Neonatal Seizure Classification

This paper has now been published, read it here:
The ILAE classification of seizures and the epilepsies: Modification for seizures in the neonate.
Ronit M. Pressler et al. Epilepsia 20211:00:1-14

Read draft document, guideline process, and comments below:


The ILAE Task Force on Neonatal Seizures has created a classification for seizures in the neonate, which is intended to become the ILAE position on the classification for seizures in this age group.  As part of the approval and adoption process the League asks for its members to review and comment on the proposed classification.  These comments from our international community will then be reviewed by an independent task force, which will then amend the proposed classification to be in line with the suggestions from the community.

Read draft Neonatal Seizure Classification document here

By policy, the manuscript will be open for comments for two months, until October 15.  All comments will be placed  on the League's website as they are received and will include the name of each writer.  At the end of that time, the comments will remain public, and an independent task force will be named by the Management Committee to review the comments and to make appropriate changes.  This task force will have six months to complete the task and then submit the amended manuscript for publication.

We have been pleased and encouraged  by the extensive and thoughtful response that we received from so many of you about the previous position papers from the League, comments that led to extensive revisions.  We look forward to learning your thoughts, concerns and suggestions about the classification of neonatal seizures.

Please send your comments to Deb Flower, who will post them to this web page as they become available.

Thank you for your help in this important effort of the ILAE.

Edward Bertram, Secretary General, ILAE


21 October 2016

First of all, I thank the ILAE Task Force on Neonatal Seizure for their hard work on this challenging topic and for give us the opportunity to review this important document. 

Secondly, after reading the classification criteria and explanations carefully, as a non-expert in neonatal seizures I think that the classification at this age range is very clear,simple and easy to practice.

Thirdly, I totally agree that “electrographic only seizures”, without clinical signs, have to be included in the classification system: therefore neonatal epileptic seizures may be 1) electro-clinical or 2) electrical only seizures.

Last, I suggest If there is only evidence of clinical seizure without electroencephalogram, it is necessary to be cautious in diagnosing epileptic seizures.

Thanks again to the authors,
Cathy Zhang

21 October 2016

I think the paper is excellent and exactly what we need to better classify neonatal seizures! I have a few questions though: 

  1. "All neonatal seizures are focal": How should we read/interpret this statement which is repeated in the paper. I am thinking about the paper of Patrizi et al from Brain and Development 2003 (attached
  2. I think it is really good that the distinction between neonatal seizures and epilepsy is made throughout the paper. However, I don't understand the rationale for calling brief spasms "epileptic spasms" - are they connected to epilepsy? how? I think it is confusing in light of the rest of the paper. 
  3. I am very happy that you mention aEEG but I think it should be mentioned that only the aEEG trend is not very good for detecting seizures, the raw-EEG should be inspected. There are three references that better show this than the present ref 94 and that is Renée Shellhaas' two papers from 2007 and a more recent study from China demonstrating very similar results (Zhang et al). I can't find these papers at the moment, will look for them.

Best wishes,
Lena Hellström-Westas, MD PhD


20 October 2016

I read with interest the proposed new classification for neonatal seizures , It looks comprehensible , I have some suggestions to the committee

The definition of electrical seizure may not be practical ( the duration of epileptic discharges has to be sufficient to demonstrate the evolution in frequency, amplitude and distribution ...) as many neonates are encephalopathic and it may be difficult in some cases to define the” adequacy “.Suggesting a minimum time for the discharge may probably be more helpful.

Inclusion of neonatal status epilepticus definition and classification may be useful and valuable

For countries with limited access to continuous EEG , the current classification will leave many countries behind , and the optimal guidelines for management of neonatal seizures might not be achievable , further subclassifications might be useful and helpful

I thank the committee for the tremendous efforts and work.

Best regards,
Amira Masri , MD

20 October 2016

Thank you for providing the opportunity to comment on this interesting paper on the classification on Neonatal seizures.

The authors have done a hard work and gave good commitment to use available evidence to create a categorization of neonatal seizures. Moreover, their valuable work allowed an harmonization with the greater ILAE terminology, reflecting unique features of seizures in the neonatal age.

We would like to comment on the definition and diagnosis of "autonomic seizures" and, in detail, on the statement in table 1 according which EEG is mandatory to confirm autonomic seizures. We do not agree with this sentence, because the majority of autonomic neonatal seizures origines from subcortical cerebral areas, in detail from the limbic and perilimbic system, surely involving the perilimbic and periinsular corteces, but they do not always involve the cortical surface. As a matter of fact, literature data showed that the cortical layers are not still completely developed in preterm neonates until 40 weeks post-menstrual age. Moreover, connections between the brainsteam and the cortical surfaces are not still complete in preterm and late preterm infants. This explains the reason why the brainstem release phenomenon remains in a subcortical area and usually does not spread through the cortex. Therefore, autonomic seizures, above all in preterm infants, may originate from the limbic system secondary involving the perilimbic cortex, but they also may not involve the cortical areas (because the connections between the brainstem and the cortex is still immature and incomplete in preterm infants ). This means that this phenomenon cannot be evidenced by simple Video-EEG studies and need a deeper study of the autonomic system, including a polygraphic Video-EEG, with Oxygen Saturation study, heart rate variability indices, thorax and abdomend movements study, nasal air flow evaluation, and electromiography for the study of minimal movements.

In conclusion, with due to respect to task force members we would like express our comments to improve the diagnosis of autonomic seizures for which EEG is NOT mandatory for confirmation, because it might give a negative result even in the presence of seizures. Therefore, we propose to use the Poligraphic Video-EEG, emphasizing the study of the autonomia functions of the neonate.

Thank you for giving us the opportunity to comment on the proposal.

Best wishes,
Raffaele Falsaperla MD Catania), Giovanna Vitaliti MD Catania, and Giovanni Corsello MD Palermo (Italy)

17 October 2018

Here are some remarks from our group, led by Prof Neubauer.

Under Historical Review, regarding references no. 18, 38, 39: I think that all these studied used classical videoEEG so the sentence... either EEG or aEEG is not completely correct 

Regarding Evaluation of the proposal: Methods, "No minimum duration was specified as long as there was sufficient demonstration of evolution in frequency and morphology of the discharge." I believe that some time-frame should be mentioned (at least 5-second duration), as "sufficient demonstration of evolution in frequency and morphology of the discharge" is rather vague, and maybe not good determination for younger neonatal EEG interpreters.

Under Discussion:

  • Regarding "Clancy and Legido described electrographic only seizures in newborns as sudden, repetitive, evolving stereotyped waveforms with a definite beginning, middle, and end and a minimum duration of 10 seconds.
  • "However, the choice of 10 seconds duration was explicitly arbitrary:" This is not completely true: van Rooij et al* are citing:“a sudden, repetitive, evolving and stereotyped ictal pattern with a clear beginning, middle and ending and a minimum duration of 5–10 seconds" *van Rooij LG, van den Broek MP, Rademaker CM, de Vries LS. Clinical management of seizures in newborns: diagnosis and treatment. Pediatr Drugs. 2013;15:9–18. 
  • "Similarly an arbitrary minimum duration of 10 seconds is also applied to the definition of a seizure in critical ill adults:" I think this is not a good reason not to mention the timing of electrographic seizures - critically ill adult EEG has nothing in common with critical ill newborn EEG.
  • "Both in neonates and critically ill adults it has been suggested that rhythmic discharges of less than 10 seconds duration (so called BIRDs)..."  I agree for the BIRDs but not for ALL EEG recordings in neonates, and again do not compare with adults
  • "aEEG may be used, although its limitations are well recognized" Notes 94, 95: You are not citing reference in JCN 2017;32:815-822, where it was proven that aEEG is a poor screening tool for neonatal seizures.

 Regarding references, I could not find any reference from Linda de Vries and her group?

Best regards,
Damjan Osredkar

16 October 2018

The American Epilepsy Society (AES) embraces the International League Against Epilepsy’s (ILAE) revised classification of neonatal seizures and commends the taskforce for their work and efforts on this important topic. This revised classification provides an excellent framework for future work in this area. We ask that the taskforce members please consider the following feedback and questions prior to completion of the final classification document:

  • A description of a process for adjudicating disagreements on EEG readings or determining semiology is not described in the methodology section. Was there a process for adjudication determined prior to the study? If so, a description of the process and how often it was used will be helpful to readers.
  • In the tables, several acronyms are used without definition. Inclusion of the definition of the acronym with the table, even if it is provided in the text, will be useful for readers who may not be as familiar with these terms.
  • In Table 3, it would be helpful to give the range of gestational age (GA) and postmenstrual (PMA) at EEG.
  • Some of the most common seizure types are not highlighted in Table 4. For example, preterm neonates with non-motor electrographic seizures due to infection.
  • It would be helpful if additional significant etiologies of neonatal seizures such as neonatal drug withdrawal are discussed.
  • Address alternative classification for lower-income countries with limited or no access to EEG.
  • It would be helpful if the updated report could address neonates that have received anti-seizure medication or sedative medications prior to EEG.
  • AES concurs with your assessment of adding EEG to diagnose clinically silent neonatal seizures in both term and pre-term neonates.

