Seizure detection algorithm for neonates based on wave-sequence analysis
Introduction
Seizures are a common sign of neurological disorder in neonates (Lombroso, 1996, Volpe, 2001), and can be caused by virtually any condition that affects neonatal brain function (Tharp, 2002). The occurrence of seizures in the newborn are associated with increased morbidity and mortality (Scher et al., 1989, Legido et al., 1991, Lombroso, 1996). The incidence of seizures is greater in the neonatal period than at any other time in life (Mizrahi, 1999, Patrizi et al., 2003), occurring in up to 0.5% of newborn infants (Tharp, 2002). However, there are major clinical challenges in the recognition of neonatal seizures, related partly to the high incidence of sub-clinical seizures (55–85%) and in that sick newborns may be pharmacologically paralysed in the intensive care setting (Hellstrőm-Westas et al., 1985, Bye and Flanagan, 1995, Gotman et al., 1997a).
Neonatal seizures have extremely variable morphology, frequency, and topography, even within the same patient, and are much more variable than seizures in adults (Shewmon, 1990, Lombroso, 1996, Patrizi et al., 2003). Due to the challenges in 24 h access to the gold standard of conventional EEG, many neonatal units now utilise amplitude-integrated EEG (aEEG) to monitor brain functions and assist in seizure recognition (e.g. Murdoch-Eaton et al., 2001, Procaccio et al., 2001, Hellstrőm-Westas et al., 2003). The amplitude-integrated EEG has historically been produced from a single EEG-channel that has been bandpass filtered between 2 and 15 Hz, with amplification of the higher frequency components, rectified, then submitted to a part-linear, part-logarithmic amplitude compression, ‘peak-smoothed’ and time-compressed (Maynard et al., 1969, Prior et al., 1973). Patterns emerging on the aEEG trace are used to classify background EEG activity and to identify seizures (e.g. Greisen and Tape-recorded, 1994, Hellstrőm-Westas et al., 2003, de Vries and Hellstrőm-Westas, 2005). However, it has been recognized that non-expert use of aEEG traces can be difficult without access to the raw EEG record (Thordstein et al., 2000), and that focal, low amplitude or shorter than 1 min duration seizures are often missed (Rennie et al., 2004). Attempts to automatically detect seizures are using the aEEG trace alone have not been made until recently (Lommen et al., 2005), and seizure length <1 min cannot be detected by this method. Another approach to seizure detection is based on visual assessment of the compressed spectral analysis graph of EEG recorded continuously at the cotside (Aziz et al., 1986).
Several automatic seizure detection methods for conventional EEG in neonates have been developed, though not many have been systematically tested. Liu et al. (1992) suggested looking for heightened regularity in the positions of the first 4 peaks in the autocorrelation function. Nagasubramanian et al. (1997) identified heightened regularity by a decrease in the spectral exponent and decreases in the variability of the power in frequency bands, approximately corresponding to ‘clinically important’ frequencies (1–15 Hz). They proposed a wavelet based calculation to efficiently perform this identification. Roessgen et al. (1998) suggested a modification of da Silva's ‘local EEG model’ (da Silva et al., 1974) with an additional saw-tooth input, representing seizure activity. Using Maximum Likelihood techniques to estimate model parameters allows the power spectrum of EEG to be described as the sum of the background power spectrum, seizure power spectrum and Gaussian noise. Seizure activity is detected when the ratio of power in the modelled seizure to background spectrum exceeds a threshold value in favour of seizure spectrum. Celka and Colditz, 2002a, Celka and Colditz, 2002b presented an alternative to Roessgen et al. (1998) model of EEG. The background activity is modelled as a non-linear transformation of a Gaussian ARMA-process which has an ‘exact’ match to the observed distribution of background EEG voltages. An EEG signal is, therefore, transformed by applying the inverse of the non-linear transformation and ARMA-filter. Finally, singular spectrum analysis is used to test if the output looks like Gaussian white noise, with a false result indicating seizure. Smit et al. (2004) used the synchronization likelihood to measure statistical interdependence between two time series, i.e. between a pair of EEG channels, assuming that during seizure activity the synchronization likelihood would rise. Boashash and Mesbah, 2001, Hassanpour et al., 2004 presented some promising findings employing joint time–frequency analysis.
The only commercially available and widely tested algorithm for neonatal seizure detection has been developed by Jean Gotman (Gotman et al., 1997a). Three methods are used with a 10 s sliding foreground window (moving by 2.5 s increments); which is compared with a preceding 20 s background window separated from it by a 1 min interval. The first method is intended for the detection of rhythmic discharges of 0.5–10 Hz and is principally based on spectral analysis (Qu and Gotman, 1997). An increase in the power of the dominant power spectrum peak indicates the presence of seizure, subject to a range of conditions (Gotman et al., 1997a). The other methods are based on half-wave decomposition of EEG signal (Gotman and Gloor, 1976). The second method utilises the detection of multiple spikes (Gotman and Gloor, 1976, Gotman et al., 1979), where half-waves of certain length, sharpness and relative amplitude are classified as spikes, and the presence of 6 spikes in the foreground window indicates a seizure (Gotman et al., 1997a). The third method is designed to detect very slow rhythmic discharges. EEG is low-pass filtered with a cut-off frequency of 1 Hz (Gotman et al., 1997a). Then, to qualify as a seizure the foreground window has to exceed thresholds for relative amplitude and coefficient of correlation of half-wave durations (Gotman, 1982, Gotman, 1990).
