Neonatal Seizure Management: Is the Timing of Treatment Critical?

OBJECTIVE: To assess the impact of the time to treatment of the first electrographic seizure on subsequent seizure burden and describe overall seizure management in a large neonatal cohort. STUDY DESIGN: Newborns (36-44 weeks of gestation) requiring electroencephalographic (EEG) monitoring recruited to 2 multicenter European studies were included. Infants who received antiseizure medication exclusively after electrographic seizure onset were grouped based on the time to treatment of the first seizure: antiseizure medication within 1 hour, between 1 and 2 hours, and after 2 hours. Outcomes measured were seizure burden, maximum seizure burden, status epilepticus, number of seizures, and antiseizure medication dose over the first 24 hours after seizure onset. RESULTS: Out of 472 newborns recruited, 154 (32.6%) had confirmed electrographic seizures. Sixty-nine infants received antiseizure medication exclusively after the onset of electrographic seizure, including 21 infants within 1 hour of seizure onset, 15 between 1 and 2 hours after seizure onset, and 33 at >2 hours after seizure onset. Significantly lower seizure burden and fewer seizures were noted in the infants treated with antiseizure medication within 1 hour of seizure onset (P = .029 and .035, respectively). Overall, 258 of 472 infants (54.7%) received antiseizure medication during the study period, of whom 40 without electrographic seizures received treatment exclusively during EEG monitoring and 11 with electrographic seizures received no treatment. CONCLUSIONS: Treatment of neonatal seizures may be time-critical, but more research is needed to confirm this. Improvements in neonatal seizure diagnosis and treatment are also needed.

No effect of a musical intervention on stress response to venepuncture in a neonatal population.

AIM: To investigate the effect of a musical intervention on neonatal stress response to venepuncture as measured by salivary cortisol levels and pain profile scores. METHODS: In a randomised control crossover trial, participants were randomised to both a control arm (sucrose) and intervention arm (sucrose and music) for routine venepuncture procedures. Salivary swabs were collected at baseline, 20 minutes post-venepuncture and 4 hours post-venepuncture. Pain levels were assessed using the Premature Infant Pain Profile (PIPP). A total of 16 preterm neonates participated in both arms to complete the study. RESULTS: Cortisol values were elevated at all timepoints in the intervention arm (baseline, 20 minutes, and 4 hours post-procedure) but not significantly so (P = .056, P = .3, and P = .575, respectively). Median change in cortisol values from baseline was +128.48 pg/mL (-47.66 to 517.02) at 20 minutes and +393.52 pg/mL (47.88-1221.34) at 4 hours post-procedure in the control arm compared to -69.564 pg/mL (-860.96 to 397.289) and +100.48 pg/mL (-560.46 to 842.99) at 20 minutes and 4 hours post-procedure in the intervention arm. There was no statistically significant difference observed between groups (P = .311 at 20 minutes, and P = .203 at 4 hours post-procedure). PIPP scores were not significantly different between study arms. CONCLUSION: Our findings did not support the additional benefit of music intervention on neonatal stress response to venepuncture in preterm infants.

Idiopathic encephalopathy related to status epilepticus during slow sleep (ESES) as a "pure" model of epileptic encephalopathy. An electroclinical, genetic, and follow-up study.

OBJECTIVE: The objective of the study was to investigate electroclinical and neuropsychological features, genetic background, and evolution of children with idiopathic encephalopathy with status epilepticus during slow sleep (ESES), including Landau-Kleffner syndrome (LKS). MATERIAL AND METHODS: All children diagnosed with idiopathic ESES at the Danish Epilepsy Centre between March 2003 and December 2014 were retrospectively reviewed. Repeated 24-hour electroencephalography (24-h EEG) recordings, neuropsychological assessments, and clinical-neurological evaluation were performed throughout the follow-up in all patients. In 13 children, genetic investigations were performed. RESULTS: We collected 24 children (14 males and 10 females). Mean age at ESES diagnosis was 6 years, and mean ESES duration was 2 years and 7 months. Twenty-one children had epileptic seizures. Three children had LKS. Topography of sleep-related EEG epileptic abnormalities was diffuse in 3 subjects, hemispheric in 6, multifocal in 9, and focal in 6. During the active phase of ESES, all children presented with a heterogeneous combination of behavioral and cognitive disturbances. In 14 children, a parallel between severity of the clinical picture and spike-wave index (SWI) was observed. We could not find a strict correlation between the type and severity of neurobehavioral impairment and the side/topography of sleep-related EEG discharges during the active phase of ESES. At the last follow-up, 21 children were in remission from ESES. Complete recovery from neurobehavioral disorders was observed in 5 children. Genetic assessment, performed in 13 children, showed GRIN2A variant in two (15.4%). SIGNIFICANCE: Our patients with idiopathic ESES showed a heterogeneous pattern of epileptic seizures, neurobehavioral disorders, and sleep EEG features. Only one-fourth of children completely recovered from the neuropsychological disturbances after ESES remission. Lack of correlation between severity/type of cognitive derangement and SWI and/or topography of sleep EEG epileptic abnormalities may suggest the contribution of additional factors (including impaired sleep homeostasis due to epileptic activity) in the neurobehavioral derangement that characterize ESES.