AES appreciates the opportunity to provide feedback on this document, and with bated breath, we look forward to receiving the final updated version of the classification of neonatal seizures.

Shlomo Shinnar, MD, PHD, AES President

16 October 2018

Thank you for collecting comments on the new neonatal seizures document.

I commend the authors for this work, which reflects their hard work and commitment to use available evidence to create a categorization of neonatal seizures that harmonizes with the greater ILAE terminology, yet reflects the unique features seen in neonates.

I share the concerns voiced regarding BRDs, and the decision to consider BRDs as equivalent to seizures. I agree completely that the current “10 second” criteria is arbitrary and we need a better understanding of whether there is a physiologic basis for a duration threshold. However, I am concerned that swinging the pendulum to the opposite end of the spectrum, and declaring all evolving discharges as seizures (even if only one or two seconds in duration?) is not clearly justified. The consequences of such a definition change could be significant, resulting both in clinical treatment of BRDs with medications, and by muddying ongoing research efforts by changing the fundamental definition of seizure such that findings from prior studies are no longer directly comparable to those from future efforts.

Perhaps a helpful solution would be a temporary categorization of BRDs that notes their uncertain significance, emphasizing the need to address this issue through well designed research. Future revisions might better be able to clarify their status using results from ongoing work. A paradigm shift of this magnitude would best be supported by targeted evidence.

Courtney Wusthoff, MD MS

16 October 2018

I congratulate all the members of ILAE task force on neonatal seizure classification for their commendable job on this intricate subject. As Bass MJ et al described in 1986, classification systems are tools for sorting the complex elements of reality into reasonable and logical entities. The major objectives of medical classification systems today are to predict outcome, to predict response to specific therapies, and to communicate experience to others in research.

With due to respect to task force members I would like express my following comments to incorporate in to this classification which will be helpful in predict outcome, might guide clinician to investigate and choose therapies.

  1. As ILAE classification system are used worldwide, emphasis should be placed clinical classification rather than electrographic (video EEG based) classification. The current proposed classification system based on video EEG is difficult to practice as neonatal EEGs service is not readily available in many NICUs. Even if EEG is available, it is not extendable beyond 20 minutes in most centers. The video monitoring capabilities with EEGs are often lacking. Even if able to perform the video EEG, the interpretation of neonatal EEGs is very subjective and mostly restrained due to lack of experienced interpreters.

  2. The committee’s assumption of neonatal seizures is always focal and placing epileptic spasms and myoclonic seizures under the broad motor seizure category undermines the importance of clinically recognizing these forms of neonatal seizures secondary to early myoclonic encephalopathies. This group includes diseases with uniformly poor prognosis such as nonketotic hyperglycinemia and monogenic causes leading to catastrophic epilepsies like ARX, CDKL5, STXBP1, and MECP2 gene abnormalities. Early recognition of these form of neonatal seizures help clinician to better investigate, guide for specific therapies and perhaps overall prognostication.

  3. The task force rightly pointed out that there are evidences to suggest that outcome of electrographic seizures and electroclinical seizures comparable. However, the electrographic seizures are not exclusive to neonates and similar effects of electrographic and electroclinical are seen in sick patients across all age groups. Is the separate classification electrographic seizure needed.

Satish Agadi, MD

15 October 2018

First of all heartiest congratulations to the team who have worked hard to put out this description of classification for neonatal seizures.

The main concern I would like to raise would be the applicability in most parts of the world due to need for continuous monitoring. In my country, none of the neonatal units have even regular EEG facility, which speaks of the limitation of facilities in many parts of the world.

The second concern would be the practical issues of using this classification by the neonatologists who are the first in contact and often in control of the babies with neonatal seizures.

While appreciating the importance of the current classification, I would really wonder whether this classification further estrange the gap in treatment of neonatal seizures in resource limited settings.

Jithangi Wanigaisnghe, MD

15 October 2018

Congratulations to the ILAE working group for the great effort to present a proposal for the classification of seizures and epilepsy in newborns, based on the ILAE 2017 classification.

1 - I agree that the definition of the term neonatal period is confusing, and it should be reconsidered at 44 weeks of postmentrual age (PMA), regardless of whether it is a term or preterm infant.

2 - I agree NOT to include clinical events without EEG correlation such as epileptic seizures; and that electrographic seizures alone without clinical signs will be relevant within this new classification.

3 - Indeed, electrographic seizures are the most frequent in neonates with encephalopathies and it is very common to find both; electrographic convulsions and brief interictal repetitive discharges (BIRDs), of which some of these evolve and others do not.

I believe they should carry out research studies in this regard, to clarify and consider if the BIRDS are of epileptic etiology and should receive treatment with antiepileptic drugs.

4 - I am not sure if the term generalized onset seizures of the new classification should be eliminated, because these are observed in the epileptic syndromes.

Efraín Olivas Peña. MD

15 October 2018

Excellent work.

I think it would be interesting to have some examples of what is an electroencephalografic seizure and what is not. The examples should be presented in the form of EEG but also in the form of aEEG, as aEEG is more frequently available in NICU.

Best regards,
Ruben Rocha, MD

15 October 2018

This will serve as a helpful guide in description of neonatal seizures and will help in improving inter-rater reliability, education, and – I hope – treatment.

The authors have taken into account many of the complexities of diagnosis of clinical and electrographic seizures. The next is less a critique of the document than an acknowledgement of the difficulty in neonatal EEG. The role of the interictal background EEG upon which selzure-like activities occur is an important feature, in my opinion, that probably affected the group’s ability to use the classification in their test sample. Background EEG is important because, in my experience, paroxysmal discharges such as BIRDs or electrographic-only seizures are more easily picked up in the severely encephalopathic infant since they stand out from suppressed or discontinuous EEG; in the healthier, perhaps more active infant, brief or every focal discharges may be “hidden” to algorithms or to the eye. Thus, the classification may be biased to the most severe cases.

Mark Quigg, MD MSc FANA

15 October 2018

I’d like to thank the ILAE Task Force on Neonatal Seizures for their proposal for a classification which is clear and easy to use in daily practice. As in the proposal for older children and adults, the graphs are very informative.

I have one question though: In the introduction of the proposal (P6; line 6) BIRDS are defined as “Brief Rhythmic Ictal Discharges”, whereas in the discussion (p15; line20-21) they are defined as “Brief Rhythmic Interictal Discharges”. This is confusing.

As the taskforce proposes in the discussion that BIRDs without evolution are not considered seizures, by definition they would be interictal, whereas following from this, in case of clear evolution they might be ictal.

Therefore, I would like to ask the taskforce to make a statement on the definition and status of these Brief Rhythmic Discharges, with the suggestion to redefine them (again) to BRD, thereby, taking the ictal or interictal meaning of this phenomenon out of the discussion, as it probably is part of the ictal-interictal continuum anyway.

Thank you for the proposal, I hope it will be implemented soon!

Dr. Vivianne van Kranen-Mastenbroek, Clinical Neurophysiologist

15 October 2018

No doubt, there is a need for a Neonatal Seizures Classification like this. I thank the ILAE Task Force on Neonatal Seizure for their hard work on this challenging topic. Even if it’s not complete.

In particular, as already underlined in the draft, the clear distinction between: Seizure in the neonatal period with acute brain damage and therefore Symptomatic Epilepsy due to symptomatic neonatal seizures, Versus, Seizures without acute brain damage, so, Neonatal Seizures / Genetic Epilepsy since first onset and consequently Epilepsy from the beginning: Syndromes, Cortical Malformations, EIEE, and so on, is the most evident evolutionary concept that has come to delineate in recent years. On the other hand the number of extremely premature neonates has increased exponentially, and they act different signs and symptoms……. The Video electroencephalography is the gold standard, of course: but we need a video-EEG collection and a more accurate semiology description to complete the classification. Neonatal movements, as we know, are a very complex aspect, very different from adults. I think It is difficult without a more specific semiology description, to hypothesize a concrete diagnosis. The key role of video-EEG in the diagnosis of neonatal seizures is emphasized in the draft, but we forget that in a big number of NICUs all around the World, not always EEG is available 24 hours a day. aEEG/CFM alone is not enough…….

