Identifying fast activation areas during AF is still challenging. Endocardial mapping only can provide sequential measurements that do not capture AF variability and take a long time to record. Recently, ECGI has been proposed as an alternative technology to measure the atrial activation frequency through the instantaneous Fourier transform. This technique allows a fast identification of the fastest activation area (FAA). Yet, artifacts from signal acquisition and the intrinsic variability of AF challenge the accuracy of the FAA location. This study evaluates a methodology to identify recurrent areas with fast activation patterns (RAFA). We compared this technique with the traditional highest dominant frequency (HDF) computed from Welch`s peri-odogram using time variant AF simulations and realistic electrical noise con-ditions. Our findings reveal that RAFA significantly outperforms the traditional HDF area. While the HDF method's sensitivity and specificity fluctuated with signal to noise ratios—showing mean sensitivities of 48.2% ± 44.3% at 20 dBs, 46.8% ± 63. 5% 10 dBs, and 31.1% ± 68.0% 3 dBs —the RAFA ap-proach demonstrated robustness with higher and more consistent mean sen-sitivities of 60.7% ± 33. 2%, 55.0% ± 40. %5, and 51.0% ± 19.5%, with simi-lar sensitivity values. These results demonstrate how RAFA can identify are-as driving AF episodes robustly and more accurately than HDF.