Affordable Access

deepdyve-link
Publisher Website

Time-Varying EEG Correlations Improve Automated Neonatal Seizure Detection.

Authors
  • Tapani, Karoliina T1, 2
  • Vanhatalo, Sampsa1
  • Stevenson, Nathan J3, 4
  • 1 1 BABA Center, Children's Hospital, HUS Medical Imaging Center, Clinical Neurophysiology, University of Helsinki, Helsinki University Hospital and University of Helsinki, Finland. , (Finland)
  • 2 2 Medical Imaging and Radiation Therapy, Kymenlaakso Central Hospital, Kymenlaakso Social and Health Services (Carea) Kotka, Finland. , (Finland)
  • 3 3 Clinicum, Faculty of Medicine, University of Helsinki, Helsinki, Finland. , (Finland)
  • 4 4 Brain Modelling Group, QIMR Berghofer Medical Research Institute, Brisbane, Australia. , (Australia)
Type
Published Article
Journal
International journal of neural systems
Publication Date
May 01, 2019
Volume
29
Issue
4
Pages
1850030–1850030
Identifiers
DOI: 10.1142/S0129065718500302
PMID: 30086662
Source
Medline
Keywords
Language
English
License
Unknown

Abstract

The aim of this study was to develop methods for detecting the nonstationary periodic characteristics of neonatal electroencephalographic (EEG) seizures by adapting estimates of the correlation both in the time (spike correlation; SC) and time-frequency domain (time-frequency correlation; TFC). These measures were incorporated into a seizure detection algorithm (SDA) based on a support vector machine to detect periods of seizure and nonseizure. The performance of these nonstationary correlation measures was evaluated using EEG recordings from 79 term neonates annotated by three human experts. The proposed measures were highly discriminative for seizure detection (median AUCSC : 0.933 IQR: 0.821-0.975, median AUCTFC : 0.883 IQR: 0.707-0.931). The resultant SDA applied to multi-channel recordings had a median AUC of 0.988 (IQR: 0.931-0.998) when compared to consensus annotations, outperformed two state-of-the-art SDAs (p < 0.001) and was noninferior to the human expert for 73/79 of neonates.

Report this publication

Statistics

Seen <100 times