We must also consider that the Neonatal Seizures are often managed by Neonatologists at first, rather than by pediatric neurologists, in almost all countries. This classification must also be able to be consulted by neonatologists. In this direction, the description of semiology would be one of the most important aspects. Otherwise phenobarbital will remain the only ”jurassic drug”………. On the contrary I think there is a big difference between the treatment of acute neonatal symptomatic seizures and the treatment of a genetic epilepsy that starts from the first days of life: specific AEDs target or precision/tailored medicine, “the future”. How could be this new classification able to guide us in this direction ? Even the neonatologists? I think we need a precision description of possible sign and symptoms of neonatal seizures according to each gestational age.

On the other hands I totally agree with:

  • Clinical events without an EEG correlate are not to be included
  • Seizures in the neonatal period have been shown to have a focal onset, and so a division into focal and generalized is unnecessary
  • The arbitrarily 10 seconds chosen as the minimum duration to define a seizure must be reviewed

With my best regards,
Federico Raviglione, MD

15 October 2018

I think overall it is a good structured proposal and is a way forward.

However a few concerns.

  1. The whole classification depends on the availability of video-EEG. This may not be available in all settings and at all times.How will it work is such situations?
  2. First symptom vs most prominent symptom – this is causing difficulties in practice when looking at videos as the implications are different.
  3.  Sequential – this term is rather confusing and will probably end up as a default state for szs which cannot be accurately classified?
  4.  Automatisms – most  such events are not epileptic , so will cause confusion in the absence of EEG.

Thank you
Dr. K.B.Das

15 October 2018

I think the proposed classification on neonatal seizure was really good because it includes aEEG exam in making diagnosis. Clinical description of seizure was also straight-forward.

Yet that might be some consideration:

1. The duration of abnormality found in EEG, is it only if we found the discharge more than 10 s, what if it less than 10 s but occur repeatedly?

2. This classification may not be applied in developing countries considering limited availability on aEEG as well as EEG. what would you suggest to overcome this problem?

Best regards
Setyo Handryastuti Soekarno MD,PhD

15 October 2018

The correct diagnosis of paroxysmal events in neonates is always difficult and very challenging without the associated video- EEG especially when they are too subtle and subclinical. This proposed neonatal seizure classification of ILAE task force is very clear and pragmatic for clinical practice. We must recognize not only the motor but also the electographic only- seizures , non- motor presentation and some particular types seizures in this age group that can give clues to the etiologies.

Thank you for the great effort of The Task Force .

Dr Ng. Thanh Thủy

14 October 2018

Very clear and rational… Simple to use, easy to understand…

Unfortunately, there are some countries where the EEG correlate could be very difficult to obtain (availability of medical equipment, lack of expertise)

Congratulations anyway!

Dr. Luis E. Briceño C.

14 October 2018

Many thanks to the authors for all the hard work to harmonize with the 2017 ILAE seizure classification while making it meaningful for neonates!  I do have some questions.

How did the authors determine whether a seizure with more than one clinical seizure type could be classified as having a predominant clinical seizure type rather than a sequential seizure? Did the authors also note the initial clinical seizure type or was only the predominant seizure type noted?  And if the initial clinical seizure type was noted, how often was it different from the predominant seizure type?

I also wonder how it was determined that a seizure was autonomic.  How were the authors able to evaluate for blood pressure fluctuations or oxygen desaturations?  Did all eegs also have a respiratory channel?  If only the video, ekg, and respiratory channel were assessed, only apnea, bradycardia, tachycardia and possibly cyanosis would be detected.

Do the authors think it is possible that the low number of seizures with predominant behavioral arrest is due to a referral bias?  These subtle behaviors are less likely to attract attention than the other clinical seizure types.

What are the genetic etiologies? Were a large proportion due to KCNQ2 (self-limited familial epilepsy or KCNQ2 encephalopathy)? What are the types of inborn errors of metabolism?

I am not sure there is enough evidence to say that a BIRD is equivalent to a seizure and therefore should be treated with seizure medication.   In reference 27, some of their BIRDS would be called seizures today since 10 patients with BIRDS had clinical seizures with at least some of the BIRDS.   The outcomes combined neonates with BIRDS and clinical seizures and BIRDS without clinical seizures so it is not clear how many neonates with BIRDS and no clinical seizures had an abnormal outcome.  In reference 28, the neonates with BIRDS only did not develop epilepsy while 22% of those with BIRDS and seizures and 48% of those with seizures developed epilepsy.  In both reference 27 and 28, the recordings were typically an hour long.  Since many neonates with BIRDS also have seizures, a longer recording might reveal that those with poor outcomes actually have BIRDS and seizures.  Until there is enough evidence, it seems more prudent to retain the definition of a seizure as a minimum 10 seconds of evolving activity and BIRDS/BRDs as less than 10 seconds of evolving activity.  Of course, if there is a clinical change along with the electrographic seizure, there is no minimum duration needed. A minor point- I think there is a typographical error and reference 91 should be used for the sentence ”It has been suggested that definite BIRDs with an evolution represent “very brief” electrographic seizures”.

With regard to table 5- Electrographic seizures are seen more commonly in preterm infants than term infants (Glass, et al Ped Neur 2107; 72:19-24) but may not be typical. In the Glass 2017 paper, they are seen in 24% (reference 6 does not specify the percentage of preterm infants with electrographic only seizures).  Only reference 11 states that the majority of preterm infants have electrographic only seizures, and it is not noted how many infants had received a seizure medication prior to eeg recording. While the majority of preterm neonates in reference 85 had electrographic only seizures, this was determined using aEEG and it is not stated how many neonates received benzodiazepines for sedation at a dose that would be used to treat seizures.

There are also a few statements that I found confusing:

  • On page 3- In the definition for the neonatal period- I am confused by the wording for how to determine whether a preterm neonate is still in the neonatal period.  Would it be easier to state the neonatal period is 44 weeks PMA and younger for term and preterm neonates?  In terms of brain maturation, it seems that the term neonate born at 37 weeks and a seizure at 6 weeks of age will be more similar to the term neonate born at 40 weeks and a seizure at 3 weeks of age than a 45 week PMA infant.
  • Table 1- Clonic- Special consideration for neonates- To make it more clear you might change the statement to- More likely to be a seizure than other clinical seizure types.

Thank you again for all your work on this!

Tammy Tsuchida, MD, PhD

14 October 2018

I’d like to congratulate and also thank the ILAE Task Force on Neonatal Seizures for proposing such a draft. I think the two main useful aspects of this draft is that it is so clear as well as very practical. The tables and especially the three graphs are very well-designed and informative.

I would like to add some comments on this invaluable draft;

  1. The gold standard of classification is EEG, and there is no doubt that it is a justified means according to the nature of the neonatal seizures, but as you know there are many countries with limited health resources to afford modern digital video-EEG monitoring facilities for their nurseries and neonatal units. Could it be modified based on cheaper devices such as aEEG?
  2. You have proposed a very nice flowchart in Fig 2 which seems to be pretty simple but very practical, but there is a dead-end after “non-seizure episodes, without EEG correlate”!
  3. I would like to propose the term “Clinical only events” vs. “EEG only events” and I would like to raise a question in this regard. “Could we ignore this important section of paroxysmal events in newborns only because they are not recordable by scalp EEG with its all known limitations?!  

I am more than happy to contribute in this great work and Thanks again for this very practical and nice proposal.

Mahmoud Mohammadi MD

14 October 2018

I would like to thank you for this great initiative for classifying seizures and epilepsies in the neonate. I consider that this is essential given the peculiarities that present in the neonate I think the key points of this classification are:

  • The importance of the EEG in the definition of seizures,
  • The important role of seizures with electrical manifestation only, which are very frequent in the newborn.
  • and finally, to establish an association with the possible etiology, given that a high percentage of patients present a symptomatic cause 

On the other hand, I consider the clinical definition of the seizures relevant, giving importance to the predominant manifestation; as well as, the novel definition of sequential seizures.

It think that it is very correct to eliminate the minimum duration (10 seconds) of the definition of seizures, given that it was an arbitrary concept

It seems to me that it might be convenient to consider adding to the classification the seizures with clinical manifestation only, especially when they appear in neonates with interictal paroxysmal activity in the EEG and there is a symptomatic etiology known.

