eeg and ecg

download eeg and ecg

of 12

  • date post

  • Category


  • view

  • download


Embed Size (px)

Transcript of eeg and ecg

  • 8/12/2019 eeg and ecg


    Combination of EEG and ECG for improved automaticneonatal seizure detection

    Barry R. Greene a,*, Geraldine B. Boylan b, Richard B. Reilly a,c,Philip de Chazal a, Sean Connolly d

    a School of Electrical, Electronic & Mechanical Engineering, University College Dublin, Irelandb Department of Paediatrics and Child Health, University College Cork, Ireland

    c Cognitive Neurophysiology Laboratory, St. Vincents Hospital, Fairview, Dublin, Irelandd Department of Clinical Neurophysiology, St. Vincents University Hospital, Dublin, Ireland

    Accepted 7 February 2007Available online 29 March 2007


    Objective:Neonatal seizures are the most common central nervous system disorder in newborn infants. A system that could automat-ically detect the presence of seizures in neonates would be a significant advance facilitating timely medical intervention.Methods: A novel method is proposed for the robust detection of neonatal seizures through the combination of simultaneously-recordedelectroencephalogram (EEG) and electrocardiogram (ECG). A patient-specific and a patient-independent system are considered,employing statistical classifier models.Results: Results for the signals combined are compared to results for each signal individually. For the patient-specific system, 617 of 633(97.52%) expert-labelled seizures were correctly detected with a false detection rate of 13.18%. For the patient-independent system, 516 of

    633 (81.44%) expert-labelled seizures were correctly detected with a false detection rate of 28.57%.Conclusions: A novel algorithm for neonatal seizure detection is proposed. The combination of an ECG-based classifier system with anovel multi-channel EEG-based classifier system has led to improved seizure detection performance. The algorithm was evaluated using alarge data-set containing ECG and multi-channel EEG of realistic duration and quality.Significance:Analysis of simultaneously-recorded EEG and ECG represents a new approach in seizure detection research and the detec-tion performance of the proposed system is a significant improvement on previous reported results for automated neonatal seizuredetection. 2007 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.

    Keywords: Neonatal seizure detection; EEG; ECG; EKG

    1. Introduction

    Seizures in the neonate require immediate medical atten-tion and represent a distinctive sign of central nervous sys-tem dysfunction. There is increasing evidence that neonatalseizures have an adverse effect on neurodevelopmental out-come, and predispose to cognitive, behavioural, or epilepticcomplications in later life (Levene, 2002). Neonatal seizures

    occur in 6% of low birth-weight infants (Volpe, 2001) and

    in approximately 2% of all newborns admitted to a neona-tal ICU (Scher et al., 1993a). Seizures in this age-group areoften subtle, difficult to diagnose and may be clinicallysilent, particularly after antiepileptic drug treatment, mak-ing diagnosis by clinical observation alone very unreliable(Boylan et al., 2002). Electroencephalography (EEG) isthe most reliable method available to detect the majorityof neonatal seizures but interpretation requires specialexpertise that is not readily available in most neonatalintensive care units least so on a 24-h basis. A systemthat could automatically detect the presence of seizures in

    1388-2457/$32.00 2007 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved.


    * Corresponding author. Tel.: +353 21 490 3793.E-mail Greene).

    Clinical Neurophysiology 118 (2007) 13481359
  • 8/12/2019 eeg and ecg


    newborn babies would be a significant advance, facilitatingtimely medical intervention.

    A number of studies have reported neonatal seizuredetection methods based on the EEG (Liu et al., 1992; Got-man et al., 1997a; Celka and Colditz, 2002; Altenburget al., 2003).Faul et al. (2005)provided a review and exper-

    imental comparison of three of the most commonly citedneonatal seizure detection algorithms. None performedsufficiently to be deemed suitable for use in the neonatalintensive care unit (ICU). Karayiannis et al. (2001)reported a video-based method for distinguishing myo-clonic from focal clonic seizures and differentiating thesetypes of seizures from normal infant behaviours. However,this approach does not provide a complete solution to theproblem, as many neonatal seizures are not accompaniedby this spectrum of body movements.

    The importance of autonomic changes may be underes-timated in neonatal seizure detection research. Neonatalseizures are often associated with changes in heart and res-

    piration rate (Greene et al., 2006b). Significant changes inheart rate may alert the clinician to the possibility of sei-zures and instigate further investigation with EEG. Thesefindings led to the development of a neonatal seizure detec-tion system based exclusively on the electrocardiogram(ECG) (Greene et al., 2006a).

