Poincaré Plot Image and Rhythm-Specific Atlas for Atrial Bigeminy and Atrial Fibrillation Detection

A detector based only on RR intervals capable of classifying other tachyarrhythmias in addition to atrial fibrillation (AF) could improve cardiac monitoring. In this paper a new classification method 2D non-linear RRI dynamics representation is presented. For aim, the concepts Poincaré Images and Atlases are introduced. Three rhythms were targeted: Normal sinus rhythm (NSR), AF bigeminy (AB). Physionet open source databases used. generated for all signals using different plot configurations: RR, dRR RRdRR. The study was computed time window lengths bin sizes. each rhythm, 80% that rhythm's patients used create reference image, Atlas. remaining 20% as test set classified into one three normalized mutual information correlation. process iterated tenfold cross-validation patient-wise dataset division. Sensitivity results obtained RRdRR configuration size 40 ms, 60 s 94.35%±3.68, 82.07%±9.18 88.86%±12.79 with specificity 85.52%±7.46, 95.91%±3.14, 96.10%±2.25 AF, NSR AB respectively. Results suggest general pattern may be captured these can classify signal segments Images. contrast studies, former generalized more does not depend rhythm-specific thresholds.