In relation to table 4 (Seizure types according to etiology): I think autonomic seizures  have been included by mistake in motor seizures. They should be included in non-motor seizures. On the other hand, electrographic only seizures would be better if they appeared in a different column than non-motor seizures. 

Thank you very much for this great proposal. Congratulations to the Authors

Patricia Navas Sánchez, Clinical Neurophysiologist

14 October 2018

As a non-expert in paediatric seizures, the new classification seems articulate and in keeping with the revised classification of adulkt seizures which was formalised last year. Great work and thanks to the team for the update.


14 October 2018

First of all, heartiest congratulations to the whole team for taking up this clinically relevant task and proposing a simple, yet concise classification in concordance with ILAE 2017 classification of seizures and epilepsy.

Many of the novel features which have been introduced by experts such as electrographic seizures only and sequential seizures are welcome decisions and scientifically justifiable. Only few suggestions I would like to make which I think could be included in the draft

  1. Clinicians all over the world have been widely using the term subtle seizures. The current proposed classification abandons that term, although many of them may be included in seizures with automatism and behavioral arrest. At least this piece of information should be mentioned in the draft to provide a continuum from the previous classification. Moreover, these subtle seizures are usually without an electrographic correlate, hence these may not even be considered seizures as per the proposed classification. Also, authors can highlight pointers in semiology (for subtle seizures) which were associated with electrographic correlate. Decision of treating subtle seizures (with electrographic correlate) will be more important in LMICs where they are more common (HIE being common). What is suggested where EEG may not be available with the neonatologists in peripheral centers in LMIC and those centers which have amplitude integrated EEG monitoring?

  2. It has been mentioned in the draft, and also a well proven a fact that many of the neonatal seizures have no electrographic correlates and electroclinical uncoupling further increases after antiepileptic drugs like Phenobarbitone. Although this is only a classification scheme, but at least briefly it should be mentioned that whether these seizures should also be treated and what are the antiepileptic drug choices in similar conditions, for guiding neonatologists all over the world as they may be the primary physician of care in neonatal seizures. On the other hand, the new term electrographic seizures have been included and the indications for doing video EEG in critically sick newborns have been mentioned. But this needs to be further specified, in which newborns we should plan for routine video EEG monitoring to detect these electrographic seizures only and what will be the treatment decision on detecting these electrographic events, without any clinical seizures. Whether brief BIRDs should be treated needs to be specified.
  1. Although there is mention of unclassifiable seizures in the classification, there may be a need to mention seizures of unknown onset to include those seizures which were noticed some time after the onset by the neonatologist or if detailed history from caregiver is not forthcoming.
  1. This is a video EEG based classification which may be difficult to stick to globally, especially in resource constrained settings in low- and middle-income countries, especially in NICUs where experts on neonatal seizure are not available easily round the clock. For such instances, an addendum of a simpler classification based only on seizure semiology in the clinical context of the newborn will be extremely helpful.

Best wishes to all of the team.

Thanks and regards,
Professor Sheffali Gulati

14 October 2018

I totally agree with proposal of classification of neonatal seizures. It's not an easy task! Well done work, which takes account of the etiology, clinical manifestation, EEG, epileptogenesis etc.

It shoud be the next step for clarifying this issue.

Thank You!

Best regards,
Beata Jędrzejczyk-Góral pediatric neurologist

14 October 2018

Thank you for your tremendous work, I totally agree with the proposal adjusting neonatal seizure classification. This classification not only makes easier in diagnosis, but also avoid misdiagnosis in neonate seizure.  

Dr Kien Minh Nguyen

14 October 2018

Thank you for your great work on simplifying a rather complex topic.

My concern is that classification of neonatal seizures based solely on EEG manifestations has several inherent limitations. 

To start with, electroclinical decoupling is a well-established phenomena; how do we account for these residual clinical seizure-like activity following start of AEDs?

In addition, a group of seizures that may be seen in neonates including myoclonic seizures and spasms does not fit the eeg criteria noted.

So a bunch of seizure types are not accounted for in this definition. 

I think classification into clinical and electrographic seizures is great but I think within clinical seizures, use of semiology may be more helpful that just relying on EEG manifestations seen.

Thank you. 

Best regards,
Abeer Hani, MD

14 October 2018

First of all, I thank the ILAE Task Force on Neonatal Seizure for their hard work on this challenging topic and for give us the opportunity to review this important document.

Secondly, after reading the classification criteria and explanations carefully, as a non-expert in neonatal seizures I think that the classification at this age range is very clear,simple and easy to practice.

Thirdly, I totally agree that “electrographic only seizures”, without clinical signs, have to be included in the classification system: therefore neonatal epileptic seizures may be 1) electro-clinical or 2) electrical only seizures.

Last,I suggest If there is only evidence of clinical seizure without electroencephalogram, it is necessary to be cautious in diagnosing epileptic seizures.

Thanks again to the Authors,
Cathy Zhang

14 October 2018

Congratulations to the team for their excellent work in this challenging domain- Neonatal Seizure Classification.

My only comment is that the new proposed term of “sequential seizures”, is confusing. As we know many seizures, in different age group, are multiple and migrating. According to Cambridge dictionary, sequential means, consecutive, successive. Accordingly, many neonatal seizures, I am afraid, are going to be classified into “sequential”.

Additionally, 2017 ILAE classification manual emphasizes the onset of seizures, in the example given which was as follows: “Sequential seizure manifestations: A seizure begins with tingling in the right arm of a 75-year-old man (Fisher RS, Cross JH, D'Souza C, et al. Instruction manual for the ILAE 2017 operational classification of seizure types)”. Although the onset of the seizures is important, the predominant manifestation is as important as the onset.

Imad Yassin Saadeldin, MD, MRCPCH, FAMAP

13 October 2018

ILAE task force on neonatal seizures has been a great initiative to integrate neonatal seizures into 2017 ILAE classification of seizures and epilepsies. The proposed neonatal classification framework emphasizes the role of video EEG in the diagnosis of neonate seizures and is likely to address poor inter-observer agreement in diagnosing and classifying these seizures. We agree that this classification is likely to have better inter-observer agreement. It would be interesting to know whether the inter-rater agreement was tested for the EEGs studied by the task force.

We would like to stress upon the following issues:

  1. The proposed classification will be difficult in resource limited settings where video EEG/ aEEG is unavailable. So, it would affect the wider applicability of this classification. This is important because the key purpose of the ILAE classification is to have a uniform and practical classification globally.
  2. Whether brief electrographic seizures (especially BRDS) need to be treated as seizures is still an enigma; thereby complicating classification if the duration is less than 10 seconds. Determining evolution if duration is very brief is difficult. Besides this, interictal epileptiform discharges are not generally an indication for escalating AEDs and whether BRDs are always ictal is also unclear.
  3. Among the neonatal epilepsy syndromes mentioned on page 11 of the document, Epilepsy with migrating focal seizures should also be included since it can also have a neonatal onset. This might have diagnostic (KCNT1) and therapeutic implications(quinidine).
  4. Among the modifiers of myoclonic seizures, Asynchrony may be another important modifier for multifocal myoclonus which has not been included (for identifying fragmentary/ erratic myoclonus, a clue to ohtahara syndrome).
  5. Among the modifiers of epileptic spasms, behavioral state (wakefulness/ deep sleep/ both) during which the spasms occur may be an important modifier to ease the classification of an underlying epilepsy syndrome (esp. differentiating ohtahara and west syndrome).
  6. Neonatal status epilepticus: Since this document highlights the role of EEG in neonatal seizures, neonatal status epilepticus will be worth mentioning. As per ACNS guidelines, it is defined to be present when the summed duration of seizures comprises ≥ 50% of an arbitrarily defined one-hour epoch. Whether frequent BRDs without clinical seizures constituting more than 50% be treated as status is another dilemma.

This great effort is likely to improve uniformity in diagnosis and classification of these seizures.

Sincere regards,
Dr Priyanka Madaan, Dr Jitendra Kumar Sahu

13 October 2018

Neonal seizure is firstly confronted by general practioner, junior pediatrician or junior neonatologist. So they need simple algorithm for differential diagnosis and management of neonatal seizures which can easily applicable.

