this with the bilateral syncronous spike-wave pattern is misleading. In the encephalopathic forms the EEG discharges are not spike waves , but they could be identified as focal unihemspheric discharges with phase reversal,changing to bilateral appearence via callosal sencondary syncronisation mechanism. So they can be derived from CTS. The presence of augmented CTS in the encephalopathic forms is a strong argument in favour of the spectral togetherness of the core and encephalitic forms.The genetic results also supported same underlying gene mutation factors for at least in the non developmental (not underlied by brain pathology) group of patients.

Peter Halász

21 May 2021

Thank you for the substantial good work done. I will restrict my comments to the aspects I feel very strongly about.

I find many changes you have made rather confusing, particularly in context of teaching and allowing one's brain (my own brain and that of people learning or training) to compartmentalise things properly.

Therefore, I will make the following points:

1. Self limiting (SeL) is a good concept and better than benign but should be applied to ALL subsets not only to SeLECTS or SeLEAS, otherwise, there will be definite confusion to all if SeL is not used before all subsets.

2. I think it is not best/optimum or helpful to take out the words occipital & Panayiotopoulos from SeLEAS. This is an occipital epilepsy syndrome and Panayiotopoulos discovered and described it. The occipital spikes seen in this syndrome are a major helping part of diagnosing it and there is nothing confusing about saying it is early onset with more Autonomic features and the Gastaut variant late onset with CLEAR visual features, I think some breach of etiquette may be seen with taking the name of Panayiotopoulos out of his own syndrome.... something to think about seriously.

3. We now know that all epilepsies have genetic basis. I find using the word Genetic Generalised epilepsies and presuming people will know we mean previous IGE is ambitious and it also means that these are the ONLY Genetic generalised epilepsies... I always use the term "Genetically influenced" for both the Focal and Generalised Epilepsies previously known as IGE and IFE to separate them from the more serious or less benign generalised and indeed sometime focal genetic epilepsies such as Dravet, CLN2 and so forth...I think it should be considered to use the term genetically influenced for IGE and IFE families (by the way, genetically influenced as a term, I was inspired to use by your group too from earlier days so I give you credit for it and ask that you retrieve it back!). It is very useful to separate the self limiting (focal-IFE) or the usually with no serious co-morbidities (Generalised-IGE) from the more serious and/or not self limiting epilepsies.

4. I wonder f you would re-consider the rather complex title and its consequently crowded abbreviation:
"D/EE-SWAS". I am a believer that abbreviations and titles are made to simplify things and make them easy to remember.

Including "hyphen" and "forward slash" to titles is best avoided. It lends itself to all sorts of spelling errors and consequent incomplete searches of repeated titles on the net or in documents. One needn't crowd or cramp in ALL components of a syndrome in the title. I believe title are meant to be brief not detailed (details come under the heading or title, not in it) It becomes more like a deconstructed title or an unnecessarily deconstructed gateaux/pie.... with full acknowledgement of your conscious attempts to "spell out" things and make them easier and clearer

Thank you again for the continuing good work but serious reconsiderations are desired.

Amre Shahwan

13 May 2021

Comments on the prominent motor symptoms related to 3 Hz spike-waves (absences), like the perioral, or those involving head movement should also be inserted to help new beginners. Moreover, the relation of absence epilepsy with photosensitivity needs further clarification.

In POLE, the number of seizures is mostly low, and therefore ictal recording is not easy to be in hand. This could not be a reasonable diagnostic criterion, it is not required in typical cases satisfying all other criteria, which are well-presented. 

Betül Baykan

11 May 2021

Regarding the differential diagnosis of Self-Limited Epilepsy with Autonomic Seizures (SeLEAS), I'd like to add not only other medical disorders associated with intermittent vomiting, but also surgical conditions. Sometimes it mimics like an intestinal obstruction clinically.

Kyaw Linn

10 May 20201

I agree with the rationale behind the changes but there needs to be some improving internal consistency in the way that titles are constructed. This is harder to see in this paper compared to the other three.

Within the self-limited group, it is confusing to switch some to a title of self-limited but not all in that group. I can understand losing the terms idiopathic, but in one instance you have switched that to childhood whereas another you have given up the term childhood and replaced with self-limited. This is confusing to me. I would prefer an approach where we can see an improved internal consistency across the titles. These types of issues make the teaching and diagnostic coding around epilepsies really tricky to implement and also make it tricky for patients when understanding or describing their epilepsy type.

