Exploiting Spatiotemporal Information for Blind Atrial Activity Extraction in Atrial Arrhythmias

Abstract

AbstractThe analysis and characterization of atrial tachyarrhythmias requires the previous estimation of the atrial activity (AA) free from any ventricular activity and other artefacts. This contribution considers a blind source separation (BSS) model to separate the AA from multilead electrocardiograms (ECGs). Previously proposed BSS methods for AA extraction exploit only the spatial diversity introduced by the multiple electrodes. However, AA typically shows certain degree of temporal correlation, featuring a narrowband spectrum. Taking advantage of this observation, we put forward a novel two-step BSS-based technique which exploits both spatial and temporal information. The spatiotemporal BSS algorithm is validated on real ECGs from a significant number of patients, and proves consistently superior to a spatial-only ICA method. In real ECG recordings, performance can be measured by the main frequency peak and the spectral concentration. The spatiotemporal algorithm outperforms the ICA method, obtaining a spectral concentration of 58.8% and 44.7%, respectively.KeywordsIndependent Component AnalysisBlind Source SeparationAtrial ArrhythmiaAtrial ActivitySpectral ConcentrationThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.