Automated Detection of Paroxysmal Atrial Fibrillation Using an Information-Based Similarity Approach

Abstract

Atrial fibrillation (AF) is an abnormal rhythm of the heart, which can increase heart-related complications. Paroxysmal AF episodes occur intermittently with varying duration. Human-based diagnosis of paroxysmal AF with a longer-term electrocardiogram recording is time-consuming. Here we present a fully automated ensemble model for AF episode detection based on RR-interval time series, applying a novel approach of information-based similarity analysis and ensemble scheme. By mapping RR-interval time series to binary symbolic sequences and comparing the rank-frequency patterns of m-bit words, the dissimilarity between AF and normal sinus rhythms (NSR) were quantified. To achieve high detection specificity and sensitivity, and low variance, a weighted variation of bagging with multiple AF and NSR templates was applied. By performing dissimilarity comparisons between unknown RR-interval time series and multiple templates, paroxysmal AF episodes were detected. Based on our results, optimal AF detection parameters are symbolic word length m = 9 and observation window n = 150, achieving 97.04% sensitivity, 97.96% specificity, and 97.78% overall accuracy. Sensitivity, specificity, and overall accuracy vary little despite changes in m and n parameters. This study provides quantitative information to enhance the categorization of AF and normal cardiac rhythms.

Detection of Atrial Fibrillation Episodes in Long-Term Heart Rhythm Signals Using a Support Vector Machine

“…In [37] the changes in RR duration during the sequence have been represented as the binary words, where value of 1 corresponds to increase of interval duration, and 0 means no change or a decrease. Then, the testing segment is classified by comparing its information-based dissimilarity index with those obtained for the templates of AF episode and normal sinus rhythm.…”

“…It can also include normalized RR intervals [29,30] or normalized RR differences [31], Shannon entropy [19] or coefficient of sample entropy [15,32]. Other form to present the RR irregularity are: the density histogram of the difference between successive RR intervals [33,34], map that plots RR intervals versus change of RR intervals [35], mapping the RR-interval time series to binary symbolic sequences [36,37] or Markov score of RR interval [16]. form to present the RR irregularity are: the density histogram of the difference between successive RR intervals [33,34], map that plots RR intervals versus change of RR intervals [35], mapping the RR-interval time series to binary symbolic sequences [36,37] or Markov score of RR interval [16].…”

“…Other form to present the RR irregularity are: the density histogram of the difference between successive RR intervals [33,34], map that plots RR intervals versus change of RR intervals [35], mapping the RR-interval time series to binary symbolic sequences [36,37] or Markov score of RR interval [16]. form to present the RR irregularity are: the density histogram of the difference between successive RR intervals [33,34], map that plots RR intervals versus change of RR intervals [35], mapping the RR-interval time series to binary symbolic sequences [36,37] or Markov score of RR interval [16]. Two HR signals expressed in beats per minute (bpm) with clinically recognized AF segments of different characteristics of HR changes in relation to normal sinus rhythm (non-AF).…”

“…Since there is no standard for the window length, many works have aimed to find the optimal length, providing the best classification performance. Some works assumed that AF episodes of less than 30 s duration are not clinically significant, which led to higher optimal number of heartbeats: 100 [33], 128 [35] and 150 [37]. Other authors claim that longer windows tend to miss short AF episodes, and thus they applied significantly shorter windows: 30 [32], 12 [15] or even 8 beats [44].…”

ECGNET: Learning Where to Attend for Detection of Atrial Fibrillation with Deep Visual Attention

“…One of the other key issues in AF detection methods is their poor performance in detecting AF episodes in short signal recordings (i.e., less than 30-s). While the majority of the stateof-the-art algorithms require a 30-s episode or at least 127/128 beats to achieve an acceptable detection performance [26], [27], [22], [28], our proposed method offers great performance on very short ECG segments of size 5-s which inhere are less than a 7-beats window.…”