Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

Methods for extracting self-similarity properties from heart rate variability signal in normal and heart failure patients

Back
Abstract:
I used four categories of methods for extracting self-similarity properties from heart rate variability (HRV) signal, obtaining in general very good results of discrimination of the two populations of 10 normal and 10 heart failure patients. Four classical methods are: the estimation of modulus, variance, periodogram and the method of Higuchi. The parameter estimated is the H exponent of Hurst, that is a global estimator of fractal characteristics of numerical series. The results obtained show that the population of normal subjects has cardiac rhythms more correlated than the population of heart failured: this is explained by the greater control action of the healthy cardiovascular system. Second Order Difference (SOD) Plot is a topological method for the analysis of numerical series. I studied and implemented three methods for extracting dispersion characteristics of the plot: the results on Brownian motions and HRV signals show that this parameter is a good estimator of H exponent. Detrended Fluctuation Analysis (DFA) is a more recent method that allow to better estimate H exponent, with a process of elimination of trends owed to noise. Results are interesting not only for high level of confidency but also for the evidence that the fractal properties can vary, so they are function of time scale. Finally, I developed a parametrical and non-linear filter to analyze fractal characteristics of HRV signal on different time scales. This decomposition confirm DFA results: the intensity of cardiovascular control on cardiac rhythm is different on short and long time scales. This method can be useful to deeper study the functionality of cardiovascular system, in order to make mathematical modelizations for simulations of healthy and pathological patterns.

Type of Seminar:
Public Seminar
Speaker:
Dipl.-Ing. Mario Bellotti
Politecnico di Milano (Italy)
Date/Time:
Sep 05, 2000   17:00
Location:

ETL K 25, ETH Zentrum,Physikstrasse 3, 8006 Zurich
Contact Person:

Dr. Milos Popovic
No downloadable files available.
Biographical Sketch: