Human neurons expressing mutations associated with neurodegenerative disease are becoming more widely available. Hence, developing assays capable of accurately detecting changes that occur early in the disease process and identifying therapeutics able to slow these changes should become ever more important. Using automated live-cell imaging, we studied human motor neurons in the process of dying following neurotrophic factor withdrawal. We tracked different neuronal features, including cell body size, neurite length, and number of nodes. In particular, measuring the number of nodes in individual neurons proved to be an accurate predictor of relative health. Importantly, intermediate phenotypes were defined and could be used to distinguish between agents that could fully restore neurons and neurites and those only capable of maintaining neuronal cell bodies. Application of live-cell imaging to disease modeling has the potential to uncover new classes of therapeutic molecules that intervene early in disease progression.