All algorithms described above have one feature in common. They apply one or other method to a fixed time interval (usually a sliding window) in the EEG trace, trying to extract interval features and compare them with background features. The aim of this article is to present and evaluate a real-time algorithm based on the analysis not of time intervals, but of the wave sequences in the EEG recording.
Section snippets
EEG dataset
The first training dataset was composed from 52 short one-channel EEG traces totalling 171 min, collected from 6 neonates in Royal Children's and Royal Women's Hospitals, Melbourne, Australia and marked by a paediatric neurologist. The EEG recordings were collected by a research nurse using the ReBRIM-1 bedside two-channel EEG monitor (BrainZ Instruments Ltd). The part of the algorithm described in 2.3.1 EEG epochs and filtering, 2.3.2 Parallel wave fragmentation: detection of critical points,
Results
Out of 97 seizures marked in the dataset, 87 seizures were detected by the Navakatikyan algorithm (Table 2). There were also 49 false positives, or 2.0 false positives per hour. The sensitivity of the present algorithm assessed by different methods was 83–95%. The positive predictive value was 48–77%. The specificity value was 87–94%.
By comparison, the Gotman algorithm (Sensa, version 5.4) with the 30 s gap-closing procedure (GCP) (as in Gotman et al., 1997a, Gotman et al., 1997b) detected 71 of
Discussion
The Navakatikyan seizure detection algorithm is based primarily on the regularity of EEG waves. Other algorithms also rely on this specific property of neonatal seizures. Gotman's algorithm (Gotman et al., 1997a) has a method assessing the power spectrum, whereas Liu's algorithm detects regularity in the autocorrelation function (Liu et al., 1992); Nagasubramanian et al. (1997) use a decrease in the spectral exponent and decrease in the variability of the power in a range of frequency bands;
Disclosure
M.A. Navakatikyan, is an employee of BrainZ Instruments Ltd. P.B. Colditz, C.J. Burke, J. Richmond and T.E. Inder have research projects partially funded by BrainZ Instruments Ltd. C.E. Williams is partially supported by BrainZ Instruments Ltd at the University of Auckland.
Acknowledgements
We are grateful to Mark Gunning from Liggins Institute (Auckland, New Zealand), Mark Stewart, Randall Britten and Ben Whitmore from BrainZ Instruments Ltd (Auckland, New Zealand) for their assistance with mathematics and programming, and to Shelly Lavery from Royal Children's Hospital (Melbourne, Australia) for acquiring the first training dataset.
References (50)
- et al.
An evaluation of neonatal seizure detection methods
Clin Neurophysiol
(2005) Automatic recognition of epileptic seizures in the EEG
Electroenceph Clin Neurophysiol
(1982)Automatic seizure detection: improvements and evaluation
Electroenceph Clin Neurophysiol
(1990)- et al.
Automatic recognition and quantification of interictal epileptic activity in the human scalp EEG
Electroenceph Clin Neurophysiol
(1976) - et al.
Automatic recognition of inter-ictal epileptic activity in prolonged EEG recordings
Electroenceph Clin Neurophysiol
(1979) - et al.
Automatic seizure detection in the newborn: methods and initial evaluation
Electroenceph Clin Neurophysiol
(1997) - et al.
Evaluation of an automatic seizure detection method for the newborn EEG
Electroenceph Clin Neurophysiol
(1997) - et al.
Detection of neonatal seizures through computerized EEG analysis
Electroenceph Clin Neurophysiol
(1992) Neonatal seizures: a clinician's overview
Brain Dev
(1996)- et al.
Value of the EEG in neonatal seizures
J Epilepsy
(1993)
Neonatal seizures: characteristics of EEG ictal activity in preterm and fullterm infants
Brain Dev
EEG diagnostics of neonatal seizures: clinical correlations and outcome
Pediatr Neurol
Seizure detection: correlation of human experts
Clin Neurophysiol
Automatic detection of epileptiform events in EEG by a three-stage procedure based on artificial neural networks
IEEE Trans Biomed Eng
Costside EEG monitoring using computerized spectral snalysis
Arch Dis Child
A time-frequency approach for newborn seizure detection
IEEE Eng Med Biol Mag
Spatial and temporal characteristics of neonatal seizures
Epilepsia
A computer-aided detection of EEG seizures in infants: a singular-spectrum approach and performance comparison
IEEE Trans Biomed Eng
Nonlinear nonstationary Wiener model of infant EEG seizures
IEEE Trans Biomed Eng
The exact ictal and interictal duration of electroencephalographic neonatal seizures
Epilepsia
Model of brain rhythmic activity: the alpha-rhythm of the thalamus
Kybernetik.
Role of cerebral function monitoring in the newborn
Arch Dis Child Fetal Neonatal Ed
Computer-aided seizure detection in newborn infants
and the cerebral function monitor: amplitude-integrated, time-compressed EEG
J Perinat Med
Cited by (123)
Neonatal EEG seizure detection using a new signal structural complexity measure based on matching pursuit decomposition with nonstationary dictionary
2022, Computer Methods and Programs in BiomedicineValidating an SVM-based neonatal seizure detection algorithm for generalizability, non-inferiority and clinical efficacy
2022, Computers in Biology and MedicineNeonatal Seizure Management: Is the Timing of Treatment Critical?
2022, Journal of PediatricsConvolutional neural networks ensemble model for neonatal seizure detection
2021, Journal of Neuroscience MethodsA machine-learning algorithm for neonatal seizure recognition: a multicentre, randomised, controlled trial
2020, The Lancet Child and Adolescent HealthAutomated EEG analysis for neonatal intensive care
2019, Encyclopedia of Biomedical Engineering