Atrial Fibrillation detection on electrocardiogram

摘要

The rapid and the irregularity of the electrical activity that characterizes Atrial Fibrillation produce an increase in both variability and complexity of RR intervals series. Based on these two characteristics, we have developed a method combining two nonlinear statistical techniques in order to detect the presence of Atrial Fibrillation. These two nonlinear statistical methods are: Poincare plot which quantifies the variability of RR intervals series and the sample entropy which characterizes the complexity of RR intervals series. We used the MIT-BIH Atrial Fibrillation database to train the algorithm to seek for thresholds and segment length and then we tested it on the MIT-BIH Arrhythmia database. Using these parameters determined by Receiver Operating Characteristic (ROC) curves, we achieved Se=99.65% and Sp=99.33% for the MIT-BIH Atrial Fibrillation database and we obtained Se=97.91% and Sp=92.72% for the MIT-BIH Arrhythmia database. The proposed method is compared to “Dash's” detection one. For the Dash's method, its sensitivity and specificity are less important than ours (Se=98.53% and Sp=95.81% for the MIT-BIH Atrial Fibrillation database. For the MIT-BIH Arrhythmia database, Se=97.43% and Sp=89.86%).