Background: Clinically important arrhythmias including atrial fibrillation, and ventricular tachycardia and fibrillation are thought to result from re-entrant excitation. These could arise from alterations not only in the magnitude but also dispersions in the direction taken by the action potential conduction, that could potentially result from anatomical change. Thus, the increased arrhythmogenicity shown by ageing Scn5a+/- hearts correlates with increased fibrosis [1,2]. Similarly, cardiomyocyte-specific deletion of Pak1 (Pak1(cko)) increases the cardiac hypertrophy, also accompanied by tissue fibrosis, during pressure overload [3]. The present report introduces multi-array epicardial measurements of both the magnitude and the dispersion of conduction. Methods and Results: 32 and 64 multi-electrode array (MEA) measurements were made from ventricles of Langendorff-perfused murine Scn5a+/-, Pak1(cko) and WT hearts. These were paced at 8 Hz from the endocardial atrioventricular (AV) node; left ventricular (LV) base, LV apex and the middle of the 64 channel arrays. MEA recordings were obtained from the free walls of the right ventricular ouflow tract (RVOT), right ventricle (RV) or LV. Local activation times (LATs) at each individual MEA electrode were obtained from the timings of the maximal negative slopes in their electrophysiological traces. Conduction properties were then assessed from the LAT measurements by two methods. First, mean conduction velocity (CV) over the entire array was determined. The effective epicardial propagation velocity in m s-1of the activation pattern was obtained from the gradient of its fitted plane. Second, a local vector analysis obtained velocity and conduction direction measurements over short (300 μm) distances. This provided the dispersion of such directions from their standard deviations. In accord with previous reports implicating the RVOT in arrhythmia [4], the RVOT showed the slowest CVs and greatest dispersions in conduction direction in Scn5a+/- compared to WT. Two-way ANOVAs demonstrated that both anatomical region and genotype independently affected CV magnitude whereas anatomical region alone affected their dispersions during pacing at the atrioventricular node. Preliminary experiments in Pak1(cko) mice similarly suggested slower CVs but greater dispersions in conduction direction compared with WT. Conclusions: Our novel quantitation of the magnitude and dispersion in direction of conduction implicates the RVOT in the increased arrhythmogenecity shown by Scn5a+/- hearts in accordance with clinical results. It provides preliminary evidence for similar findings in Pak1(cko) hearts. It may accordingly provide a potentially useful means of assessing for the presence and extent of arrhythmic substrate.