Prediction of sudden cardiac death based on fundamental changes in nonlinear characteristics of cardiac signals

To quickly detect sudden cardiac death (SCD), it is decisive to gather suitable information and enhance the accuracy of the diagnosis algorithms. Consequently, in the present study, the heart rate variability (HRV) signal of subjects who experience sudden cardiac death (SCD) is studied. We looked at people's heart signals for one hour before something happens to see if there are any noticeable changes. The patients' HRV signals are segregated into ۵-minute parts in the suggested approach. Each section is divided into four shorter signals. Thereupon, the energy and instant amplitude of each sub-signal are examined. The information flows between signal strengths and measuring the complexity of energy sub-signals are checked. A significant change from its former section is identified. A support vector machine classifier benefits from detecting individuals exposed to SCD by considering significant changes as indicators of the SCD process. It can anticipate SCD ۱۵ minutes before it happens. Not restricted to any special subclass of cardiac diseases, this technique has priority. To evaluate the specificity of the algorithm, it has been used not only with patients having SCD but also with individuals who are healthy, as well as those with coronary artery disease (CAD) and congestive heart failure (CHF), analyzing their HRV signals. The specificity values for normal, CHF, and CAD patients are ۱۰۰%, ۹۳.۳%, and ۹۵.۶%, respectively, in the results.

کلیدواژه ها:

Sample entropy ، Teager-Kaiser operator ، Transfer entropy ، Local characteristic decomposition ، Heart rate variability