Abstract Background An assessment of positive outcome probability of ablation therapy based on the comprehensive signal complexity analysis is a promising working hypothesis while electrocardiographic imaging (ECGI) can detect and visualize zones of phase singularities (PS) associated with stable sources of atrial fibrillation (AF). Methods Ten consecutive patients with persistent AF (three female, median (min–max) – 63.5 (45–75) years) underwent ECGI using “Amycard 01C EP lab” system with cardiac MRI (1.5-T Magnetom Avanto) followed by pulmonary vein isolation. Each T-Q segment with a length >800 ms during AF was processed to find PS. Sites with rotations around stable pivot points were considered as PS and then marked and visualized on the reconstructed anatomical 3D atrial model. Finally, a signal complexity cluster analysis was performed to define and depict phase-aggregation zones. Results ECGI analysis identified a total number of 410 PS, with 196 (47.8%) occurring in the LA and 214 (52.2%) in the RA. The median (25–75% IQR) number of revealed PS per patient was n=20 (14–30) for RA and n=20 (11–22) for the LA. The majority of the PS in the LA was located on the inferior wall n=66 (min-max 1–17). In eight patients, comprehensive signal complexity analysis revealed stability of phase-clustered zones over time. The mean number (min-max) of PS in a clustered area was 10 (6–15). In two patients, PS were distributed disordered on the entire LA and RA surface. Distribution of phase singularities Conclusions This is the first clinical study demonstrating signal complexity analysis capability of clustering noninvasively mapped PS and relating them to specific atrial anatomical regions. Thereby obtained clusters may be a potential zones of conduction block, and could contribute to a better understanding of the temporal AF complexity.