http://sofa2016.org/subm.php Specialized Software System for Heart Rate Variability Analysis: An Implementation of Nonlinear Graphical Methods E. Gospodinova, M. Gospodinov, Nilanjan Dey, I. Domuschiev, Amira S. Ashour, Sanda Balas     Abstract The Heart Rate Variability (HRV) analysis is a non-invasive/effective tool to demonstrate the influence of the autonomic nervous system over the heart rhythm regulation. The current work presents a novel created software system for HRV analysis based on 24-hour Holter ECG signals of group of healthy and unhealthy subjects. The nonlinear analysis of heart intervals was performed with the implementation of original high performance algorithms and software to quantify the heart rate irregularity. The experimental results established that the designed software system for analysis of 24-hour Holter recordings is appropriate for diagnostic, forecast and prevention of the pathological cardiac statuses. The developed and implemented graphical representation and visualization approaches for the results can be stored in specialized data base. Introduction Various cardiovascular diseases are characterized by different graphical images after processing the heartbeat of software programs. Ischemic heart disease, stroke, chronic obstructive lung have remained the top major causes of death. The HRV is known as a diagnostic parameter that defined by the ECG through measuring time intervals between the heartbeats. Methods of HRV analysis are divided in two groups, namely linear and non-linear methods. Linear methods can be used in time- or frequency- domain for HRV analysis. Nevertheless, significant characteristics of the signal dynamics are missed during the use of linear methods. Therefore, recently increased interest is directed toward the non-linear analysis, where the the HRV measurements are non-linear and non-stationary. In addition, a considerable part of information is coded in the dynamics of their fluctuation in different time periods. Consequently, the current work is conducted on selected control groups of patients with cardiovascular diseases, each with 24-hour Holter recordings. Each record contains information of around 100,000 heartbeats. Fluctuations of the physiological signals possess hidden information in the form of self-similarity, scale structure and fractality. Due to the large volume of research information, it is important to correctly determine the trend of the disease in each patient. Such tests are performed periodically to compare the graphic characteristics of images from clinical studies undertaken as a result of treatment and to give an idea for the patient's condition and treatment quality. Nonlinear analysis of cardiology data is a relatively new scientific approach that provides new approach to assess dynamics of heart activity. The main objectives of this work are described, as follows: Development of software for HRV analysis of 24-hour Holter ECG records of healthy and unhealthy subjects. The software offered analysis in time-domain, frequency-domain, nonlinear and wavelet analysis. The nonlinear analysis results are illustrated by applying the Detrended Fluctuation Analysis (DFA), Rescaled adjusted range Statistics plot (R/S) and Poincaré plot. Demonstration of the results’ graphical representation using different approaches on clinical studies of HRV and assessment of health status of patients with cardiovascular disease with multiple and periodic 24-hour Holter studies of treatment. Methodology In the present work, two groups of signals are analysed, namely the RR time series of 16 normal subjects and 16 congestive heart failure (CHF) patients. These signals are consisting of around 100,000 data points corresponding to 24-hour Holter ECG recordings. The proposed software system deployed several methods as follows, where these methods are realized by the Matlab software developed in research project for nonlinear analysis of ECG signals. The DFA (Detrended Fluctuation Analysis) is a technique for detecting correlations in time series [6]. These functions are able to estimate several scaling exponents from the RR time series being analysed. The scaling exponents characterize short or long-term fluctuations. The relationship on a double-log graph between the investigated signal’s. The rescaled range is a statistical measure of the time series variability. The Hurst exponent is one closely associated method with the R/S (Rescaled adjusted range Statistics plot). This exponent is a measure that has been widely used to evaluate the self-similarity and correlation properties of fractional Brownian noise, the time series produced by a fractional Gaussian process. In order to estimate the Hurst exponent, versus in log-log axes is plotted. The slope of the regression line approximates the Hurst exponent, where is the standard deviation estimated from the observed data X(n).Poincaré plot The Poincaré plot analysis is a graphical nonlinear method to assess the dynamic of HRV. This method provides summary information as well as detailed beat-to-beat information on the behaviour of the heart. Consequently, these methods are used to implement a new software tool for the HRV based classification of healthy subjects and CHF patients, where several researchers were interested with the HRV analysis and cardiac diseases. Results & Conclusion Figure 2(a) and (b) illustrate values of scaling exponents and the line slope F(n) on double logarithmic plot obtained by using the DFA method for the investigation of signals. It depicts a significant difference between patients with CHF and healthy controls in short- and long- time scales. Healthy subjects typically show fractal behavior of heartbeat dynamics, while patients with CHF show an alteration in fractal correlation properties. Moreover, Fig.3 (a) and (b) demonstrate the obtained results of the R/S method that applied to the studied signals to determine the value of the Hurst exponent. It establishes that the RR time series are correlated, i.e. they are fractal time series. For normal subject, the value of Hurst exponent is high due to the variation being chaotic, and for CHF patient this value decreases due to the low RR variation. The results of the Poincaré plot analysis of RR time series for healthy subject and CHF patient are presented in Fig.4 (a) and (b). The analysed data from the medical study of patients were combined into two groups of signals and records, namely for 16 CHF (Congestive Heart Failure) patients and 16 normal subjects. The main software menu for each patient is illustrated in Fig. 1. Figure 1 shows the distribution of the QRS complexes, RR intervals and results of time-domain analysis (the values of investigated parameters, RR and HR histograms) in the menu. This page allows the selection of the required analysis type including the time-domain, frequency-domain, nonlinear and wavelet analysis. This work is interested with the nonlinear analysis results. Figure 4 illustrates that the Poincaré plot for healthy subject is a cloud of points in the shape of an ellipse (‘comet’ shaped plot). However, points for the CHF patient are a cloud of points in shape of a circle (‘complex’ shaped plot). The geometry of these plots can be used to distinguish between healthy and unhealthy subjects. The obtained Poincaré plot parameters are directly related to the physiology of the heart. Table 1 reports the investigated parameters values (mean ± standard deviation). It illustrates the significant difference between the different parameter’s values for the corresponding method for the healthy subjects and the CHF patients. It is noticeable that values of all parameters have higher values in the healthy subjects compared to the CHF patients (which have lower values) except in the vase of the ratio. Consequently, the proposed software is significant to distinguish between the two classes of the healthy and CHF patients. References 1. Force, T., Task Force. Heart rate variability. Standards of measurement, physiological interpretation, and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Eur. Heart J, Vol.17, pp.354-381 (1996). 2. Dimitrova, M.; Lahtchev, L.; Lozanova, S.; Roumenin, C. Cloud Computing Approach to Novel Medical Interface Design. In Handbook of Medical and Healthcare Technologies (pp. 245-265). Springer New York (2013). 3. Ernst, G., Heart Rate Variability, Springer-Verlag London (2014). 4. Acharya, U-R.; Suri, J-S.; Spaan, J-E.; Krishnan, S-M., Advances in Cardiac Signal Processing, Springer-Verlag Berlin Heidelberg (2007). 5. Gospodinova, E.; Gospodinov, M.; Dey, N.; Domuschiev, I.; Ashour, A-S.; Sifaki-Pistolla, D. Analysis of Heart Rate Variability by Applying Nonlinear Methods with Different Approaches for 6. Graphical Representation of Results. Analysis, Vol.6, No.8 (2015). Gospodinova, E.; Gospodinov, M.; Dey, N.; Domuschiev, I.; Ashour, A-S.; Sifaki-Pistolla, D. Analysis of Heart Rate Variability by Applying Nonlinear Methods with Different Approaches for Graphical Representation of Results. Analysis, Vol.6, No.8 (2015).