Adaptable Preprocessing Units and Neural Classification for the Segmentation of EEG Signals

Abstract

In this contribution, a methodology for the simultaneous adaptation of preprocessing units (PPUs) for feature extraction and of neural classifiers that can be used for time series classification is presented. The approach is based upon an extension of the backpropagation algorithm for the correction of the preprocessing parameters. In comparison with purely neural systems, the reduced input dimensionality improves the generalization capability and reduces the numerical effort. In comparison with PPUs with fixed parameters, the success of the adaptation is less sensitive to the choice of the parameters. The efficiency of the developed method is demonstrated via the use of quadratic filters with adaptable transmission bands as preprocessing units for the segmentation of two different types of discontinuous EEG: discontinuous neonatal EEG (burst-interburst segmentation) and EEG in deep stages of sedation (burst-suppression segmentation).

Keywords

EEG Processing - Neural Networks - Classification - Intensive Care

• References

• 1 Widrow B, Stearns SD. Adaptive Signal Processing . Prentice-Hall signal processing series. Prentice-Hall; 1985
• 2 Baum EB, Haussler D. What size net gives valid generalization?. Neural Computation 1989; 1: 151-60.
• 3 Murata N, Yoshizawa S, Amari S. Network information criterion – determining the number of hidden units for artificial neural network models. IEEE Transactions on Neural Networks 1994; 5: 865-72.
• 4 Witte H, Putsche P, Eiselt M. et al. Analysis of the interrelations between a low-frequency and a high-frequency signal component in human neonatal EEG during quiet sleep. Neuroscience Letters 1997; 236: 175-9.
• 5 Witte H, Schelenz C, Specht M. et al. Interrelations between EEG frequency components in sedated intensive care patients during burst-suppression period. Neuroscience Letters 1999 (in press).
• 6 Ripley BD. Pattern Recognition and Neural Networks . Cambridge University Press; 1996
• 7 Bishop CM. Neural Networks for Pattern Recognition . Oxford: Clarendon Press; 1995
• 8 Doering A. Optimization of Feature Extraction and Classifier Structure for Pattern Recognition with Neural Networks . PhD thesis, TU Dresden; 1997
• 9 Deczky AG. Synthesis of recursive digital filters using the minimum P error criterion. IEEE Trans. Audio Electroacoustics 1972; 20: 257-63.
• 10 Blum JR. Multivariate stochastic approximation procedure. Annals Math Statistics 1954; 25: 737-44.
• 11 Parks TW, McClelland JU. Chebycheff approximation for nonrecursive digital filters with linear phase. IEEE Trans ET 1997; 19: 189-94.
• 12 Oppenheim AV, Schafer RW. Discrete-Time Signal Processing . Prentice Hall; 1989
• 13 Arnold M, Doering A, Witte H, Dörschel J, Eiselt M. Use of adaptive Hilbert transformation for EEG segmentation and calculation of instantaneous respiration rate in neonates. J Clin Monitoring 1996; 12: 43-60.
• 14 Galicki M, Witte H, Dörschel J, Eiselt M, Griessbach G. Common optimization of adaptive preprocessing units and a neural network during the learning period. Application in EEG pattern recognition. Neural Networks 1997; 10 (Suppl. 06) 1153-63.
• 15 Doering A, Witte H. Extended neural networks for signal detection and classification: An approach for simultaneous optimization of parameterized preprocessing and neural networks. In: Brender J. et al. Medical Informatics Europe 96 . IOS Press; 1996: 977-81.
• 16 Proakis JG, Manolakis DG. Introduction to digital signal processing . Macmillan Publishing Company; 1988
• 17 Eiselt M, Schendel M, Witte H, Dörschel J, Curzi-Dascalova L, D’Allest AM. Quantitative EEG analysis in premature and full-term newborns during quiet sleep Electro-enceph Clin Neurophysiol. 1997 103. 528-34.
• 18 Beydoun A, Yen CE, Drury I. Variance of interburst intervals in burst suppression. Electroenceph Clin Neurophysiol 1991; 79: 435-8.
• 19 Akrawi WP, Drummond JC, Kalkman CJ, Patel PM. A comparison of the electrophysiologic characteristics of EEG burst-suppression as produced by isoflurane, thiopental, etomidate and propofol. J Neurosurg Anesthesiol 1996; 8: 40-6.
• 20 Todd MM, Wu B, Warner DS, Maktabi M. The dose-related effects of nitric oxide synthase inhibition on cerebral blood flow during isoflurane and pentobarbital anesthesia. Anesthesiology 1994; 80: 1128-36.
• 21 Jäntti V, Yli-Hankala A. Correlation of instantaneous heart rate and EEG suppression during enflurane anaesthesia: synchronous inhibition of heart rate and cortical electrical activity. Electroenceph Clin Neurophysiol 1990; 76: 476-9.
• 22 Yli-Hankala A, Jäntti V. EEG burst-suppression patterns correlates with the instantaneous heart rate under isoflurane anaesthesia. Acta Anaesthesiol Scand 1990; 34: 665-8.