Mohamed Romih

12 October 2018

Overall I think the classification is great and will be helpful for clinicians. I have the following comments:

  • I like the way the neonatal classification has incorporated electrographic seizures into the 2017 seizure classification
  • I am not convinced that getting rid of the focal/generalized aspect is the way to go. Although I acknowledge the rationale for this, I am concerned that instead of enforcing the fact that all neonatal seizures are focal, leaving that out of the classification will actually weaken this message in the wider community. Clinicians with less neonatal epilepsy knowledge who are aware of the ILAE 2017 classification may not appreciate that point (ie that all neonatal seizures are all focal) if it is actually completely left out. I would prefer that the focal/ generalized be put back in to the figures 2 and 3 and that where the generalized is - the box says not found in neonates or something of that nature– or alternatively just say focal and don’t have generalized so that the point is made and the neonatal classification is differentiated from the 2017 ILAE seizure classification of seizures in older children and adults.
  • I also think the focal epilepsy should go back in – not all neonatal epilepsy will fit into a defined syndrome and so there needs to be something to call those other neonatal epilepsies – focal epilepsy would be the most appropriate.
  • The definitions sections – parts of this seem a bit unnecessary as you don’t seem to actually use some of the terms in the document apart from in Table 3 - perhaps they could just go in the table legend? GA, CA, PMA
  • I assume the committee was blinded to aetiology when they reviewed the neonatal seizure? – it would be good to say that somewhere (I couldn’t see it) as it would add to the robustness of the observation of seizure type associated with aetiology.
  • It would be helpful to have an EEG figure of the BIRD to go along with the proposal for electrographic seizures in the discussion – specifically showing BIRDs that fulfil the proposed definition of electrographic seizures and those that do not.
  • Table 1:
    • Tonic seizures under the specific considerations for neonates – I think that using the term generalized here ( “Generalized Tonic Posturing”) is not helpful and potentially confusing given the use of the word generalized in the 2017 classification– I would prefer it said “bilateral symmetric tonic posturing”
    • Behavioral arrest under the specific considerations for neonates – I don’t understand what you mean by “may be focal”? Using the term focal here is confusing.
  • Figure 2 – I really like this but I personally think you need to put the word focal back in to emphasize the point or it will be lost – people will be less inclined to read the text , they will ultimately just look at the figures - perhaps just add the word Focal before Seizures in that seizure box (see below Figure).
  • Figure 3 – again I think you need to add back in focal seizures and focal epilepsy as mentioned above I would suggest something like the following:
    Neonatal Seizure and Epilepsy Classification

Thank you for giving me the opportunity to comment on the proposal. Overall I think it is very good.

Best wishes,
Lynette Sadleir

12 October 2018

We really appreciate the effort of Task Force.

We only have some concerns about the electroencephalographic correlate: At all ages, true clinical epileptic seizures without electroencephalographic correlate can occur. We have occasionally observed focal epileptic seizures in neonates without an electroencephalographic epileptic activity despite a technically adequate EEG record.

A classification system should allow to allocate all the possible results even there is a low possibility of occurrence of these events. In that sense, our proposal is to include clinical epileptic seizures without electroencephalographic correlate in this classification.The criteria to consider them in this group could be that a experienced child neurologist can to corroborate or to establish such a diagnosis. The video-register of the event may be the principal tool to this diagnosis.

Best Regards,
Dr. Ivan Espinoza Quinteros

12 October 2018

I agree to the rest comments, that there is a tremendous effort done by the task force for this issue. If one comes to practise neurology in resource limited country the first questions crossing the mind is that how to recognize the subtle forms and bizarre features of NS; Is it epileptic or symptomatic one; Is it easy to manage or intractable as we see often. If the current classification task force address this and respond to it this will be of help to this group of epileptogist from these countries.

تم الإرسال من البريد لـ

12 October 2018

As a non-expert in neonatal seizures I have the impression that his is a very clear and rational way to classify seizures at this age range.

I also see that there are good pragmatic reasons to recommend to be cautious when reacting to clinical paroxysmal events which do not have ictal EEG correlates, and to accept paroxysmal EEG-events as seizures even when appearing "subclinical" at first glance.

Focal seizures with origin in basal or mesial brain regions can, however, occur without scalp EEG correlates at any age, and if remaining localized or obscured by artifacts can be impossible to capture using scalp EEG even under optimal recording conditions.

I would suggest to reconsider if a classification system should offer an option to classify events without ictal scalp EEG correlated as seizures provided that there is a high plausibility for this, e.g. based on concordance of structural and functional imaging with the observed paroxysmal behavioral events.

With best regards,
Andreas Schulze-Bonhage

12 October 2018

Dear colleagues:

I appreciate the presentation of this classification.  I think it's simple and pragmatic.

The work they present in the observation of the videos is very important for de etiologic orientation.
In the reality of most neonatal units the possibility of integrating the EEG video as a tool of daily use is still difficult, timidly the use of the aEEG begins in some units.

This classification gives greater prominence to the joint interpretation of clinical and electrographic events. I think that in some countries with less favorable conditions, the integral approach will be later.

It's a good start. Congratulations

Dra. Andrea Rey

11 October 2018

I think the team had done a tremendous job so far. I thank them for making video EEG as the gold standard for the diagnosis of neonatal seizure. Particulaly in a developing country like us (Banglades), video EEG is a big challenge for us! You may find it surprising that we have only a few video EEG in our country. Your recommendations may increase the availability of video EEG in our country. I think 'sequential seizure' needs more discussed and clarified!

Best regards,
Dr. Yamin Shahriar Chowdhury

11 October 2018

Congratulations to the team for a this great job.

However this can not applied to my own setting where both the machine and EEG expertise are lacking.

I suggest if can further develop the proposal to include "clinical neonatal seizure'' only without EEG recordings. This is for the benefit of those dealing with neonates and neonatal seizures in the developing world.

Thank you,
Halima Adamu

11 October 2018

Thank you very much for the excellent work. The "Classification of Seizures & the Epilepsies in the Neonate" proposal answers many of the questions that are presented in newborns with neonatal seizures, which are often a clinical challenge.

From the clinical point of view, the classification seems simple and very applicable in different contexts, in addition to emphasizing the role of the EEG in the diagnosis. There is still a gap in seizures that are semiologically "tipical seizures" and that do not accompany an EEG correlate. Should they be defined as neonatal seizures? The presentation of repeated "focal clonic" or "generalized tonic" seizures, with an evident lesional substratum, but that have no EEG correlate ... can be defined as a seizure according to this proposal? On the other hand, brief electrographic events of less than 10 seconds, evolutive or rhythmic but without clinical correlate, should be considered neonatal seizures? or would it be interictal activity? These questions are presented in the usual clinical practice.

I really appreciate the effort, and I think this proposal simplifies the description of the observed events. If EEG is available, it can help in a better classification, but there is still work to be able to define events without EEG correlate, especially when the diagnosis defins therapeutic behaviors, considering environments where EEG availability in neonatal units is scarce.

Dr Viviana Venegas

11 October 2018

Dear team,

'First of all I congratulate the neonatal classification team for undertaking this task which would provide a common language in classifying neonatal seizures across the globe.

Some of the seizure-related clinical events in neonates are morphologically similar to non-seizure related events, such as apnoea. I agree with author neonatal seizures should be underpinned with electrographic evidence which is absolutely essential.'

Gopalakrishnan Venkatachalam

10 October 2018

I thank the ILAE Task Force on Neonatal Seizure for their hard work on this challenging topic and for give us the opportunity to review this important document. It integrates the 2017 ILAE Classification and accepts and reinforces the concept (ILAE report 2010) that neonatal convulsions should no longer be considered as a separate entity.

The following comments might be useful for the discussion:

- First of all I totally agree with the statement that clinical events without an EEG correlate have not to be considered as epileptic seizures, therefore neonatal epileptic seizures may be 1) electroclinical or 2) electrical only seizures. According to Mizrahi and Kellaway (1987) this proposed classification states the importance of the EEG in the diagnosis of seizures in neonates.

- This classification transposes the concept of “seizures type” as defined in the 2001 ILAE Commission report on classification and terminology: “Axis 2: is the epileptic seizure type, or types, experienced by the patient…… that represent diagnostic entities with etiologic, therapeutic, and/or prognostic implications. Localization within the brain should be specified when this is appropriate”. However on discussing about neonates a “List of Accepted Epileptic Seizure Types” has to be still defined and this issue is the effort of the present work, highly appreciable. Such an approach also in neonates is applicable and useful for the management.

- The fact that arbitrarily 10 seconds have been chosen as the minimum duration to define a seizure must be reviewed, taking into account that not only spasms, but also focal seizures, may have a shorter duration, although they are clearly epileptic in nature.

