Abstract

Cardiovascular diseases are the leading cause of death in the world, and among them Atrial Fibrillation is the most common cardiac arrhythmia affecting up to 2% of the world population. During the early stage of Atrial Fibrillation, the events appear and disappear randomly which makes the diagnosis challenging. Many approaches are based on the R-R variability to perform the mostion of the signal as Atrial Fibrillation, but it is well known the R-R variability is a common feature in most of the arrhythmias and it can lead to a miss- classifications. Instead, Atrial activity (P wave) contains the most relevant information about the atrial conduction, and therefore approaches based on extracting features from P wave are more reliable in the diagnosis of atrial arrhythmias. In this work, a robust methodology for the detection of the P wave is presented. The approach is based on exploiting spatial and temporal information contained in 12-lead ECG when the atrial and ventricular activity is correlated. The result shows a positive predictive value of 100% in the 70% of the cases and a minimum of 90% for a few cases in normal sinus rhythm signals.

Keywords:Spatio-temporal filtering; Atrial fibrillation; P wave; ECG signal processing; QRST cancellation