Evaluation of quasi periodicity in cardiac rhythm

Summary form only given. Variability of cardiac activity reflected in electrocardiograms has been of considerable interest and the subject of different studies. In particular, the presented project included development of a scheme for presentation of cardiac activity and a comparative evaluation of heart rate variability. The graphical representation of cardiac rhythm variability is demonstrated in a form of circular displays, which could be easily used in a clinical setting. Specifically, two different approaches were used to implement the desired displays. The technique implemented involves breaking down an ECG signal into particular segments, using segment lengths 300, 400 and 500 data points. The first set of plots represents selected ECG waveforms in segments with each segment defined by the user. The segments are then plotted on a circular path with overlapping segments. This overlapping of segments allows demonstrating variability in a patient´s cardiac activity. Evaluated ECG data was recorded using BioCapture 3.0, system from Cleveland Medical Devices which was interfaced with a PC. In parallel a MATLAB code was developed to load the recorded signals, evaluate numerically cardiac rhythm, and obtain a set of circular plots representing varying levels of cardiac rhythm changes. A separate set of developed circular displays to evaluate additional features of cardiac rhythm variations. This involved plots with some R peaks eliminated from the display. The peak elimination involved R peaks that are within a constant rate category (if duration is within the same time period then they are eliminated). This is achieved by deleting a corresponding block size and corresponding slicing of a section (block) of the RR-interval. The size of the block is determined by the sampling frequency and the mean of the selected RR-intervals. Suppose one has an ECG data of 1000 points, which is divided into 10 blocks with each block having a width of 100 points. The formula for fi- nding the block size: Block size = (mean of RR-intervals) times (Fs). In this case for the data used the mean RR-interval = 0.82, and Fs = 128 Hz. Block size = (0.82) times (128 Hz) = 103.936, rounding it off to 100. The plotting is done using user selectable intervals of variability at 2%, 5%, 10%, 15% and 20%. The parameters calculated in this section are (number of peaks detected, number of peaks skipped, number of peaks remaining, means of RR intervals, and the number of possible segments). A combination of circular displays and corresponding computation of listed characteristics compare favorably with other approaches to cardiac rhythm evaluation including Welch´s methods of analysis and Poincare plot, and provides significantly improved documentation of cardiac rhythm changes.