- The novel seizure type “sequential” has to be discussed. In agreement with the works of Luders et al (2005) and Nagarajan et al (2011), rather than a single predominant sign it is important to define the succession of single individual critical signs (i.e. from Loddenkemper et al 2005: “2. Semiology: abdominal aura -> automotor seizure ->right versive seizure -> generalized (now bilateral) tonic-clonic seizure”). The sequence of semiological signs should always be described. I agree with Nagarajan that the clinical feature presenting early in the seizure may reflect the seizure onset region in the brain.

- I totally agree that “electrographic only seizures”, without clinical signs, have to be included in the classification system: both in the results of the Commission and in the clinical practice (especially using long term EEG monitoring) the most common seizure type is electrographic only seizure.

- Moreover, also oral automatism, eyes deviation, apnea and autonomic signs might be frequently recorded as focal seizures, regardless of the degree of CNS maturation.

- It has been criticized that this classification proposal cannot be used where the EEG registration is not available and for this reason it cannot be applicable in all health care settings. This claim could seems obvious, but we have to keep in mind what we have learned on ictal semiology by means of the video EEG recordings, that allow us to recognize critical events with a good accuracy, even without EEG registration. Obviously electrical only seizures can not be recognized, but this seizure type do exist and has to be included in the classification system.

- Last but not least, as soon as this Proposal Classification will be accepted by the ILAE community, it will be mandatory to share it with the community of neonatologists, who, to date, generally keep using the classification proposed by Volpe (1973, 1977, 2001), and do not consider the EEG an indispensable tool in the diagnosis and management of NS.

Thanks again to the Authors,
Massimo Mastrangelo

10 October 2018

Thank you for your proposal. I agree and consider it very useful.

Lidia Caceres

9 October 2018

Neonatal seizures should classify as


  1. Acute Symptomatic
  2. Neonatal Epilepsy

B :

  • On EEG
  • Clinical Seizure without EEG documentation or normal EEG
  • Clinical seizures with Electrographic documentation



Acute Symptomatic

  • acute metabolic
  • genetic metabolic



  • Ischaemic or hemorrhagic stroke
  • vascular malformation


  • Central nervous system
  • Systemic infection


Benign seizures

Dr Surendra Khosya

9 October 2018

Thank you for your work on this important topic. This report forms an excellent framework for future work in this area.

When reading the methodology section, a description of a process for adjudicating disagreements on EEG readings or determining semiology is not described. Was a process for adjudication determined prior to the study? If so, a description of the process and how often it was used will be helpful to readers.

In the tables several acronyms are used without definition. Inclusion of the definition of the acronym with the table, even if it is provided in the text, will be useful for readers who may not be as familiar with these terms.

In Table 4, not all of the most common seizure types are highlighted. For example, 2 of 4 preterm neonates in an infectious diagnosis had non-motor electrographic seizures. In Table 2 it would be helpful to give the range of GA and PMA at EEG.

Timothy E Welty PharmD FCCP BCPS

9 October 2018

The classification of "seizures" and "epilepsies" has always been a challenging task in neonates and young infants, primary reasons being the immaturity of nervous system and the heterogenous etiological profile. In this background, the ILAE Task Force's effort to develop guidelines for classification is definitely welcome and greatly appreciated. This should be welcomed, as this will contribute to uniformity in reporting, choosing homogenous populations for therapeutic trials, and help streamline management decision maing across multiple levels of health care. But, I have the following comments to make.

1. The results from study (of 147 seizure episodes carried out for this purpose by authors) shows "electrographic only" as the commonest seizure type both in term and pre-term neonates. This was noted across all etiologies except genetic. This fact highlights the underrecognised burden of "electrographic only" seizures in critically ill neonates. This also means, it will leave out many neonates untreated in resource-constrained settings, and large number of facilities without access to continuous EEG. When the stated objectives of these guidelines is developing a classification system with "implications for management and treatment", and "applicable to all healthcare settings", the emphasis on EEG, though ideal, is not pragmatic. The authors/guidelines should balance this by adding modifiers for resource-constrained settings, with the sole purpose of optimising management and treatment of these neonates. This can be part of main document, and not a passing mention/reference.

2. The shift from "initial" manifestation to "predominant" manifestation in the proposed classification, is frought with certain challenges/dangers. For example, the Neonatologist may view apnea as "dominant" manifestation, while Neurologist may see 'focal tonic' as "dominant" manifestation. With the current evidence (basic sciences and clinical studies), there is limited data to suggest that "predominant" manifestation has correlation with either etiology or outcome, as opposed to "initial" manifestation. The guidelines should address this ambiguity to minimize reporting bias, and subjectivity in classification.

3. In most health care settings across the world, treatment of seizures precedes the detailed evaluation (includes EEG monitoring). In such a scenario, a good proportion of neonates would have already received anti-epileptic drug(s). How does the guidelines address this?

4. Doing away with seizure onset (focal, generalized) from the proposed classification, is debatable. Although the onset of seizures is often difficult to ascertain in neonates, attempting to identify seizure onset, is definitely useful in clinical practice. This is so because, etiology is the most definitive predictive factor for outcome, and seizure onset can help direct the investigations/ etiological work-up. The authors may consider in retaining the seizure onset in classification, but can be kept as optional/ wherever applicable/feasible.

All the best, and hoping for widely useful and acceptable classification system.

Dr. Ramesh Konanki

8 October 2018

I attach a diagram for a proposal of a classifcation of Neonatal seizures:
Neonatal Seizures diagram from Ibrahim Shoukry

With my best regards,
Ibrahim Shoukry

8 October 2018

Thank you for the opportunity to comment for ILAE Task Force on Neonatal seizure.

First, I express my respect for the great effort made in the proposed classification of seizures and the Epilepsies ; modification for seizures in the neonate, especially because the aim of the report is to use terminology consistent with the 2017 ILAE Classification of Seizures and Epilepsies which we know that neonatal seizure differ from those of older children and adult.

I agree that in Neonatal period, the initial division into focal and generalized is unnecessary because as newborns have been shown to have seizures with exclusively focal onset. I agree that in Neonate, the second division into aware and unaware seizure is not applicable because it is not possible to confidently assess awareness and responsiveness in this age group. The following division into motor and non- motor seizures and finally by the seizure type is difficult due to lack of verbal and limited non- verbal communication. Seizures in neonates can present with a variety of clinical signs and in some situations, it may be difficult to identify the dominant feature. The seizure classification according to the predominant clinical sign manifestation, as this is more likely to have clinical implication for etiology than seizure onset zone.

For the reasons above, I agree with the proposed diagnostic framework of seizures in the neonatal period (figure 2) which emphasizes the necessity of EEG diagnosis of seizures in the neonatal period. Ictal EEG finding  is gold standard for seizure diagnosis in the neonate. Using EEG video analysis, some clinical events previously reported as seizures were in fact non epileptic. Electrographic seizure without clinical correlates are frequent , particularly in critically ill neonates.

I agree about the definition of electrographic event with a pattern characterized by sudden, repetitive, evolving stereotyped waveforms with a beginning and end. I didn't have idea about the time spent of electrographic  event but I think the duration of epileptic discharges has to be sufficient to demonstrate the evolution in frequency, amplitude and distribution (if focal secondary generalized) of physiologic wave. These evolution show that the epileptic discharges are ictal. The duration has to be sufficient to determine the ictal epileptic discharges are focal / focal secondary generalized or generalized.

In conclusion, we consider  that the proposed diagnostic framework of seizures in the neonatal period including classification of seizures which emphasizes the necessity of EEG diagnosis (electrographic only and with clinical sign correlation) is agreeable.

Thanks to the authors,
Anak Agung Ayu Meidiary, MD, Neurologist

8 October 2018

This is tremendous work, very helpful, thank you!

Monika Olischar, MD

7 October 2018

I read the document with interest. The text is clear and informative. I have just one comment regarding the classification of seizures by the panel. The way agreement was obtained must be clarified further. Was inter-rater agreement tested?

If not, did the members of the panel reach a formal consensus?

Ettore Beghi

3 October 2018

Neonatal seizure classification remains a very difficult topic.  It is clear that the Task Force managed to improve the classification a lot and made it much more easy to use!

I agree fully with the electrographic only term.  The idea of BIRDS remains controversial, as some short patterns can show evolution, really resemble seizures, but are shorter than 10 seconds.

The 10 seconds is useful, since we have learned that seizure burden remains important in treatment decisions, but I am not sure that these BIRDS are very different from longer seizures in origin.

