Combined Nonlinear Analysis of Atrial and Ventricular Series for Automated Screening of Atrial Fibrillation

Abstract

Atrial fibrillation (AF) is the most common cardiac arrhythmia in clinical practice. It often starts with asymptomatic and short episodes, which are difficult to detect without the assistance of automatic monitoring tools. The vast majority of methods proposed for this purpose are based on quantifying the irregular ventricular response (i.e., RR series) during the arrhythmia. However, although AF totally alters the atrial activity (AA) reflected on the electrocardiogram (ECG), replacing stable P-waves by chaotic and time-variant fibrillatory waves, this information has still not been explored for automated screening of AF. Hence, a pioneering AF detector based on quantifying the variability over time of the AA morphological pattern is here proposed. Results from two public reference databases have proven that the proposed method outperforms current state-of-the-art algorithms, reporting accuracy higher than 90%. A less false positive rate in the presence of other arrhythmias different from AF was also noticed. Finally, the combination of this algorithm with the classical analysis of RR series variability also yielded a promising trade-off between AF accuracy and detection delay. Indeed, this combination provided similar accuracy than RR-based methods, but with a significantly shorter delay of 10 beats.

Detection of Brief Episodes of Atrial Fibrillation Based on Electrocardiomatrix and Convolutional Neural Network

“…Using windows of seven beats from 20 records of the AFDB, the authors reported Se = 98.09% and Sp = 91.66%. On the contrary, Ródenas et al ( 2015 ) and Ródenas et al ( 2017 ) assessed the presence of the P-wave by measuring the wavelet entropy (WE) from the median TQ interval and the variability of the TQ interval series, respectively. For the former, the WE of 10 beats was used to make the detection with an average delay of 5 beats, resulting in Acc = 95.28%.…”

“…One problem when investigating P-wave morphology is dealing with noisy records. In fact, for the studies from Ladavich and Ghoraani ( 2015 ), Ródenas et al ( 2015 ), and Ródenas et al ( 2017 ) manual revision of the data was needed before running the experiments.…”

“…For instance, de Carvalho et al ( 2012 ) applied the spectral entropy for determining f wave presence; Babaeizadeh et al ( 2009 ) investigated the absence of P waves by measuring the similarity between the samples in two consecutive PR intervals. Conversely, Ladavich and Ghoraani ( 2015 ), Ródenas et al ( 2015 ), and Ródenas et al ( 2017 ) only explored whether P waves are absent and left out all rhythm information. However, the detectors accounting for morphological information have not been completely capable of characterizing the atrial activity, which is commonly obscured by noise, leading to similar performances as those accounting for rhythm (Sörnmo, 2018 ).…”

A Deep Learning Approach for Featureless Robust Quality Assessment of Intermittent Atrial Fibrillation Recordings from Portable and Wearable Devices

“…In the present study, 50 signals were discarded because they did not present any data. The remaining 250 ECGs were processed with an automatic algorithm to detect AF episodes within the clean ECG intervals [ 45 ] and later revised by an expert physician. Finally, the PDB comprises 36 single-lead ECG recordings lasting between 1 and 2 h from patients with intermittent AF which were recorded using a Nuubo™ textile wearable Holter.…”

“…The signals were recorded at a sampling frequency of 250 Hz and 12 bits of resolution over a dynamic range of ±5 mV. As before, an automatic algorithm was used to detect AF episodes [ 45 ], and two experts independently visualized the recordings to identify sufficiently clean and noisy intervals. ECG excerpts were labelled as noisy when either physician was unable to confidently distinguish all R-peaks.…”