PROJECT SUMMARY/ABSTRACT Application of multiscale computer modeling to help guide and elucidate heart disease treatments is emerging. Computational modeling, however, has not been exploited for optimizing cardiac resynchronization therapy (CRT). While CRT has emerged as a powerful treatment for heart failure (HF) to restore normal activation pattern in the heart, about 30% of patients still do not improve after therapy (non-responders). Improvement of responder rate therefore remains a crucial clinical challenge and the holy grail of CRT. We believe that computational modeling can help optimize CRT and improve the responder rate. Equally important, the development of a multiscale computational framework that considers the key physics of the heart can help understand several novel pacing therapies (e.g., conduction system pacing (CSP) including HIS bundle pacing and left branch bundle (LBB) pacing) that have been developed recently to improve the responder rate. Specifically, computational modeling can help elucidate the key factors affecting the long and short-term effectiveness of these pacing therapies in patients with different intraventricular conduction delay and/or LV scar/ischemia. Here, the overall goal here is to develop computational approaches that combine machine learning algorithms and physics-based modeling to fundamentally understand the short and long-term effects of CRT that includes CSP, optimize CRT, and to elucidate the advantages and disadvantages of CSP over standard CRT. The following specific aims are constructed to accomplish this goal. First, we will develop an experimentally-validated multiscale cardiac electro-mechanics-perfusion (EMP) computational framework to simulate the chronic effects of CRT and CSP in treating mechanical dyssynchrony in LBBB + ischemia. Second, we will integrate the computational modeling framework with efficient machine learning and optimization algorithms to optimize CRT with LV epicardial and endocardial pacing in ischemia. Third, we will use the validated multiscale computational EMP framework to elucidate the effects and factors affecting the response of CSP in ischemia. The proposed approach and methodologies are innovative. More importantly, successful completion will directly translate the findings to the clinic for optimization of CRT therapy to reduce non-responder rates as well as patient identification for different pacing therapies. This would have substantial impact on improving the treatment and reducing the cost of HF epidemic.