Still not completely sure whether generalized seizure should not be considered at all, especially in severe encephalopathy patients eg EIEE.  And should we not consider some spasms or myoclonic seizures as generalized?  Although it is clear that most of the seizures remain focal.

Sequential and predominant are more difficult terms to use in practice.   Is there still a place for migrating seizures, or should we replace these by sequential, even with same type of semeiology?

From the etiologic point of view, there are some nice new insights

I agree that clonic seizures are very often seen in localized pathology eg stroke or hemorrhage , whereas patients with encephalopathy (HIE, EIEE, EME) mostly present with tonic, myoclonic or EEG only seizures.  Still much more to discover here from an pathophysiological point of view

Nice work,
Katrien Jansen, MD, Ph

3 October 2018


I congratulate the team for this great and timely work. I am sharing some comments from my neonatology colleagues.

  1. Page 4/line 44: Although the starting date for GA is frequently defined as the first day of the last menstrual period, for many others it is defined from early fetal ultrasound or (most accurately) 2 weeks before the date of artificial insemination. Please considering adding these other methods when defining GA.
  2. Page 4/line 47 typo: Change “deliver” to “delivery”
  3. Page 4/line 57: By this confusing definition of “neonatal period”, a preterm baby born at or before 36 weeks would be outside the neonatal age beginning at 40 weeks PMA, whereas a term baby born at 41 weeks would be in the neonatal age range until 45 weeks PMA. To be consistent, please consider using 44 weeks’ PMA to define the end of the “neonatal period” for babies born at any gestational age.
  4. The new neonatal seizure classification proposal looks fine to me. I’m glad to hear that describing the body part first affected by a seizure is not important in the neonate.

Edward J. Novotny, Jr., M.D

2 October 2018

I think this material will be very important to support all the neuropediatrics who work with  the neonatal population. In my country, Brazil, not all the hospitals where I workhave EEG or a Poligrafy and it is very dangerous, in my opnion, to continuos work like that. With this material finally we can discuss with the manager of these hospitals about the safety to identify seizures and treat the seizures correctly.

Thank you for all the members of these group, mainly Prof. Ronit that I have the opportunity to meet at the Summer school in Dianalund and Prof Magda, a Brazilian professor part of this time!!

Best regards
Daniela Bezerra

30 September 2018

Well written paper and agree with their recommendations.

Batool F Kirmani, MD, FAAN, FAES

23 September 2018

Congratulations to the team for their hard work and excellence in this challenging domain- Neonatal Seizure Classification. The neonatal seizure classification proposed emphasizes the role of EEG in the diagnosis of seizures in the neonate, this is essential, as electrographic only seizures as well as electroclinical seizures are important to recognise, identify and quantify in the neonate.

In the ILAE operational classification of seizures (2017), the seizure type was defined by the initial manifestation, not the most obvious. It is important to do so with neonatal seizures as well, whereas in this proposed classification the seizure type is determined by the predominant clinical feature. How does one decide what is the dominant feature – a significant desaturation with apnoea or a motor phenomenon (positive or negative) or the feature that persists for the longest duration? The proposed category of sequential seizures (as most seizures have multiple features, a lot of seizures may be lumped/dumped in this category) is unlikely to help with better understanding of the underpinning networks, aetiological diagnosis, prognostication or therapeutic interventions.   Co-relationship between seizure type and aetiology is far from absolute – most electroclinical seizures, irrespective of aetiology, may have multiple clinical features as shown by us (2012) and others.

What is the specific purpose of this classification of clinical epileptic seizures? Our understanding of seizure networks in the developing brain is developing and evolving. Seizures in the neonate could originate in neocortical, thalamocortical, limbic, and? even in cerebellar and brainstem networks. The clinical feature presenting early in the seizure may reflect the seizure onset region of the brain. By not looking at the initial manifestation of the electroclinical seizure, the evolution of the semiology, the electroclinical patterns at onset and propagation, the opportunity to learn more about the neonatal brain circuits involved in seizure generation may be lost.

The proposed classification is excellent in outlining the importance of the EEG in the diagnosis of seizures in the neonate. My suggestion is that the issue of classification of the clinical manifestations of neonatal seizures, be discussed further and perhaps modified. Thank you, once again, to this ILAE task force on Neonatal Seizures, for this commendable and important work.

Lakshmi Nagarajan, MBBS, MD, FRACP  

19 September 2018

I congratulate the team for the hard work put in and trying to classify neonatal seizures, which have otherwise protean manifestations.

I have the following comments to offer regarding the proposed neonatal seizure classification..

1. I partially agree with the statement that most neonatal seizures are of focal origin, but in the case of pyridoxine dependent, hyperglycinemia etc. the predominant seizure type will be generalized ( myoclonic) as well as in Ohtahara syndrome due to structural malformations, you may get a clinical generalised tonic spasm..hence completely removing generalised category at the initial level may be wrong. We can have : Focal (motor/non motor), generalised and electrographic as three major groups to begin with.

2. Is there any reference for the statement: "Clinical events without abnormalities on EEG are not actual seizures" because in the same neonate, we do observe similar movements with and without electrical correlates. So should there be an entity of ' clinical seizures/ suspected seizures even if no electrographic abnormalities are seen during the seizures. Of course in such babies, we must have an etiology (e.g. HIE), abnormal neurological state during the event and at least some background abnormalities on the EEG even without the events.

3. In the etiology, it will be better to have vascular and structural as separate entities, as the prognosis, connotations, and need for prolonged AEd may be very different in a dysplasia versus hemorrhage/stroke.

4. As rightly pointed out by the authors, this scheme will be difficult to use in developing countries without access to continuous EEG. The scheme used in reference no. 92 can be attached as an appendix for ease of use by such centres.

5. When we talk of sequential seizure, it will be better if the task force specifies whether major importance should be given to the beginning of such a seizure or to the predominant manifestation. Otherwise everyone will classify it differently leading to confusion.

With regards,
Dr. Kavita Srivastava, MBBS, MD

16 September 2018

In last year's the general classification of seizure in older children and adult was widely accepted,followed by most of the clinicians and epileptologists  in most of the countries. The major change which were  made -- the  introduction of word awareness in-spite of the word consciousness either aware or unaware of the event and focal onset vs generalized onset or unknown & non motor vs motor manifestation, all these  components may not be applicable in neonates, large majority of neonatal seizures are not caused by epilepsy as a chronic disorder.

The idea is understandable that classifying neonatal seizures by the “predominant” clinical manifestation is more likely to have clinical implications for etiology. The novel seizure type “sequential” is consistent with the features of neonatal seizures that seizure foci can migrate from one site to another during a single seizure and neonatal seizure are exclusively focal onset a bold conclusion indeed  therefore the terminology generalised in not necessary  here  fully agreed upon The proposed neonatal classification framework emphasizes the role of EEG in the diagnosis of seizures in the neonate and includes a classification of seizure types relevant to this age group. The seizure type is typically determined by the predominant clinical feature. As seizures in neonates are often electrographic only with no clinical feature, therefore three separate groups of classification like clinical, electroclinical and electrographic groups need to be included.

The concept of “sequential seizure” highlights the difference in seizure manifestation between neonates, and older children and adults. As mentioned in the evaluation of the proposal, these ideas will be the application of new classification easier even to non-specialists and nursing staffs working in ICU -- an extremly essential part in critical care management. It is expected that the diagnosis and treatment of neonatal seizures will be advanced by the wide application of this new classification of neonatal seizures.

A sample of proposed classification

Etiological classification

1. Genetic/Congenital

  • epileptic syndromes genetically determined
  • structural /congenital malformations
  • inborn errors of metabolism

2. Acquired:

  • hypoxic ischemic encephalopathy(HIE) Commonest
  • vascular event like intraventricular hemorrhage
  • sepsis and CNS infection
  • metabolic encephalopathy(trasient errors) and enzymopathies

Auto Immune and Unknown etiologies are concerned as they are not operating in neonatology.

Clinical ,Electroclinical and Electrographic classification needs more elaboration:

  1. West`s syndrome is classical example of electroclinical diagnosis which perhaps merges with LGS in later age group
  2. Transient metabolic errors ,sepsis or in HIE(early stage) where only clinical manifestation may be recorded in video with nonspecific EEG changes
  3. Clinically silent or sub-clinical seizures where only EEG changes are seen

Epilepsy syndromes sould be under a separate heading although it may genetically determined Syndromes presenting in the neonatal period include: self-limited (benign) familial neonatal epilepsy, early myoclonic encephalopathy (EME), and early infantile epileptic encephalopathy (Ohtahara syndrome)

Frame work is excellent in the light of 2017 classification:

Neonatal Seizure proposed classification GP Burman
Thanks a lot to the ILAE
GP Burman

14 September 2018

I thank the elements of this Task Force for the efforts to create a useful and widely accepted classification of neonatal seizures. It's certainly not an easy task.