Given the pragmatic difficulties of ever increasing title lengths and the difficulties around how much of epilepsy syndromes characteristics to capture in the title, the retention of eponymous or acronym titles as legitimate alternative titles to me does seem sensible and where possible I would favour retaining these even if their more complicated non-eponymous title needs improving. The retention of some eponymous titles as practical alternative handles would help professional and patient communication in my view..

Given this, I would prefer an approach such as either

  • Self-Limited Epilepsy with Centrotemporal Spikes (SeLECTS)
  • Self-Limited Epilepsy with Autonomic Seizures (SeLEAS, Panayiotopoulos syndrome)
  • Self-Limited Occipital Visual Epilepsy
  • Self-Limited Photosensitive Occipital Lobe Epilepsy;


  • Childhood Epilepsy with Centrotemporal Spikes (CECTS)
  • Childhood Epilepsy with Autonomic Seizures (Panayiotopoulos syndrome)
  • Childhood Occipital Visual Epilepsy (COVE)
  • Photosensitive Occipital Lobe Epilepsy; (POLE)

Both of these approaches have an internal consistency in keeping with your rationale. On balance I favour the second list as it retains your rationale, avoids unnecessarily changing CECTS and I can see it working more easily in practice.

Colin Dunkley

7 May 2021

1. Regarding table 2 on Self-Limited Epilepsy with Centrotemporal Spikes (SeLECTS)

In the alert column its written usual seizure frequency more than daily..Doesn't it imply that SeLECTS has normally daily seizure frequency which we know isn't true and if a SeLECTS has a seizure frequency of daily episodes it should be considered as a major alert.

2. Regarding Developmental and/or Epileptic Encephalopathy with spike-wave activation in sleep (D/EESWAS)

I particularly liked its new nomenclature but in the description its written "The EEG pattern previously required for this syndrome was known as continuous spike-wave in sleep (CSWS) and the clinical correlate previously known as electrical status epilepticus in sleep (ESES). Isn't it the other way round? The EEG correlate was called ESES(obviously as term electrical is used) and the clinical correlate was called Epileptic encephalopathy with CSWS.

Regarding EEG in this syndrome its written "SWAS is usually diffuse but may occur more focally (typically
frontally) or multifocally"

I would like to point out that we have analysed EEG of more than 35 children of ESES and followed them for more than 3 years with serial EEG and neurocognitive assessment and their EEG were analysed in special source localisation and mapping software and we found that these apparently "diffuse" looking pseudo generalised spikes are infact in 100 % cases due to “almost”synchronous bilateral activation of spikes (with a lead-in of 20-40 ms from one hemisphere with shifting asymmetry ) with a tangential/oblique dipole .The rules of mapping localised all these spikes around Rolandic region(anterior and posterior walls).Morover the frontal predominance which you have mentioned is also seen due to the wrong interpretation of the spikes originating from posterior margins of the rolandic sulcus which has frontal negativity and tempero-occipital positivity with a positive phase reversal.

We have sent this study and accepted for poster presentation in AOEC 2021.I hope you consider this finding before concluding that the EEG pattern in this syndrome as "diffuse with frontal predominance".

One more thing which I wanted to add is that significance of the so called "developmental epileptic encephalopathy " in this syndrome should be infact more stressed. This in fact comprises the entity which we previously called symptomatic ESES(structural causes) and its more and more recognised than before.

3. Regarding Self-Limited Epilepsy with Autonomic Seizures (SeLEAS), "generalized discharges may also be seen"

It's seen that most of these "generalised" discharges are infact the bilateral occipito-frontal spikes which is seen in this syndrome and I hope you add this as an EEG finding in this syndrome. Also sometimes bilateral occipito-frontal spikes seen in this syndrome sometimes mimic ESES pattern if we are not aware of the clinical history.Therfore I feel that this valuable EEG finding be added in description.



5 May 2021

This series of four position papers represent a monumental contribution to the establishment of well defined epilepsy syndromes.

Corrective Feedback:
The Onset in Childhood paper has a typo on Page 60 (Figure 1) in that Occipital Spikes currently shows 4 to 7 years, but should be 4 to 17 as in the narrative. 

Regards, John McGinley