TY - GEN
T1 - A review of techniques for automatic detection of neonatal seizure using EEG
AU - Ryan, N.
AU - Mesbah, M.
AU - Boashash, B.
PY - 1999
Y1 - 1999
N2 - This investigation is part of a major project of the Signal Processing Research Centre (SPRC) to develop a technique to automatically detect epileptic seizure in newborns using EEG. Currently there are three published techniques under examination by the SPRC that aim to achieve this. The technique of Roessgen et al. (1998) is model based and uses parameter estimation for detection. The two other methods are non-parametric. The technique of Gotman et al. (1997) uses frequency analysis to detect changes in the dominant peak of the frequency spectrum of short epochs of EEG data. The technique of Liu et al. (1992) performs analysis in the time domain and is based on the autocorrelation function of short epochs of EEG data. Despite varying approaches, the techniques investigated here all attempt to detect periodicity in the EEG. This periodicity is the main characteristic of EEG seizure waveforms. The underlying methodologies of the three published techniques are discussed. Implementation of the three techniques is also discussed. Further work will involve the comparison of the three implementations on a common set of neonatal EEG recordings. It is anticipated that time-frequency analysis of neonatal EEG will be pursued as the basis for future detection techniques.
AB - This investigation is part of a major project of the Signal Processing Research Centre (SPRC) to develop a technique to automatically detect epileptic seizure in newborns using EEG. Currently there are three published techniques under examination by the SPRC that aim to achieve this. The technique of Roessgen et al. (1998) is model based and uses parameter estimation for detection. The two other methods are non-parametric. The technique of Gotman et al. (1997) uses frequency analysis to detect changes in the dominant peak of the frequency spectrum of short epochs of EEG data. The technique of Liu et al. (1992) performs analysis in the time domain and is based on the autocorrelation function of short epochs of EEG data. Despite varying approaches, the techniques investigated here all attempt to detect periodicity in the EEG. This periodicity is the main characteristic of EEG seizure waveforms. The underlying methodologies of the three published techniques are discussed. Implementation of the three techniques is also discussed. Further work will involve the comparison of the three implementations on a common set of neonatal EEG recordings. It is anticipated that time-frequency analysis of neonatal EEG will be pursued as the basis for future detection techniques.
UR - http://www.scopus.com/inward/record.url?scp=84904433427&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84904433427&partnerID=8YFLogxK
U2 - 10.1109/ISSPA.1999.818157
DO - 10.1109/ISSPA.1999.818157
M3 - Conference contribution
AN - SCOPUS:84904433427
SN - 1864354518
SN - 9781864354515
T3 - ISSPA 1999 - Proceedings of the 5th International Symposium on Signal Processing and Its Applications
SP - 239
EP - 242
BT - ISSPA 1999 - Proceedings of the 5th International Symposium on Signal Processing and Its Applications
PB - IEEE Computer Society
T2 - 5th International Symposium on Signal Processing and Its Applications, ISSPA 1999
Y2 - 22 August 1999 through 25 August 1999
ER -