Techniques for the analysis of long-term cardiorespiratory recordings from infants
The paper describes a series of techniques for the analysis of long-term cardiorespiratory recordings from infants. The analysis performed is based on analysing the respiratory rate, respiratory-rate variability, heart rate and the heart-rate variability measured for nonoverlapping time epochs of approximately 100s duration throughout the recordings. Three descriptors of each of these four data series have been used. First, the mean value and standard deviation; secondly, the probability density function, which was then processed by a principal-components factor analysis to achieve data reduction and classification to make comparisons between different groups of infants; and finally, the power spectral density function. These techniques provide a comprehensive method for analysing long-term cardiorespiratory recordings which should greatly extend their value for both routine clinical use and for research purposes.
KeywordsData classification Heart rate Long-term recordings Respiratory rate Time-series analysis
Unable to display preview. Download preview PDF.
- Bendat, J. S. andPiersol, A. G. (1971)Random data: analysis and measurement. Wiley-Interscience.Google Scholar
- Guilleminault, C. andCoons, S. (1983) Sleep states and maturation of sleep: a comparative study between full term normal controls and near miss SIDS infants. InSudden infant death syndrome.Tildon, J. T., Roeder, L. M. andSteinschneider, A. (Eds.), Academic Press, New York.Google Scholar
- Harper, R. M., Schlabassi, R. J. andEstrin, T. (1974) Time series analysis and sleep research.IEEE Trans.,AC-19, 932–943.Google Scholar
- Morrison, H. R. (1970)Multivariate statistical methods. McGraw Hill, New York.Google Scholar
- Wilson, A. J. andFranks, C. I. (1982) The Sheffield respiration analysis system.IEE Proc. A,129, 702–706.Google Scholar