EEG - A Valuable Biomarker of Brain Injury in Preterm Infants.

This review focuses on the role of electroencephalography (EEG) in monitoring abnormalities of preterm brain function. EEG features of the most common developmental brain injuries in preterm infants, including intraventricular haemorrhage, periventricular leukomalacia, and perinatal asphyxia, are described. We outline the most common EEG biomarkers associated with these injuries, namely seizures, positive rolandic sharp waves, EEG suppression/increased interburst intervals, mechanical delta brush activity, and other deformed EEG waveforms, asymmetries, and asynchronies. The increasing survival rate of preterm infants, in particular those that are very and extremely preterm, has led to a growing demand for a specific and shared characterization of the patterns related to adverse outcome in this unique population. This review includes abundant high-quality images of the EEG patterns seen in premature infants and will provide a valuable resource for everyone working in developmental neuroscience.

Electrographic Seizures during the Early Postnatal Period in Preterm Infants.

OBJECTIVE: To investigate the frequency and characteristics of electrographic seizures in preterm infants in the early postnatal period. STUDY DESIGN: Infants <32 weeks gestational age (GA) (n = 120) were enrolled for continuous multichannel electroencephalography (EEG) recording initiated as soon as possible after birth and continued for approximately up to 72 hours of age. Electrographic seizures were identified visually, annotated, and analyzed. Quantitative descriptors of the temporal evolution of seizures, including total seizure burden, seizure duration, and maximum seizure burden, were calculated. RESULTS: Median GA was 28.9 weeks (IQR, 26.6-30.3 weeks) and median birth weight was 1125 g (IQR, 848-1440 g). Six infants (5%; 95% CI, 1.9-10.6%) had electrographic seizures. Median total seizure burden, seizure duration, and maximum seizure burden were 40.3 minutes (IQR, 5.0-117.5 minutes), 49.6 seconds (IQR, 43.4-76.6 seconds), and 10.8 minutes/hour (IQR, 1.6-20.2 minutes/hour), respectively. Seizure burden was highest in 2 infants with significant abnormalities on neuroimaging. CONCLUSION: Electrographic seizures are infrequent within the first few days of birth in very preterm infants. Seizures in this population are difficult to detect accurately without continuous multichannel EEG monitoring.

Remission of encephalopathy with status epilepticus (ESES) during sleep renormalizes regulation of slow wave sleep.

OBJECTIVE: In previous studies, we showed an altered overnight decrease of non-rapid-eye-movement (NREM) sleep slow waves in children with encephalopathy related to status epilepticus during sleep (ESES). Here, we test the hypothesis that these alterations renormalize after remission of ESES. Because overnight decrease of slow waves has been linked to brain recovery and cognition, we investigate whether cognitive outcome is related to overnight changes of slow waves. METHODS: We performed a retrospective analysis of longitudinal overnight electroencephalography (EEG) in 10 patients with idiopathic ESES. Automated slow wave detection and calculation of slope of slow waves during the first and last hour of NREM sleep were employed. Intraindividual comparisons were undertaken of the slope during active phase and after remission of ESES, and between patients after remission of ESES and healthy controls. Explorative analysis of the relationship between slow wave slope and cognitive outcome was performed. RESULTS: The slope of slow waves did not decrease significantly across the night during active ESES, particularly at the spike focus. After remission of ESES, the slope decreased significantly overnight. Compared to controls, there was no difference in overnight slope decrease. Association between slope and neuropsychological outcome showed best cognitive outcome after remission in those children (n = 3) who showed some degree of slope decline during active ESES. SIGNIFICANCE: This study provides evidence that alterations of overnight changes of NREM-sleep slow waves during active ESES are reversible when ESES resolves, and that the severity of neuropsychological compromise might be related to the extent of slow wave impairment during ESES. Our findings suggest that analysis of slow waves might serve as a prognostic factor regarding cognitive outcome. ESES may serve as disease model of pathologic slow wave sleep and our results might be expanded to epilepsies with spike wave activation in slow wave sleep not only in children but also in adults.

A review of important electroencephalogram features for the assessment of brain maturation in premature infants.

This review describes the maturational features of the baseline electroencephalogram (EEG) in the neurologically healthy preterm infant. Features such as continuity, sleep state, synchrony and transient waveforms are described, even from extremely preterm infants and includes abundant illustrated examples. The physiological significance of these EEG features and their relationship to neurodevelopment are highlighted where known. This review also demonstrates the importance of multichannel conventional EEG monitoring for preterm infants as many of the features described are not apparent if limited channel EEG monitors are used. CONCLUSION: This review aims to provide healthcare professionals in the neonatal intensive care unit with guidance on the more common normal maturational features seen in the EEG of preterm infants.

Reflex Myoclonic Epilepsy of Infancy: Seizures Induced by Tactile Stimulation.

Myoclonic epilepsy with reflex seizures in infancy is an extremely rare condition, in which seizures are provoked mainly by auditory or auditory-tactile stimuli. To increase the awareness of pediatricians regarding this underrecognized condition, we describe a child with seizures provoked only by the tactile stimulation of specific areas of the head and face.

Predicting 2-y outcome in preterm infants using early multimodal physiological monitoring.

BACKGROUND: Preterm infants are at risk of adverse outcome. The aim of this study is to develop a multimodal model, including physiological signals from the first days of life, to predict 2-y outcome in preterm infants. METHODS: Infants <32 wk gestation had simultaneous multi-channel electroencephalography (EEG), peripheral oxygen saturation (SpO2), and heart rate (HR) monitoring. EEG grades were combined with gestational age (GA) and quantitative features of HR and SpO2 in a logistic regression model to predict outcome. Bayley Scales of Infant Development-III assessed 2-y neurodevelopmental outcome. A clinical course score, grading infants at discharge as high or low morbidity risk, was used to compare performance with the model. RESULTS: Forty-three infants were included: 27 had good outcomes, 16 had poor outcomes or died. While performance of the model was similar to the clinical course score graded at discharge, with an area under the receiver operator characteristic (AUC) of 0.83 (95% confidence intervals (CI): 0.69-0.95) vs. 0.79 (0.66-0.90) (P = 0.633), the model was able to predict 2-y outcome days after birth. CONCLUSION: Quantitative analysis of physiological signals, combined with GA and graded EEG, shows potential for predicting mortality or delayed neurodevelopment at 2 y of age.

Phenobarbital for Neonatal Seizures: Response Rate and Predictors of Refractoriness.

Background Phenobarbital is the first-line choice for neonatal seizures treatment, despite a response rate of approximately 45%. Failure to respond to acute anticonvulsants is associated with poor neurodevelopmental outcome, but knowledge on predictors of refractoriness is limited. Objective To quantify response rate to phenobarbital and to establish variables predictive of its lack of efficacy. Methods We retrospectively evaluated newborns with electrographically confirmed neonatal seizures admitted between January 1999 and December 2012 to the neonatal intensive care unit of Parma University Hospital (Italy), excluding neonates with status epilepticus. Response was categorized as complete (cessation of clinical and electrographic seizures after phenobarbital administration), partial (reduction but not cessation of electrographic seizures with the first bolus, response to the second bolus), or absent (no response after the second bolus). Multivariate analysis was used to identify independent predictors of refractoriness. Results Out of 91 newborns receiving phenobarbital, 57 (62.6%) responded completely, 15 (16.5%) partially, and 19 (20.9%) did not respond. Seizure type (p = 0.02), background electroencephalogram (EEG; p ≤ 0.005), and neurologic examination (p ≤ 0.005) correlated with response to phenobarbital. However, EEG (p ≤ 0.02) and seizure type (p ≤ 0.001) were the only independent predictors. Conclusion Our results suggest a prominent role of neurophysiological variables (background EEG and electrographic-only seizure type) in predicting the absence of response to phenobarbital in high-risk newborns.

Optimizing Detection Rate and Characterization of Subtle Paroxysmal Neonatal Abnormal Facial Movements with Multi-Camera Video-Electroencephalogram Recordings.

