Building an Open Source Classifier for the Neonatal EEG Background: A Systematic Feature-Based Approach From Expert Scoring to Clinical Visualization

Overview

Journal Front Hum Neurosci

Specialty Neurology

Date 2021 Jun 17

PMID 34135744

Citations 6

Authors

Saeed Montazeri

Elana Pinchefsky

Ilse Tse

Viviana Marchi

Jukka Kohonen

Minna Kauppila

Manu Airaksinen

Karoliina Tapani

Paivi Nevalainen

Cecil Hahn

Emily W Y Tam

Nathan J Stevenson

Sampsa Vanhatalo

Affiliations

Soon will be listed here.

Abstract

Neonatal brain monitoring in the neonatal intensive care units (NICU) requires a continuous review of the spontaneous cortical activity, i.e., the electroencephalograph (EEG) background activity. This needs development of bedside methods for an automated assessment of the EEG background activity. In this paper, we present development of the key components of a neonatal EEG background classifier, starting from the visual background scoring to classifier design, and finally to possible bedside visualization of the classifier results. A dataset with 13,200 5-minute EEG epochs (8-16 channels) from 27 infants with birth asphyxia was used for classifier training after scoring by two independent experts. We tested three classifier designs based on 98 computational features, and their performance was assessed with respect to scoring system, pre- and post-processing of labels and outputs, choice of channels, and visualization in monitor displays. The optimal solution achieved an overall classification accuracy of 97% with a range across subjects of 81-100%. We identified a set of 23 features that make the classifier highly robust to the choice of channels and missing data due to artefact rejection. Our results showed that an automated bedside classifier of EEG background is achievable, and we publish the full classifier algorithm to allow further clinical replication and validation studies.

Citing Articles

Machine learning techniques for predicting neurodevelopmental impairments in premature infants: a systematic review.

Ortega-Leon A, Urda D, Turias I, Lubian-Lopez S, Benavente-Fernandez I Front Artif Intell. 2025; 8:1481338.

PMID: 39906903 PMC: 11788297. DOI: 10.3389/frai.2025.1481338.

Automated assessment of EEG background for neurodevelopmental prediction in neonatal encephalopathy.

Lagace M, Montazeri S, Kamino D, Mamak E, Ly L, Hahn C Ann Clin Transl Neurol. 2024; 11(12):3267-3279.

PMID: 39543820 PMC: 11651191. DOI: 10.1002/acn3.52233.

Quantitative EEG features during the first day correlate to clinical outcome in perinatal asphyxia.

Tuiskula A, Pospelov A, Nevalainen P, Montazeri S, Metsaranta M, Haataja L Pediatr Res. 2024; 97(1):261-267.

PMID: 38745028 PMC: 11798844. DOI: 10.1038/s41390-024-03235-y.

Deep Learning for Generalized EEG Seizure Detection after Hypoxia-Ischemia-Preclinical Validation.

Abbasi H, Davidson J, Dhillon S, Zhou K, Wassink G, Gunn A Bioengineering (Basel). 2024; 11(3).

PMID: 38534490 PMC: 10968073. DOI: 10.3390/bioengineering11030217.

Networks of cortical activity show graded responses to perinatal asphyxia.

Syvalahti T, Tuiskula A, Nevalainen P, Metsaranta M, Haataja L, Vanhatalo S Pediatr Res. 2023; 96(1):132-140.

PMID: 38135725 PMC: 11258028. DOI: 10.1038/s41390-023-02978-4.

References

Raurale S, Boylan G, Mathieson S, Marnane W, Lightbody G, OToole J . Grading hypoxic-ischemic encephalopathy in neonatal EEG with convolutional neural networks and quadratic time-frequency distributions. J Neural Eng. 2021; 18(4). PMC: 8208632. DOI: 10.1088/1741-2552/abe8ae. View

Stevenson N, Lauronen L, Vanhatalo S . The effect of reducing EEG electrode number on the visual interpretation of the human expert for neonatal seizure detection. Clin Neurophysiol. 2017; 129(1):265-270. DOI: 10.1016/j.clinph.2017.10.031. View

Weeke L, Dix L, Groenendaal F, Lemmers P, Dijkman K, Andriessen P . Severe hypercapnia causes reversible depression of aEEG background activity in neonates: an observational study. Arch Dis Child Fetal Neonatal Ed. 2017; 102(5):F383-F388. DOI: 10.1136/archdischild-2016-311770. View

Marics G, Cseko A, Vasarhelyi B, Zakarias D, Schuster G, Szabo M . Prevalence and etiology of false normal aEEG recordings in neonatal hypoxic-ischaemic encephalopathy. BMC Pediatr. 2013; 13:194. PMC: 4222766. DOI: 10.1186/1471-2431-13-194. View

