Advances in technical computing enable larger and more detailed neural models that can incorporate ever more of the rapidly expanding body of quantitative neuroscience data. In principle, such complex network models that are strongly constrained by experimental data could direct experimental research and provide novel insights into experimental observations. However, as network models grow in complexity and scale, the necessary tasks of development and organization become unwieldy. Further, the models risk becoming inaccessible to experimentalists and other modelers, and their results may then be seen as less relevant to experimental work. To address these obstacles, we developed a tool for managing simulations called SimTracker. It supports users at each step of the modeling process, including execution of large scale parallel models on supercomputers. SimTracker is suitable for users with a range of modeling experience. SimTracker can be a valuable modeling resource that promotes iterative progress between experiment and model.
