Introduction: Elucidation of the mechanism of atrial fibrillation (AF) will lead to improved catheter ablation outcomes. However, the electrical substrates on which AF drivers such as meandering rotors and multiple wavelets are likely to appear are still unclear. Methods: Twenty-nine patients (age 68±10 years, eight females) with persistent AF were included in this study. Real-time phase mapping (ExTRa mapping) was performed in the left atrium during AF. After that, electroanatomical mapping such as voltage and activation mapping with HD-grid catheter was performed after conversion to sinus rhythm with cardioversion. According to the nonpassively activated ratio (%NP: the ratio of the form of rotors and/or multiple wavelets to the recording time), the recorded three highest %NP (Active-area) and three lowest %NP areas (Passive-area) were collected in each subject. Electrical property and overlapping with the low-voltage zone (LVZ, voltage amplitude <0.5mV) or mid-voltage zone (MVZ, 0.5-1.0mV) within these areas were assessed and compared between the Active- and the Passive-areas. Results: We analyzed a total of 87 Active- and 84 Passive-areas and the mean %NP were 61±8% and 13±8%, respectively. Signals with ≥4 fractionated components (fractionated electrograms) were seen significantly more in the Active-areas than in the Passive-areas (53% vs. 26%, respectively; p<0.001, Figure A ). However, there was no difference in the conduction velocity (0.70±0.34m/s vs. 0.73±0.30m/s; p=0.577), mean voltage amplitude (1.78±1.54mV vs. 1.71±1.53mV; p=0.794) and coinciding with LVA/MVA (43% vs. 41%; p=0.785) between the Active- and the Passive-areas ( Figure B-D ). Conclusion: AF drivers tended to appear in the area with fractionated electrograms. High %NP areas partly overlapped, but not always coincided with LVA/MVA, as much as low %NP areas.