Congestive heart failure (CHF) is a chronic progressive condition impacting the ability of the heart muscles to pump blood properly. With the growing complexities in lifestyle and unhealthy food habits, susceptibility to this disease is increasing across the population and this is one of the major cardiovascular disorders1. While hypertension and coronary artery diseases remain the main factors leading to congestive heart failure, other factors like obesity and diabetes add up as potential causes of CHF. Heart beat variablity (RR Interval) time series belonging to two groups, diseased (congestive heart failure) and normal subjects have been analyzed with the fractals. The increase of chaotic nature as captured from the analysis is, in general, a sign of a better and stable functioning of the human heart. The overall results show significant quantitative differences between the diseased subjects and the normal subjects as well as different stages of the disease with the normal subjects always having a higher value than the diseased subjects. Decrease in the values is an indicator of dysfunction of the heart and the extent of deviation is an indicator of the degree of dysfunction. The analysis leads to formulation of biomarkers which may be used to detect abnormalities of the human heart sufficiently ahead of the time before it turns out to be fatal. The data when split into periods of around 1 hour and analyzed separately also shows the same consistent behaviour. Eventually mobile apps may be prepared to compute the bio markers on a real time basis. The app will read data from an implanted loop recorder, compute and compare it with a benchmark preloaded for a healthy heart. Whenever the computed value overshoots/ falls below the preloaded value for healthy subjects by a pre-defined tolerance value, some audio visual markers will alert the individual of the impending cardiac disorder.

Study of RR interval time series for Congestive Heart Failure had been an area of study with different methods including non-linear methods. In this article the cardiac dynamics of heart beat are explored in the light of complex network analysis, viz. visibility graph method. Heart beat (RR Interval) time series data taken from Physionet database [46, 47] belonging to two groups of subjects, diseased (congestive heart failure) (29 in number) and normal (54 in number) are analyzed with the technique. The overall results show that a quantitative parameter can significantly differentiate between the diseased subjects and the normal subjects as well as different stages of the disease. Further, the data when split into periods of around 1 hour each and analyzed separately, also shows the same consistent differences. This quantitative parameter obtained using the visibility graph analysis thereby can be used as a potential bio-marker as well as a subsequent alarm generation mechanism for predicting the onset of Congestive Heart Failure.

Authors: Anirban Bhaduri, Susmita Bhaduri and Dipak Ghosh
Physica A: Statistical Mechanics and its Applications (Probable Impact Factor: 2.12)
DOI: 10.1016/j.physa.2017.04.091
Physica A: Statistical Mechanics and its Applications Volume 482, 15 September 2017, Pages 786-795
Status :Published
Area : Biomedical Data analysis