Objectives We retrospectively analyze the diagnostic accuracy for paroxysmal abnormal facial movements, comparing one camera versus multi-camera approach. Background Polygraphic video-electroencephalogram (vEEG) recording is the current gold standard for brain monitoring in high-risk newborns, especially when neonatal seizures are suspected. One camera synchronized with the EEG is commonly used. Methods Since mid-June 2012, we have started using multiple cameras, one of which point toward newborns' faces. We evaluated vEEGs recorded in newborns in the study period between mid-June 2012 and the end of September 2014 and compared, for each recording, the diagnostic accuracies obtained with one-camera and multi-camera approaches. Results We recorded 147 vEEGs from 87 newborns and found 73 episodes of paroxysmal facial abnormal movements in 18 vEEGs of 11 newborns with the multi-camera approach. By using the single-camera approach, only 28.8% of these events were identified (21/73). Ten positive vEEGs with multicamera with 52 paroxysmal facial abnormal movements (52/73, 71.2%) would have been considered as negative with the single-camera approach. Conclusions The use of one additional facial camera can significantly increase the diagnostic accuracy of vEEGs in the detection of paroxysmal abnormal facial movements in the newborns.

Xp11.22 Microduplications Including HUWE1: Case Report and Literature Review.

Xp11.22 microduplications have been reported in different patients with X-linked intellectual disability. Comparing the duplicated segments, a minimum region of overlap has been identified. Within this region, only one gene, the HUWE1 gene, coding the E3 ubiquitin protein ligase, turned out to be duplicated in all previously described patients. We provide a review of the literature on this topic, making a comparison not only of genetic aspects, but also of clinical, neurophysiological, and neuroradiological findings. Furthermore, we describe the phenotypic and molecular characterization of a case of intellectual disability in a child carrying one of the smallest Xp11.22 microduplications reported, involving the whole sequence of HUWE1 gene. Unlike previously described cases, our patient's neuroimaging showed abnormal findings; he also experienced one seizure and showed interictal electroencephalogram (EEG) epileptiform abnormalities. Given the fact that HUWE1 duplications and mutations have previously been described in several patients with X-linked cognitive impairment, our findings support the hypothesis that HUWE1 gene might be implicate in the pathogenesis of intellectual disability. Nevertheless, further investigations and a more detailed examination of patients' clinical history are needed to clear up other eventual genotype-phenotype correlations, such as the presence of epilepsy/epileptiform EEG abnormalities.

Febrile and postinfectious myoclonus: case reports and review of the literature.

PURPOSE: Myoclonus associated with fever or following an infectious event (parainfectious myoclonic jerks) is a transient event with an abrupt onset. It is not common and often not recognized. In these reports, its clinical features and its favorable outcome are described. PATIENTS AND METHODS: We identified three pediatric patients with parainfectious myoclonic jerks who were treated at Children's Neuropsychiatry Unit at the University-Hospital of Parma (Italy). RESULTS: The myoclonic jerks presented an abrupt onset and were characterized by brief involuntary muscle contractions temporally related to febrile episodes or following infectious illnesses. The outcome was positive as they resolve spontaneously with cessation of febrile/infective events. CONCLUSION: Parainfectious myoclonic jerks are temporary benign phenomena that could occur in association with the fever or following an infectious event. It is quite important to make a correct diagnosis as early as possible to avoid unnecessary hospitalizations or investigations.

Children's Headache: Drawings in the Diagnostic Work Up.

OBJECTIVES: This study aims to evaluate the drawings effectiveness in childhood headache assessment. BACKGROUND: Headache is a common cause of pain in children. Although drawings have been used in childhood to recognize psychological insights and pain perception, they were rarely used for headache characterization. METHODS: We collected drawings from 67 subjects with cephalalgia during a 22-month timeframe. The clinical diagnosis was made according to the 2nd edition of The International Headache Classification. Drawings were independently categorized as migraine or tension-type headache (TTH) by two child neuropsychiatrists blinded to the clinical data. Cohen kappa for interrater agreement, sensitivity, specificity, and positive predictive value (PPV) were calculated. Subjects were also divided into three age groups to assess the influence of age. Finally, a control group of 90 subjects was collected and K-means cluster analysis was performed. RESULTS: The drawings had a sensitivity of 85.71 and 81.48%, a specificity of 81.48 and 85.71%, and a PPV of 85.71 and 81.48%, for migraine and TTH diagnosis, respectively. Drawings by the older age group showed the highest predictability degree. Finally, by mean of cluster analysis, 59 of the 67 patients were correctly classified, whereas control subjects were similarly distributed between the two clusters. CONCLUSIONS: Drawings are a useful instrument for migraine and TTH differential diagnosis. Thus, we suggest their inclusion in childhood headache diagnostic assessment.

