A H Ansari1, P J Cherian2, A Dereymaeker3, V Matic4, K Jansen5, L De Wispelaere6, C Dielman7, J Vervisch8, R M Swarte9, P Govaert10, G Naulaers11, M De Vos12, S Van Huffel13. 1. Department of Electrical Engineering (ESAT), KU Leuven, Leuven, Belgium; iMinds Medical Information Technology, Leuven, Belgium. Electronic address: amirhossein.ansari@kuleuven.be. 2. Section of Clinical Neurophysiology, Department of Neurology, Erasmus MC, University Medical Center Rotterdam, The Netherlands; Division of Neurology, Department of Medicine, McMaster University, Hamilton, Canada. Electronic address: perumpij@mcmaster.ca. 3. Department of Development and Regeneration, University Hospitals Leuven, Neonatal Intensive Care Unit, KU Leuven, Leuven, Belgium. Electronic address: anneleen.dereymaeker@uzleuven.be. 4. Department of Electrical Engineering (ESAT), KU Leuven, Leuven, Belgium; iMinds Medical Information Technology, Leuven, Belgium. Electronic address: Vmatic@singidunum.ac.rs. 5. Department of Development and Regeneration, University Hospitals Leuven, Neonatal Intensive Care Unit, KU Leuven, Leuven, Belgium; Department of Development and Regeneration, University Hospitals Leuven, Child Neurology, KU Leuven, Leuven, Belgium. Electronic address: katrien.jansen@uzleuven.be. 6. Section of Neonatology, Department of Pediatrics, Sophia Children's Hospital, Erasmus MC, University Medical Center Rotterdam, The Netherlands. Electronic address: a.dewispelaere@erasmusmc.nl. 7. ZNA Koningin Paola Kinderziekenhuis, Antwerp, Belgium. Electronic address: charlotte.dielman@zna.be. 8. Department of Development and Regeneration, University Hospitals Leuven, Neonatal Intensive Care Unit, KU Leuven, Leuven, Belgium. Electronic address: jan.vervisch@uzleuven.be. 9. Section of Neonatology, Department of Pediatrics, Sophia Children's Hospital, Erasmus MC, University Medical Center Rotterdam, The Netherlands. Electronic address: r.swarte@erasmusmc.nl. 10. Section of Neonatology, Department of Pediatrics, Sophia Children's Hospital, Erasmus MC, University Medical Center Rotterdam, The Netherlands; ZNA Koningin Paola Kinderziekenhuis, Antwerp, Belgium. Electronic address: govaert@icloud.com. 11. Department of Development and Regeneration, University Hospitals Leuven, Neonatal Intensive Care Unit, KU Leuven, Leuven, Belgium. Electronic address: gunnar.naulaers@uzleuven.be. 12. Institute of Biomedical Engineering, Department of Engineering, University of Oxford, Oxford, UK. Electronic address: maarten.devos@eng.ox.ac.uk. 13. Department of Electrical Engineering (ESAT), KU Leuven, Leuven, Belgium; iMinds Medical Information Technology, Leuven, Belgium. Electronic address: Sabine.VanHuffel@esat.kuleuven.be.
Abstract
OBJECTIVE: After identifying the most seizure-relevant characteristics by a previously developed heuristic classifier, a data-driven post-processor using a novel set of features is applied to improve the performance. METHODS: The main characteristics of the outputs of the heuristic algorithm are extracted by five sets of features including synchronization, evolution, retention, segment, and signal features. Then, a support vector machine and a decision making layer remove the falsely detected segments. RESULTS: Four datasets including 71 neonates (1023h, 3493 seizures) recorded in two different university hospitals, are used to train and test the algorithm without removing the dubious seizures. The heuristic method resulted in a false alarm rate of 3.81 per hour and good detection rate of 88% on the entire test databases. The post-processor, effectively reduces the false alarm rate by 34% while the good detection rate decreases by 2%. CONCLUSION: This post-processing technique improves the performance of the heuristic algorithm. The structure of this post-processor is generic, improves our understanding of the core visually determined EEG features of neonatal seizures and is applicable for other neonatal seizure detectors. SIGNIFICANCE: The post-processor significantly decreases the false alarm rate at the expense of a small reduction of the good detection rate.
OBJECTIVE: After identifying the most seizure-relevant characteristics by a previously developed heuristic classifier, a data-driven post-processor using a novel set of features is applied to improve the performance. METHODS: The main characteristics of the outputs of the heuristic algorithm are extracted by five sets of features including synchronization, evolution, retention, segment, and signal features. Then, a support vector machine and a decision making layer remove the falsely detected segments. RESULTS: Four datasets including 71 neonates (1023h, 3493 seizures) recorded in two different university hospitals, are used to train and test the algorithm without removing the dubious seizures. The heuristic method resulted in a false alarm rate of 3.81 per hour and good detection rate of 88% on the entire test databases. The post-processor, effectively reduces the false alarm rate by 34% while the good detection rate decreases by 2%. CONCLUSION: This post-processing technique improves the performance of the heuristic algorithm. The structure of this post-processor is generic, improves our understanding of the core visually determined EEG features of neonatal seizures and is applicable for other neonatal seizure detectors. SIGNIFICANCE: The post-processor significantly decreases the false alarm rate at the expense of a small reduction of the good detection rate.
Authors: Ana Borovac; Steinn Gudmundsson; Gardar Thorvardsson; Saeed M Moghadam; Paivi Nevalainen; Nathan Stevenson; Sampsa Vanhatalo; Thomas P Runarsson Journal: IEEE J Transl Eng Health Med Date: 2022-08-23
Authors: Andriy Temko; Achintya Kr Sarkar; Geraldine B Boylan; Sean Mathieson; William P Marnane; Gordon Lightbody Journal: IEEE J Transl Eng Health Med Date: 2017-09-11 Impact factor: 3.316
Authors: Gabriel Fernando Todeschi Variane; João Paulo Vasques Camargo; Daniela Pereira Rodrigues; Maurício Magalhães; Marcelo Jenné Mimica Journal: Front Pediatr Date: 2022-03-23 Impact factor: 3.418
Authors: Edgar Guevara; Wyston C Pierre; Camille Tessier; Luis Akakpo; Irène Londono; Frédéric Lesage; Gregory A Lodygensky Journal: Front Neurosci Date: 2017-07-04 Impact factor: 4.677