Detection of regularities in heart rate variations by linear and non-linear analysis: power spectrum versus...

Regularity or complexity characteristics of RR interval variations have attained importance in the evaluation of a variety of pathological conditions. Regular periodic fluctuations in heart rate are considered indicators of loss of control in the cardiovascular system and of disturbances in pulmonary function. Methods to detect such loss of cardiopulmonary function have been developed mainly through linear and non-linear system analysis. Both approaches were tested in the present study. MBW, which is the ratio between maximal power spectrum (PS) amplitude and its 3-db bandwidth (BW) in the 0-0.1 Hz frequency range, was calculated and compared for different examples derived from 24 h Holter recordings and for sinusoidal function mixed with varying degrees of noise. Approximate entropy (ApEn), a measure derived from non-linear system analysis has been tested for the same data and compared with the MBW ratio. The results show that MBW effectively differentiates regular and non-regular epochs while ApEn differentiates them in a less optimal manner especially in cases where higher degrees of noise are present but regularity is still observed visually.