A painful stiff neck following an ear, nose, and throat surgical procedure: case report.

Grisel syndrome is a rare, nontraumatic atlantoaxial subluxation, typical of developmental ages and characterized by head flexion/rotation and painful fixation. Neurological symptoms may occur. It is secondary to head/neck infections and ear, nose, and throat surgery (adenoidectomy, tonsillectomy, and mastoidectomy). Here, we report the case of a child who presented a painful stiff neck following an adenotonsillectomy, with imaging evidencing an atlantoaxial subluxation. The child showed improvement in his condition following a conservative treatment with antibiotics, anti-inflammatory, and analgesic therapy and cervical collar. We believe it is of great significance for clinicians taking into account this peculiar condition in the differential diagnosis of a stiff neck in pediatric patients, thus avoiding misdiagnosis and delays. Indeed, its diagnosis is mainly based on a focused anamnesis associated with the detection of the typical neuroradiological findings.

Deep Learning for EEG Seizure Detection in Preterm Infants.

EEG is the gold standard for seizure detection in the newborn infant, but EEG interpretation in the preterm group is particularly challenging; trained experts are scarce and the task of interpreting EEG in real-time is arduous. Preterm infants are reported to have a higher incidence of seizures compared to term infants. Preterm EEG morphology differs from that of term infants, which implies that seizure detection algorithms trained on term EEG may not be appropriate. The task of developing preterm specific algorithms becomes extra-challenging given the limited amount of annotated preterm EEG data available. This paper explores novel deep learning (DL) architectures for the task of neonatal seizure detection in preterm infants. The study tests and compares several approaches to address the problem: training on data from full-term infants; training on data from preterm infants; training on age-specific preterm data and transfer learning. The system performance is assessed on a large database of continuous EEG recordings of 575[Formula: see text]h in duration. It is shown that the accuracy of a validated term-trained EEG seizure detection algorithm, based on a support vector machine classifier, when tested on preterm infants falls well short of the performance achieved for full-term infants. An AUC of 88.3% was obtained when tested on preterm EEG as compared to 96.6% obtained when tested on term EEG. When re-trained on preterm EEG, the performance marginally increases to 89.7%. An alternative DL approach shows a more stable trend when tested on the preterm cohort, starting with an AUC of 93.3% for the term-trained algorithm and reaching 95.0% by transfer learning from the term model using available preterm data. The proposed DL approach avoids time-consuming explicit feature engineering and leverages the existence of the term seizure detection model, resulting in accurate predictions with a minimum amount of annotated preterm data.

A standardised assessment scheme for conventional EEG in preterm infants.

OBJECTIVE: To develop a standardised scheme for assessing normal and abnormal electroencephalography (EEG) features of preterm infants. To assess the interobserver agreement of this assessment scheme. METHODS: We created a standardised EEG assessment scheme for 6 different post-menstrual age (PMA) groups using 4 EEG categories. Two experts, not involved in the development of the scheme, evaluated this on 24 infants <32 weeks gestational age (GA) using random 2 hour EEG epochs. Where disagreements were found, the features were checked and modified. Finally, the two experts independently evaluated 2 hour EEG epochs from an additional 12 infants <37 weeks GA. The percentage of agreement was calculated as the ratio of agreements to the sum of agreements plus disagreements. RESULTS: Good agreement in all patients and EEG feature category was obtained, with a median agreement between 80% and 100% over the 4 EEG assessment categories. No difference was found in agreement rates between the normal and abnormal features (p = 0.959). CONCLUSIONS: We developed a standard EEG assessment scheme for preterm infants that shows good interobserver agreement. SIGNIFICANCE: This will provide information to Neonatal Intensive Care Unit (NICU) staff about brain activity and maturation. We hope this will prove useful for many centres seeking to use neuromonitoring during critical care for preterm infants.

Reliability and accuracy of EEG interpretation for estimating age in preterm infants.

