Symbolic Recurrence Analysis of RR Interval to Detect Atrial Fibrillation.

Jesús Pérez-Valero,M Victoria Caballero Pintado,Joan Garcia-Haro,Arcadio García Alberola,Paul Curtin,Francisco García Córdoba,Antonio-Javier García-Sánchez,Francisco Melgarejo,Mariano Matilla-García,Eduardo Pinar,Manish Arora,José A García Córdoba,Manuel Ruiz Marín

doi:10.3390/jcm8111840

Abstract

Atrial fibrillation (AF) is a sustained cardiac arrhythmia associated with stroke, heart failure, and related health conditions. Though easily diagnosed upon presentation in a clinical setting, the transient and/or intermittent emergence of AF episodes present diagnostic and clinical monitoring challenges that would ideally be met with automated ambulatory monitoring and detection. Current approaches to address these needs, commonly available both in smartphone applications and dedicated technologies, combine electrocardiogram (ECG) sensors with predictive algorithms to detect AF. These methods typically require extensive preprocessing, preliminary signal analysis, and the integration of a wide and complex array of features for the detection of AF events, and are consequently vulnerable to over-fitting. In this paper, we introduce the application of symbolic recurrence quantification analysis (SRQA) for the study of ECG signals and detection of AF events, which requires minimal pre-processing and allows the construction of highly accurate predictive algorithms from relatively few features. In addition, this approach is robust against commonly-encountered signal processing challenges that are expected in ambulatory monitoring contexts, including noisy and non-stationary data. We demonstrate the application of this method to yield a highly accurate predictive algorithm, which at optimal threshold values is 97.9% sensitive, 97.6% specific, and 97.7% accurate in classifying AF signals. To confirm the robust generalizability of this approach, we further evaluated its performance in the implementation of a 10-fold cross-validation paradigm, yielding 97.4% accuracy. In sum, these findings emphasize the robust utility of SRQA for the analysis of ECG signals and detection of AF. To the best of our knowledge, the proposed model is the first to incorporate symbolic analysis for AF beat detection.

Highlights

Atrial fibrillation (AF) is the most common sustained arrhythmia associated with a high risk of an ischemic stroke [1]
We introduce the application of symbolic recurrence quantification analysis (SRQA) for the study of ECG signals and detection of AF events, which requires minimal pre-processing and allows the construction of highly accurate predictive algorithms from relatively few features
The organization of vertical structures in the SRPs, indicative of periods of persistent symbolic dynamics, were not useful predictors of AF. These findings indicate that the separation of AF signals from NS signals is driven by diverging sequential dynamics and symbolic diversity rather than differences in persistent stable states

Summary

Introduction

Atrial fibrillation (AF) is the most common sustained arrhythmia associated with a high risk of an ischemic stroke [1]. For this reason, timely diagnosis of arrhythmia is of crucial importance. The diagnosis of AF can be difficult since patients may be asymptomatic, in cases of intermittent AF episodes that end spontaneously (paroxysmal atrial fibrillation, PAF). Given the significant risk of mortality and morbidity, and the fact that asymptomatic AF is not detected unless looked for, there is a strong impetus for ambulatory monitoring. With the greater need for ambulatory monitoring, accurate and automated detection of asymptomatic AF becomes a relevant task. AF is characterized in the ECG by loss of the normal atrial depolarization waves (P waves), rapid and irregular atrial fibrillatory waves with an undulating baseline and typically irregular RR intervals (time between two consecutive ventricular beats) [2]

Results

Discussion

Conclusion