Sudden cardiac death annually claims approximately 300,000 American lives. Its very name underscores the urgent need for an accurate methodology to assess an individual's risk of experiencing unstable cardiac propagation that may result in malignant arrhythmias. Our initial work has identified an inexpensive, sensitive diagnostic procedure to perform that assessment. Our proposed research will further explore the biophysical characteristics of unstable cardiac electrical propagation and suggest criteria useful for non-invasively assessing the risk of ventricular arrhythmias. QT and RR interval spatial and temporal variations are known to be intimately related to the stability of cardiac conduction. Our preliminary theoretical and experimental studies demonstrate that the stability can be assessed by measuring QT and RR intervals collected during an ECG stress test employing a proprietary quasi-stationary exercise protocol (i.e., with gradually changing exercise loads). This is a hypothesis that we propose to test by studying a mutual interdependence of the QT and RR interval fluctuations, which accompany the dynamics of the heart rate gradual trends during the quasi-stationary exercise. We will evaluate quantitative criteria for unstable cardiac conduction that are suggested by the physics and may serve as the basis for a new, non-invasive diagnostic tool for evaluating the risk of cardiac arrhythmias.