OBJECTIVES: To determine the accuracy of, and agreement among, EEG and aEEG readers' estimation of maturity and a novel computational measure of functional brain age (FBA) in preterm infants. METHODS: Seven experts estimated the postmenstrual ages (PMA) in a cohort of recordings from preterm infants using cloud-based review software. The FBA was calculated using a machine learning-based algorithm. Error analysis was used to determine the accuracy of PMA assessments and intraclass correlation (ICC) was used to assess agreement between experts. RESULTS: EEG recordings from a PMA range 25 to 38 weeks were successfully interpreted. In 179 recordings from 62 infants interpreted by all human readers, there was moderate agreement between experts (aEEG ICC = 0.724; 95%CI:0.658-0.781 and EEG ICC = 0.517; 95%CI:0.311-0.664). In 149 recordings from 61 infants interpreted by all human readers and the FBA algorithm, random and systematic errors in visual interpretation of PMA were significantly higher than the computational FBA estimate. Tracking of maturation in individual infants showed stable FBA trajectories, but the trajectories of the experts' PMA estimate were more likely to be obscured by random errors. The accuracy of visual interpretation of PMA estimation was compromised by neurodevelopmental outcome for both aEEG and EEG review. INTERPRETATION: Visual assessment of infant maturity is possible from the EEG or aEEG, with an average of human experts providing the highest accuracy. Tracking PMA of individual infants was hampered by errors in experts' estimates. FBA provided the most accurate maturity assessment and has potential as a biomarker of early outcome.

A machine-learning algorithm for neonatal seizure recognition: a multicentre, randomised, controlled trial.

BACKGROUND: Despite the availability of continuous conventional electroencephalography (cEEG), accurate diagnosis of neonatal seizures is challenging in clinical practice. Algorithms for decision support in the recognition of neonatal seizures could improve detection. We aimed to assess the diagnostic accuracy of an automated seizure detection algorithm called Algorithm for Neonatal Seizure Recognition (ANSeR). METHODS: This multicentre, randomised, two-arm, parallel, controlled trial was done in eight neonatal centres across Ireland, the Netherlands, Sweden, and the UK. Neonates with a corrected gestational age between 36 and 44 weeks with, or at significant risk of, seizures requiring EEG monitoring, received cEEG plus ANSeR linked to the EEG monitor displaying a seizure probability trend in real time (algorithm group) or cEEG monitoring alone (non-algorithm group). The primary outcome was diagnostic accuracy (sensitivity, specificity, and false detection rate) of health-care professionals to identify neonates with electrographic seizures and seizure hours with and without the support of the ANSeR algorithm. Neonates with data on the outcome of interest were included in the analysis. This study is registered with ClinicalTrials.gov, NCT02431780. FINDINGS: Between Feb 13, 2015, and Feb 7, 2017, 132 neonates were randomly assigned to the algorithm group and 132 to the non-algorithm group. Six neonates were excluded (four from the algorithm group and two from the non-algorithm group). Electrographic seizures were present in 32 (25·0%) of 128 neonates in the algorithm group and 38 (29·2%) of 130 neonates in the non-algorithm group. For recognition of neonates with electrographic seizures, sensitivity was 81·3% (95% CI 66·7-93·3) in the algorithm group and 89·5% (78·4-97·5) in the non-algorithm group; specificity was 84·4% (95% CI 76·9-91·0) in the algorithm group and 89·1% (82·5-94·7) in the non-algorithm group; and the false detection rate was 36·6% (95% CI 22·7-52·1) in the algorithm group and 22·7% (11·6-35·9) in the non-algorithm group. We identified 659 h in which seizures occurred (seizure hours): 268 h in the algorithm versus 391 h in the non-algorithm group. The percentage of seizure hours correctly identified was higher in the algorithm group than in the non-algorithm group (177 [66·0%; 95% CI 53·8-77·3] of 268 h vs 177 [45·3%; 34·5-58·3] of 391 h; difference 20·8% [3·6-37·1]). No significant differences were seen in the percentage of neonates with seizures given at least one inappropriate antiseizure medication (37·5% [95% CI 25·0 to 56·3] vs 31·6% [21·1 to 47·4]; difference 5·9% [-14·0 to 26·3]). INTERPRETATION: ANSeR, a machine-learning algorithm, is safe and able to accurately detect neonatal seizures. Although the algorithm did not enhance identification of individual neonates with seizures beyond conventional EEG, recognition of seizure hours was improved with use of ANSeR. The benefit might be greater in less experienced centres, but further study is required. FUNDING: Wellcome Trust, Science Foundation Ireland, and Nihon Kohden.