On the Use of Neural Network Techniques to Analyze Sleep EEG Data

This is the third communication on the use of neural network techniques to classify sleep stages. In our first communication we presented the algorithms and the selection of the feature space and its reduction by using evolutionary and genetic procedures. In our second communication we trained the evolutionary optimized networks on the basis of multiple subject data in context with some smoothing algorithms in analogy of Rechtschaffen and Kales (RK). In this third communication we could demonstrate that the robustness concerning individual specific features of automatically generated sleep profiles could be reasonably improved by an additional modification of the procedure used by SASCIA (Sleep Analysis System to Challenge Innovative Artificial Networks). The outputs of nine different networks that were created by the data of 9 different subjects were used simultaneously for classification. The medians of the values obtained in each output measure were selected for the allocation to a sleep stage. The fitness criteria of 16 automatically generated sleep profiles showed reasonable concordance with the expert profile. Even though in single cases the concordance between conventional RK classifications and automatically generated profiles were a few percentages lower, the average correct classification of the 12 classified subjects improved substantially, thus proving that the classifier is more robust against individuum-specific variability. Despite the fact that the expert generally employs three channels (EEG, EMG and EOG), at least to build up sleep profiles, the SASCIA system was able to produce profiles on the basis of only one EEG channel with 80% concordance and a correlation coefficient of 0.86. The feature selections were performed by genetic algorithms and the topologies of the networks were optimized by evolutionary algorithms. This algorithm will now be used for larger sample forward classification.

[1]  A. Rechtschaffen.  A manual of standardized terminology, techniques and scoring system for sleep of human subjects , 1968 .

[2]  T M Itil,et al.  Digital computer classifications of EEG sleep stages. , 1969, Electroencephalography and clinical neurophysiology.

[3]  E. Wolpert.  A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. , 1969 .

[4]  G. Clauss,et al.  Grundlagen der Statistik für Psychologen, Pädagogen und Soziologen , 1971 .

[5]  J M Gaillard,et al.  Principles of automatic analysis of sleep records with a hybrid system. , 1973, Computers and biomedical research, an international journal.

[6]  J Hasan,et al.  Differentiation of normal and disturbed sleep by automatic analysis. , 1983, Acta physiologica Scandinavica. Supplementum.

[7]  R. Hoffmann,et al.  Quantitative description of sleep stage electrophysiology using digital period analytic techniques. , 1984, Sleep.

[8]  D. Kupfer,et al.  Application of automated REM and slow wave sleep analysis: I. Normal and depressed subjects , 1984, Psychiatry Research.

[9]  J Hasan,et al.  Automatic analysis of sleep recordings: a critical review. , 1985, Annals of clinical research.

[10]  J. Pilcher,et al.  Automatic analysis overcomes limitations of sleep stage scoring. , 1986, Electroencephalography and clinical neurophysiology.

[11]  H. Kuwahara,et al.  Automatic real-time analysis of human sleep stages by an interval histogram method. , 1988, Electroencephalography and clinical neurophysiology.

[12]  Hans Fleischer,et al.  Grundlagen der Statistik , 1988 .

[13]  B.H. Jansen,et al.  Knowledge-based approach to sleep EEG analysis-a feasibility study , 1989, IEEE Transactions on Biomedical Engineering.

[14]  R Ferri,et al.  Comparison between the results of an automatic and a visual scoring of sleep EEG recordings. , 1989, Sleep.

[15]  T Penzel,et al.  Integrated sleep analysis, with emphasis on automatic methods. , 1991, Epilepsy research. Supplement.

[16]  M Molnár,et al.  Correlation dimension changes of the EEG during the wakefulness-sleep cycle. , 1991, Acta biochimica et biophysica Hungarica.

[17]  J. C. Woestenburg,et al.  Removal of ocular artifacts from the REM sleep EEG. , 1992, Sleep.

[18]  J. Röschke,et al.  A nonlinear approach to brain function: deterministic chaos and sleep EEG. , 1992, Sleep.

[19]  MARC JOBERT,et al.  Wavelets—a new tool in sleep biosignal analysis , 1994, Journal of sleep research.

[20]  A. Gunzinger,et al.  Correlation Dimension of the Human Sleep Electroencephalogram: Cyclic Changes in the Course of the Night , 1994, The European journal of neuroscience.

[21]  J Röschke,et al.  Nonlinear dynamical aspects of the human sleep EEG. , 1994, The International journal of neuroscience.

[22]  W M Herrmann,et al.  On the dimensionality of sleep-EEG data. Using chaos mathematics and a systematic variation of the parameters of the Corex program to determine the correlation exponents of sleep EEG segments. , 1995, Neuropsychobiology.

[23]  J Hasan,et al.  Past and future of computer-assisted sleep analysis and drowsiness assessment. , 1996, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[24]  W. Herrmann,et al.  The future of computer-assisted investigation of the polysomnogram: sleep microstructure. , 1996, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[25]  R Baumgart-Schmitt,et al.  On the use of neural network techniques to analyse sleep EEG data. First communication: application of evolutionary and genetic algorithms to reduce the feature space and to develop classification rules. , 1997, Neuropsychobiology.