Non-invasive detection of higher frequency atrial sources during atrial fibrillation

This study proposes a non-invasive methodology to detect higher atrial frequencies in AF patients, which may be related to localized drivers, responsible of AF maintenance. The proposed algorithm extracts a signal from a Body Surface Potential Mapping (BSPM) recording from the linear combination that maximizes a cost function that measures periodicity within a given range. By applying iteratively this algorithm in 1.5Hz-width bands from 5Hz to 20Hz in steps of 0.25Hz, several candidates for AF signals with increasing frequency are obtained. Signals with high spectral concentration, low kurtosis and repeatability in 5 consecutive steps were considered compatible with AF. Among these signals, the one with the highest frequency was selected as candidate for AF driver. This frequency was compared with spectral analysis of surface signals in segments under adenosine effects, i.e. with absence of ventricular activity. The algorithm was applied to 18 BSPM recordings from AF patients. In all cases, the high frequency was consistent with the spectral analysis of the segments with adenosine. High atrial frequency values were 9.61 ± 2.12, spectral concentration was 0.415±0.096 and kurtosis 0.60±0.81. The non-invasive detection of high frequency atrial sources may help to define the most appropriate therapy, e.g. for ablation of atrial regions with high activation rates.