Two comments came to mind:

  • The 1998 semiological seizure classification proposed the term "hypomotor" for seizures with predominant motor arrest, when consciousness cannot be assessed (as is obviously the case here). Why not retain that term, instead of "behavioral arrest?"
  • The sequence of semiological signs is relevant and should always be described. The proposed category of "sequential seizures" reduces the internal logic of the classification and risks being used as a "whatever" category, where all longer and more complex seizures are included.

Best regards,

Ricardo Rego

13 September 2018

I think this is an important paper. The authors have done a terrific work which provides many clarifications on the topic.

I'll be glad if upon publication, the authors would consider to write a short summary of their work to publish in Wikipedia in the frame of the Wikipedia Epilepsy Initiative. I'll be happy to assist.

Nicola Maggio, MD, PhD

11 September 2018

I agree with the proposal of this important document.  We think that it is the product of a better understanding of the epileptogenesis, a more detailed seizure semiology and a better electroencephalographic correlation, that applies to the majority of cases.  It is an excellent effort to guide the clinical approach and management of epileptic disorders.

David Dondis, MD, FACP

10 September 2018

I believe that this new classification on neonatal seizures is a great effort to unify the diagnostic criteria, and thus to better understand the pathophysiology and the evolutionary course. I think it will be very useful for timely intervention and avoid overmedication.

Rosa Alvarado Merino, pediatric neurologist

9 September 2018

A usefull classification would be:

1 genetic :

  • epileptic syndromes
  • structural congenital malformations
  • inborn errors of metabolism

2 acquired:

  • hypoxic ischemic encephalopathy
  • stroke
  • sepsis and CNS infection
  • tansient metabolic errors

The above classification is a practical cot side or NICU translation approach to ILAE classification into:

  • Genetic
  • Metabolic
  • Structural
  • Infection

Omitting Auto Immune and Unknown etiologies as they are not operating in neonatology. Highlighting 2 types ol metabolic dérangements namely enzymopathies and transient errors.

Ibrahim Shoukry

4 September 2018

First of all, we express our great respect for accomplishing the hard work, proposing a new classification of neonatal seizures. All of us have read the new classification and discussed about it using e-mail.

In general, we consider that the new classification of neonatal seizures successfully incorporates their features into the 2017 ILAE classification of seizures in older children and adults. We strongly agree with the brave statement that neonatal seizures are exclusively of focal onset. The idea that neonatal seizures will be best to be classified according to the “predominant” clinical manifestation seems quite practical, as mentioned by the authors. This unique idea will contribute to better applicability of the new classification.

In epilepsies in older children and adults, determination of site of seizure onset is important. In contrast, a large majority of neonatal seizures are not caused by epilepsies as a chronic disorder. The idea is understandable that classifying neonatal seizures by the “predominant” clinical manifestation is more likely to have clinical implications for etiology. The novel seizure type “sequential” is consistent with the features of neonatal seizures that seizure foci can migrate from one site to another during a single seizure. The concept of “sequential seizure” highlights the difference in seizure manifestation between neonates, and older children and adults. As mentioned in the evaluation of the proposal, these ideas will the application of new classification easier even to non-specialists and nursing staffs. It can be expected that the diagnosis and treatment of neonatal seizures will be advanced by the wide application of this new classification of neonatal seizures.

However, we are afraid that there are some issues to be solved in this new classification. At first, there may be substantial inter-observer variability determining the predominant manifestation within a single seizure. The predominant manifestation can be different according to the observers, because what “predominant” means is not sufficiently clear. Some may think clinical manifestation which is observed for the longest period within a seizure, whereas other may think clinical manifestation which attracted most strong attention of the observer. It is desirable to show clearly how to determine “predominant” seizure manifestations using several representative scenarios.

Second, a large majority of neonatal seizures will have sequential evolution of seizure manifestations during a single seizure. This can be often in infants with asphyxia, the most often cause of neonatal seizures. Therefore, we are afraid that a large majority of neonatal seizures may be classified into “sequential”. In such case, the significance of seizure classification will be reduced. Some method should be provided to describe sequential changes of clinical manifestations by preparing descriptors. We consider that these two issues should be examined in the future.

Future agenda regarding neonatal seizure includes the definition of ictal EEG findings, the gold standard of the diagnosis. “The 10-seconds rule” lacks biological evidence, but its appropriateness has been recognized until now. We also thing that “the 10-seconds rule” is acceptable. On the other hand, seizures with a shorter duration such as spasms do exist. Ictal EEG findings of neonatal seizures with a shorter duration should be established. In addition, a significance of BIRDs (brief interictal rhythmic discharges) should be investigated in the future, as the authors stated.

In conclusion, we consider that the proposed new classification of neonatal seizures proposed is agreeable, but there are some issues to be solved in the future. We hope that vigorous researches will be performed toward future revision.

Akihisa Okumura MD, PhD

3 September 2018

This paper is a terrific addition and provides excellent clarification on many issues regarding neonatal seizures.

For the authors to consider: there is a certain amount of redundancy between the introduction, review, results and discussion which, if pared down, would make the document a bit shorter.

Found one typo page 16 line 55 should be Bedside maneuvers not Beside maneuvers.

Thanks to the authors.

David. E. Mandelbaum, MD, PhD

12 August 2018

I have read the neonatal seizure classification. It is a clear and most welcome addition to the ILAE classifications.

I think it is really helpful to consider and give prominence to a concerning entity that is titled within the paper as 'electrographic only seizure'. This does create some confusion by referring to this entity as a seizure given that the current ILAE definition of seizure referred to implies a clinical manifestation. This to me implies that either the overarching ILAE definition of seizure needs to be updated to allow for epileptic events without clinical manifestation or the concerning electrographic only entity in neonates should be called something else.

Otherwise why shouldn't we refer to epileptiform changes without clinical correlate in the older child as electrographic only seizures.

I think figure 1 is a bit confusing particularly when trying to correlate with figure 3. I think it would be useful to have 'unknown' and 'structural' also shown in figure 1. In figure 3 structural seems to be defined as encompassing infarction and haemorrhage where in fig 1 structural seems be missing. Fig 1 the size of the ovals seem to be not proportionate to the likelihood. I think this is important as HIE may be in practice to readily considered before other other important causes of neonatal encephalopathy. Also other overlaps can occur other than those shown. I wonder whether figure 1 is best omitted as i cant imagine how it might look to correctly reflect the intention!!?

Colin Dunkley

11 August 2018

The Neonatal Seizure Classification has done an excellent job of conforming (with appropriate variations) with the overall 2017 classification of seizure types by the ILAE. This makes sense, since there was coordination and overlap of both Task Forces. Perhaps the biggest differences are minimization of subjective symptoms (not reportable in neonates), lack of generalized seizures in neonates and increased emphasis on the EEG pattern. The EEG pattern is important for classification of some non-neonatal seizures (e.g., focal impaired awareness versus absence), but the 2017 seizure classification is mostly based on signs and symptoms.

The motor classifications of neonatal seizures include one novel word, which is “sequential.” This one word opens up the whole slippery question about whether a propagation pattern is one unified seizure or multiple “sequential” seizures. This is not an easy issue in any age group, but the 2017 classification defined the seizure type by how it started, and not by its propagation – which is infinitely variable.

The 2017 classification classified a seizure on the basis of its first manifestation, even if not the most prominent. Exceptions were impaired awareness at any time during a seizure and the special case (not applying to neonates) of focal to bilateral tonic-clonic seizures. The neonatal classification reverses this approach and classifies by the most prominent feature. All neonatal seizures are admittedly focal, but it seems that which type of focal should be important. I am not a neonatologist, so I cannot speak to whether onset or prominence are more important for localization, etiology, prognosis and treatment. However, in adults onset is most linked to these. If a different system is used for neonatal seizures, at what age are we supposed to switch how we classify propagating seizures?

This classification system should be a major step forward for establishing clarity and consistency of terminology in this area.

Robert S. Fisher, MD, PhD

9 August 2018

I have to congratulate all doctors involved in this great effort made in the proposed Classification of Seizures & the Epilepsies in the Neonate, since this subjetc has been forgotten long time ago and now this documente will become a great tool for clinical practice and new researches in this field. 


N. Barajas, fellow pediatric neurology