Inter-cellular electrical coupling in the heart is a major determinant of excitation patterns in health and disease. Cardiac arrhythmias, in particular rapid arrhythmias such as tachycardia and fibrillation, are characterised by spatially complex and heterogeneous conduction patterns. In disease conditions, these complex excitation patterns may result from electrophysiological and structural remodelling, for example, remodelling of inter-cellular gap junctions and/or the proliferation of fibrosis. Evaluation of the precise mechanisms by which local tissue structure determines global arrhythmic excitation patterns is a major challenge, yet may be critically important for the development of effective treatment strategies. Computational modelling is a viable tool to address this challenge, but the established approaches for organ-scale simulations are unsuitable to capture the impact of local conduction heterogeneities. In this study, we present a novel network model of inter-cellular coupling for anisotropic organ-scale simulation of electrical dynamics that enables cellular connections to be heterogeneously interrupted. The presented approach is both simple and powerful. Intercellular coupling strength is weighted based on local myocyte orientation and can be easily modified by sampling from continuous distributions and/or removing individual cellular connections entirely; both approaches can be differentially applied to axial and transverse cellular connections for full control over the conduction substrate. Preliminary simulations demonstrate the value of such an approach and indicate that perturbing cellular connections in this manner can facilitate lower global conduction velocities and capture wave breakdown and the development of re-entry in a way that is not possible with the previously established approaches. Therefore, we have presented a useful model to simulate the impact of local conduction heterogeneity in organ-scale models that expands the scope and possibilities for computational models to elucidate arrhythmia mechanisms and generate genuine subject-specificity, which may be particularly relevant in disease conditions.