Download PDF
Neuropediatrics 1980; 11(4): 303-322
DOI: 10.1055/s-2008-1071399
Original articles

© Georg Thieme Verlag KG Stuttgart · New York

ANSÄTZE ZUR MASCHINELLEN ERKENNUNG DER ALTERSGEMÄSSEN ENTWICKLUNG DES KINDLICHEN EEG IM RAHMEN EINER 5-JAHRES-LÄNGSSCHNITTSTUDIE

AUTOMATIC RECOGNITION OF AGE-SPECIFIC DEVELOPMENT OF THE EEG IN INFANCY AND EARLY CHILDHOOD - A FIVE YEAR FOLLOW-UP STUDYM.  Lechle1 , P.  Michels1 , W. S. Tirsch1 , S. J. Pöppl1 , H.  Mann2 , E.  Müllner2 , H. -M. Weinmann3
  • 1MEDIS-Institut für Medizinische Informatik und Systemforschung der Gesellschaft für Strahlen- und Umweltforschung, Arabellastraße 4/III, D-8000 München 81
  • 2Institut für Soziale Pädiatrie und Jugendmedizin der Universität München, Lindwurmstraße 131, D-8000 München 2
  • 3Kinderklinik der Technischen Universität München, Kölner Platz 1, D-8000 München 40
Further Information

Publication History

Publication Date:
19 March 2008 (online)

 PDF Download  Permissions and Reprints

Abstract

This study was initiated with a view to early detection of dysfunctions of the central nervous system (CNS) and to observe the physiological development of the CNS by means of EEG obtained in the age-groups: new-born, half, 1, 2, 3, 4 and 5 years. This was supported by standardized clinical examinations for neurological and somatic findings and by standardized tests of psychomotoric and intellectual development. In this investigation the data as used for automatic processing consisted of 110 EEGs in the age-groups: half, 1, 2, 3 and 4 years. Bipolar EEGs were recorded using "ten-twenty" method with the following leads: F4-C4, P4-O2, F3-C3 and P3-O1.
For feature extraction three methods of data reduction were applied, i. e. interval-amplitude, spectral analysis and the autoregressive model. From the features thus obtained, age-specific frequency parameters were selected by statistical methods; further evaluation was performed by cluster and discriminant analysis. The former showed an unequivocal case grouping for each age-group. Using unmatched samples of only clinically healthy children from the three age-groups œ, 1 and 2 years, linear discriminant analysis were applied to the parameters of the three methods of data reduction.
This procedure yields in mean recognition rates of 95 %, 98 % resp. 98 % in hold-one-out classification. Similar results are obtained with samples from four or five age groups (œ to 4 years). The results of this type of automatic EEG analysis show that discriminant analysis can be used to allocate EEGs to specific age-groups; hence, it may readily be ascertained whether differences between chronological age and development age as evidenced by the EEG exist.

Key words

Physiological development of the CNS - interval - amplitude - analysis spectral analysis, autoregressive model, Cluster-, discriminantanalysis