The management of cardiac arrhythmia remains the largest problem in cardiac electrophysiology. The prevalence of the most frequent arrhythmia, atrial fibrillation (AF), is expected to rise steeply due to the ageing population. In spite of intensive research, the mechanism of atrial fibrillation remains unclear, leading to poor results in its treatment. Ablation of AF often results in complex atrial tachycardia (AT), which are difficult to treat. Also ventricular tachycardias (VT) and fibrillations (VF) are a major cause of sudden cardiac death. Again, eliminating  VTs with ablation has achieved only modest success in complex cases. Therefore, there is an urgent need to better understand and localize the sources of arrhythmia in order to improve its treatment. I propose a radical new approach of applying network theory to study the mechanisms of AT, VT and AF. Currently, network theory is known for being the basis for the Google search engine other online social networks, and has myriad applications throughout biology, physics, and social sciences. However, it has never been applied to the heart. In this proposal, based on my invention and preliminary work, I propose to apply network theory to clinical data of cardiac arrhythmia, backed-up by in-silico simulations. A new set of research tools will be created to automatically detect the source of the arrhythmia for complex AT and AF, which will identify possible ablation targets. For VT a substrate analysis is proposed, in order to reveal the structure of the heart to also determine the ablation target. My preliminary results already show that network analysis is able to automatically predict sites of ablation, prior to surgery in AT, largely exceeding the most recent technologies currently used in clinics. Therefore, this translational project will not only provide novel insights into the mechanism of cardiac arrhythmia, but will actually lead to an improved treatment for the patient.