New Feature Set for Better Representation of Dynamic of RR Intervals in Poincare Plot ?

Abstract

Traditional features extracted from Poincare plot (e.g., SD1 and SD2) IntroductionHeart Rate Variability (HRV) is the variation in the time series of RR intervals [1] and is an indicator of heart's condition [2]. It has been proved that nonlinear analysis of HRV provides more valuable information for the physiological interpretation of heart rate fluctuations compared to linear HRV measures [3]. One of the useful nonlinear analysis of HRV is Poincare plot, plot of each RR interval against next RR interval [4]. Woo et al. were the first group who used Poincare plot for evaluating the differences between healthy subjects and heart failure patients [5]. Poincare plot is a geometrical representation of RR time series to demonstrate patterns of heart rate dynamics resulting from nonlinear processes [6]. Tulppo et al. [7] fitted an ellipse to the point distribution in the Poincare plot and defined two standard descriptors (SD1 and SD2) for quantification of the Poincare plot geometry. These standard descriptors represent the minor axis and the major axis of the ellipse and guide the visual inspection of the distribution. Brennan et al. proved that although Poincare plot is a nonlinear representation of RR intervals, but SD1 and SD2 cannot describe these nonlinear behavior and are linear and statistic parameters [8]. Furthermore, these descriptors ignore temporal information and only quantify point distribution [9].In addition to geometric features that was proposed by our team previously [10-13], a new feature set is proposed in this article to capture the temporal information in a Poincare plot by considering the angle between consecutive points, the direction of the trajectory, and the position of points in relation to the line of identity. These features were used to create ADP Map (angle, direction, and position map) for a new representation of RR points.Performance of proposed feature set was evaluated in distinguishing four cardiac condition including normal sinus rhythm (NSR), acute myocardial infarction (MI), congestive heart failure (CHF), and atrial fibrillation (AF). 2.Method and Data MethodFor extraction of the new feature set and creating new map, quantification of temporal information/dynamics of the points in Poincare plot were the focus of this article. For capturing dynamic information, three features were calculated between every three consecutive points in Poincare plot: angle, direction of the trajectory, and the location of middle point in relation to the line of identity. For visualization purpose, these features were used to create a 3D map (ADP map).