    The aim of this study was to attempt to improve theneonatal seizure detection rate by combining simulta-

    neously-acquired ECG and EEG data. To the best of ourknowledge this is the first method to combine the ECGwith the EEG for seizure detection.

    2. Data-set

    A data-set of 12 records from 10 term neonates contain-ing 633 labelled seizure events, with mean seizure durationof 4.60 min, were recorded and analysed. The records had amean duration of 12.84 h. Each record contained 712channels of EEG and one channel of simultaneously-acquired ECG. Ten records, sampled at 256 Hz, were madein the neonatal intensive care units of the Unified Mater-nity Hospitals in Cork, Ireland, using the Viasys NicOnevideo-EEG system. The remaining recording, sampled at200 Hz, was recorded at Kings College Hospital, London,on a Telefactor Beehive video-EEG system. A total of154.1 h of EEG and ECG were analyzed.

    The data-set used in this research is a resource of con-

    tinuously-recorded digital video-EEG data and otherphysiological parameters in newborns with seizures inthe first 3 days from birth. All newborns were full term(GA: 4042 weeks) and had hypoxic ischaemic encepha-lopathy (HIE). All the data for each recording wereincluded in the analysis regardless of record length orquality. Electrographic seizures were identified and anno-tated by an expert in neonatal EEG (GBB). Fig. 1 shows

    Fig. 1. Example of a multi-channel electrographic seizure. Seizure onset and duration are marked.

    B.R. Greene et al. / Clinical Neurophysiology 118 (2007) 13481359 1349

  • 8/12/2019 eeg and ecg


    an example of an electrographic seizure from record 12,detected by both the patient-specific and patient-indepen-dent systems.

    Annotations give information on the time of onset andthe duration of each electrographic seizure.Table 1detailsthe number of seizure events per record, the duration ofeach record and the mean seizure duration for each record.As the ECG and EEG signals were recorded simulta-neously these annotations can be related directly in timeto the ECG signal.

    The data-set contained a wide variety of seizure dura-

    tions and seizure types. While the mean seizure durationacross the data-set was 4.60 min, the mean seizure durationfor each patient ranged from 1.05 min to 11.64 min. Thedata-set contained Electrographic-only seizures as wellas Electroclinical seizures. Four records 2, 3, 10, 12 con-tained only Electrographic-only seizures. Two records 9and 11 contained only Electroclinical seizures. Theremaining recordings contained both Electrographic-onlyand Electroclinical seizures. Furthermore, the data-setcontained focal, multi-focal and generalized seizures.

    3. Method

    The combination of EEG and ECG for neonatal sei-zure detection was considered in the context of bothpatient-specific and patient-independent seizure detectionclassifiers. While the ideal scenario for this applicationis a patient-independent system capable of identifyingall seizures from any patient with a zero false detectionrate, a patient-specific system might also represent anadvance in neonatal ICU monitoring. The algorithmsconsidered in this study are epoch-based, so each seizureevent was rounded to the nearest epoch length whenmapping time annotations to epochs. An epoch contain-ing P50% electrographic seizure activity was labelled as

    a seizure epoch.

    3.1. ECG

    The algorithm reported in this paper utilises the sameECG features described previously, based on the RRintervals for 60-s epochs of ECG (Greene et al., 2006a).

    3.1.1. ECG pre-processingAll ECG signals were filtered with a 20th order FIRband-pass filter (corner frequencies 8 and 18 Hz) to removebaseline wander, power-line noise and out of band noise.Before filtering, the mean of the ECG was removed fromthe signal.

    3.1.2. RR interval calculation

    The RR interval is defined as the time in secondsbetween adjacent R-wave maximum (QRS) points. Robustdetection of the QRS point is determined using a QRSdetection algorithm as described byBenitez et al. (2001).

    The Hilbert transform of the first derivative of the signal

    was used to emphasize the R peaks. A moving windowpeak search was carried out with an adaptive threshold.As neonatal ECG often manifests elevated P-wave, a stepback search was performed to isolate the P peak ensuringrobust detection of the R-wave maximum. Correction formissing and extra QRS points was implemented asdescribed byde Chazal et al. (2003).

    3.1.3. ECG feature extraction

    The six ECG feature types co