Background: Atrial fibrillation (AF) is associated with diminished cardiac functions. This study describes a novel methodology for non-invasive measurement of left atrial (LA) function in patients with AF. Methods: A total of 55 patients (81% male, 55% paroxysmal, 60.4±10.2 years) underwent cardiac MRI prior to AF ablation. Invasive LA pressure was measured during sinus rhythm following trans-septal puncture for AF ablation. LA pressure and volume (MRI derived) loops were prepared for all patients. LA diastolic function was defined as the ratio of the change in LA pressure to the change in LA volume during passive LA filling. Additionally, the ratio of pulmonary vein flow velocities during ventricular systole and diastole were obtained using phase contrast MRI (Figure top panel, S/D ratio). Patients with ventricular diastolic dysfunction (as seen by E/A ratio) were excluded. LA systolic functions were calculated from the active emptying fraction during atrial contraction. Diastolic dysfunction scores greater than the 90th %tile (>1.6mmHg/ml) were considered to signify severe diastolic dysfunction. Results: The mean diastolic dysfunction score was 0.76±0.7 mmHg/ml. The score was higher in patients with persistent versus paroxysmal AF (1.0±0.9 vs. 0.60±0.5,p=0.02). The mean S/D ratio was 1.04±0.6; and was lower in patients with persistent versus paroxysmal AF (0.9±0.5 vs. 1.2±0.7, p=0.04). There was a negative linear association between diastolic dysfunction scores and S/D ratio measures (Figure bottom panel, p <0.001). In a multivariate analysis adjusting for age, LA systolic function and CHADS2 score, the severe diastolic dysfunction was independently associated with lower S/D ratio (p-0.03). Conclusion: The non-invasive MRI derived ratio of pulmonary vein flow velocities during systole and diastole are associated with the MRI/invasive derived diastolic dysfunction score, and appear to be a fair surrogate of LA diastolic function.