Ter Horst H, Brouwer O, Bos A . Burst suppression on amplitude-integrated electroencephalogram may be induced by midazolam: a report on three cases. Acta Paediatr. 2004; 93(4):559-63. DOI: 10.1080/08035250410022882. View

Stevenson N, Tataranno M, Kaminska A, Pavlidis E, Clancy R, Griesmaier E . Reliability and accuracy of EEG interpretation for estimating age in preterm infants. Ann Clin Transl Neurol. 2020; 7(9):1564-1573. PMC: 7480927. DOI: 10.1002/acn3.51132. View

Toet M, Lemmers P . Brain monitoring in neonates. Early Hum Dev. 2009; 85(2):77-84. DOI: 10.1016/j.earlhumdev.2008.11.007. View

Olischar M, Klebermass K, Kuhle S, Hulek M, Messerschmidt A, Weninger M . Progressive posthemorrhagic hydrocephalus leads to changes of amplitude-integrated EEG activity in preterm infants. Childs Nerv Syst. 2003; 20(1):41-5. DOI: 10.1007/s00381-003-0809-y. View

Stevenson N, OToole J, Korotchikova I, Boylan G . Artefact detection in neonatal EEG. Annu Int Conf IEEE Eng Med Biol Soc. 2015; 2014:926-9. DOI: 10.1109/EMBC.2014.6943743. View

10.

de Vries L, Hellstrom-Westas L . Role of cerebral function monitoring in the newborn. Arch Dis Child Fetal Neonatal Ed. 2005; 90(3):F201-7. PMC: 1721888. DOI: 10.1136/adc.2004.062745. View

11.

Ansari A, Cherian P, Dereymaeker A, Matic V, Jansen K, De Wispelaere L . Improved multi-stage neonatal seizure detection using a heuristic classifier and a data-driven post-processor. Clin Neurophysiol. 2016; 127(9):3014-3024. DOI: 10.1016/j.clinph.2016.06.018. View

12.

Basheer I, Hajmeer M . Artificial neural networks: fundamentals, computing, design, and application. J Microbiol Methods. 2000; 43(1):3-31. DOI: 10.1016/s0167-7012(00)00201-3. View

13.

Raurale S, Boylan G, Lightbody G, OToole J . Grading the severity of hypoxic-ischemic encephalopathy in newborn EEG using a convolutional neural network. Annu Int Conf IEEE Eng Med Biol Soc. 2020; 2020:6103-6106. PMC: 7613058. DOI: 10.1109/EMBC44109.2020.9175337. View

14.

Temko A, Marnane W, Boylan G, Lightbody G . Clinical implementation of a neonatal seizure detection algorithm. Decis Support Syst. 2015; 70:86-96. PMC: 4394138. DOI: 10.1016/j.dss.2014.12.006. View

15.

Kamino D, Almazrooei A, Pang E, Widjaja E, Moore A, Chau V . Abnormalities in evoked potentials associated with abnormal glycemia and brain injury in neonatal hypoxic-ischemic encephalopathy. Clin Neurophysiol. 2020; 132(1):307-313. PMC: 7855101. DOI: 10.1016/j.clinph.2020.09.024. View

16.

Ansari A, De Wel O, Pillay K, Dereymaeker A, Jansen K, Van Huffel S . A convolutional neural network outperforming state-of-the-art sleep staging algorithms for both preterm and term infants. J Neural Eng. 2019; 17(1):016028. DOI: 10.1088/1741-2552/ab5469. View

17.

Thoresen M, Hellstrom-Westas L, Liu X, de Vries L . Effect of hypothermia on amplitude-integrated electroencephalogram in infants with asphyxia. Pediatrics. 2010; 126(1):e131-9. DOI: 10.1542/peds.2009-2938. View

18.

Korotchikova I, Stevenson N, Walsh B, Murray D, Boylan G . Quantitative EEG analysis in neonatal hypoxic ischaemic encephalopathy. Clin Neurophysiol. 2011; 122(8):1671-8. DOI: 10.1016/j.clinph.2010.12.059. View

19.

Murray D, Boylan G, Ryan C, Connolly S . Early EEG findings in hypoxic-ischemic encephalopathy predict outcomes at 2 years. Pediatrics. 2009; 124(3):e459-67. DOI: 10.1542/peds.2008-2190. View

20.

Banville H, Chehab O, Hyvarinen A, Engemann D, Gramfort A . Uncovering the structure of clinical EEG signals with self-supervised learning. J Neural Eng. 2020; 18(4). DOI: 10.1088/1741-2552/abca18. View