Improvement and Validation of an Automated Neonatal Seizure Detector

We present the improvements made to and subsequent validation of an automated approach to detect neonatal seizures. The evaluation of the algorithm has been performed on a new and extensive data set of neonatal EEGs. Previously, we have classified neonatal seizures visually into two types: the spike train and oscillatory type of seizures and developed two separate algorithms that run in parallel for their automated detection. The first algorithm analyzes the correlation between high-energetic segments of the EEG, whereas the second one detects increases in low-frequency activity (<8 Hz) and then uses an autocorrelation. An improved version of our automated system (called 'NeoGuard') uses more informative features for classification and optimized parameters for thresholding. The validation was performed on 756 hours of 'unseen' continuous EEG monitoring data from 24 neonates with encephalopathy and recorded seizures. The seizure detection system showed a median sensitivity of 86.9 % per patient, positive predictive value (PPV) of 89.5 % and false positive rate of 0.28 per hour. The modified algorithm has a high sensitivity combined with a good PPV whereas false positive rate is much lower compared to the previous version of the algorithm.

[1]  Paul B. Colditz,et al.  A computer-aided detection of EEG seizures in infants: a singular-spectrum approach and performance comparison , 2002, IEEE Transactions on Biomedical Engineering.

[2]  Mostefa Mesbah,et al.  An optimal feature set for seizure detection systems for newborn EEG signals , 2003, Proceedings of the 2003 International Symposium on Circuits and Systems, 2003. ISCAS '03..

[3]  D B Vigneron,et al.  Seizure-associated brain injury in term newborns with perinatal asphyxia , 2002, Neurology.

[4]  J Connell,et al.  Continuous EEG monitoring of neonatal seizures: diagnostic and prognostic considerations. , 1989, Archives of disease in childhood.

[5]  A. Liu,et al.  Detection of neonatal seizures through computerized EEG analysis. , 1992, Electroencephalography and clinical neurophysiology.

[6]  J. Gotman,et al.  Automatic seizure detection in the newborn: methods and initial evaluation. , 1997, Electroencephalography and clinical neurophysiology.

[7]  Y. Ben-Ari,et al.  Consequences of neonatal seizures in the rat: Morphological and behavioral effects , 1998, Annals of neurology.

[8]  J. Frost,et al.  A Multistage System for the Automated Detection of Epileptic Seizures in Neonatal Electroencephalography , 2009, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[9]  D A Shewmon,et al.  What Is a Neonatal Seizure? Problems in Definition and Quantification for Investigative and Clinical Purposes , 1990, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[10]  S. Huffel,et al.  Automated neonatal seizure detection mimicking a human observer reading EEG , 2008, Clinical Neurophysiology.

[11]  R. Reilly,et al.  Combination of EEG and ECG for improved automatic neonatal seizure detection , 2007, Clinical Neurophysiology.

[12]  Ping Zhou,et al.  Teager–Kaiser Energy Operation of Surface EMG Improves Muscle Activity Onset Detection , 2007, Annals of Biomedical Engineering.

[13]  C. Lombroso,et al.  Neonatal seizures: a clinician's overview , 1996, Brain and Development.

[14]  R. Grebe,et al.  Automated neonatal seizure detection: A multistage classification system through feature selection based on relevance and redundancy analysis , 2006, Clinical Neurophysiology.

[15]  Sean Connolly,et al.  Interobserver agreement in neonatal seizure identification , 2009, Epilepsia.

[16]  Perumpillichira J. Cherian,et al.  Technical standards for recording and interpretation of neonatal electroencephalogram in clinical practice , 2009, Annals of Indian Academy of Neurology.