Electroencephalography in neonatal epilepsies. Issue 9 (30th September 2020)

• Record Type: Journal Article
• Authors: Okumura, Akihisa
• Is Part Of: Pediatrics international. Volume 62:Issue 9(2020)
• Page Start: 1019
• Page End: 1028
• Publication Date: 2020-09-30
• Subjects: amplitude‐integrated EEG -- conventional EEG -- ictal changes -- suppression‐burst; Pediatrics -- Periodicals; 618.92
• DOI: 10.1111/ped.14227
• Languages: English
• ISSNs: 1328-8067

Abstract

Abstract: Neonatal epilepsies – neonatal seizures caused by remote symptomatic etiologies – are infrequent compared with those caused by acute symptomatic etiologies. The etiologies of neonatal epilepsies are classified into structural, genetic, and metabolic. Electroencephalography (EEG) and amplitude‐integrated EEG (aEEG) are essential for the diagnosis and monitoring of neonatal epilepsies. Electroencephalography / aEEG findings may differ substantially among infants, even within infants with variants in a single gene. Unusual EEG/aEEG findings, such as downward seizure patterns on aEEG, can be found. Neonatal seizures are exclusively of focal onset. An International League Against Epilepsy task force proposed that the seizure type is typically determined by the predominant clinical feature and is classified into motor or non‐motor presentations. Ictal EEG usually demonstrates a sudden, repetitive, evolving, and stereotyped activities with a minimum duration of 10 s. In epileptic spasms and myoclonic seizures, the cut‐off point of 10 s cannot be applied. One must always be aware of electro–clinical dissociation in neonates suspected to have seizures. Amplitude‐integrated EEG is also useful for the diagnosis and monitoring of neonatal epilepsies but aEEG cannot be recommended as the mainstay because of its relatively low sensitivity and specificity. At present, EEG findings are not pathognomonic, although some characteristic ictal or interictal EEG findings have been reported in several neonatal epilepsies. Deep learning will be expected to be introduced into EEG interpretation in near future. Objective EEG classification derived from deep learning may help to clarify EEG characteristics in some specific cases of neonatal epilepsy.