Application of the Stationary Wavelet Transform to Reduce Power-line Interference in Atrial Electrograms

Abstract

The atrial electrical activity analysis is nowadays playing a key role to improve current knowledge about the mechanisms triggering and maintaining atrial arrhythmias, such as atrial fibrillation (AF). Indeed, the intra-atrial electrograms (EGMs) provide essential information to guide the prevailing cornerstone treatment of AF, i.e., catheter ablation. Bearing in mind that cardiac electro-physiology laboratories are plenty of technology, such a kind of operating rooms are a very adverse environment when it comes to avoid power-line interference in EGM recordings. However, the reduction ofthis nuisance signal deserves more attention. Thus, aimed at reducing powerline noise but, at the same time, preserving the original EGM morphology, a novel algorithm based on the stationary Wavelet transform (SWT) is proposed. To validate the method, 150 bipolar EGMs with 10 seconds in length have been synthesized with different noise levels. The original and denoised EGMs have then been compared by means of an adaptive signed correlation index (ASCI), computed both in time and frequency domains. The obtained results have shown improvements between 9% and 17% for the proposed method regarding a reference notch filtering. Consequently, this new algorithm may enable more reliable and truthful further analyses of atrial EGMs.