Linear and nonlinear methods in analysis of sleep stages with respect to heart rate

Abstract

Insomnia is a health disorder affecting a large fraction of population. For the diagnosis of sleep related disorders and diseases it is important to establish the sleep structure of patient. Contemporary diagnostics is based on polysomnography, however this methods is extremely time-consuming and expensive. Results of recent investigations show dependence between some characteristics of heart rate and sleep stages. The procedure of recording of heart rate is comparatively cheap. However, the computer aided conclusions about sleep stages based on analysis of the heart rate still not sufficiently reliable. In the present paper several methods of analysis of heart rate are considered to assess their relevance to the classification of sleep stages. Most intriguing question is on combination of nonlinear and standard methods, e.g. Fourier transforms, for the analysis of the heart rate variability. Some recommendations concerning such combinations are presented