Reconstructing the connectome from an ensemble of measurements

While connectomics paradigms have been undergoing rapid development in the experimental community, the problem of analyzing the resulting data has remained largely unaddressed. Recently, the mesoscale connectome of the mouse was made available from the Allen Brain Institute. This connectome was constructed by way of using enhanced green fluorescent protein (EGFP) expressing adeno-associated viral vectors to discover the connectivity strength between brain areas. Herein, we will attempt to show that the problem of discovering removed entries from connectivity data in a large neural system from an ensemble of such measurements can be formulated naturally in terms of nuclear norm minimization techniques. It is our belief that the presented methods will allow the acquisition of future connectomes with an order of magnitude reduction in experimental effort, as well as significantly outperform the simpler inference techniques used in prior work, and function well with few data observations.