A mathematical model for data-driven synthesis of neuron morphologies based on random walks

Recent advances in computational resources have enabled the development of large-scale, biophysically detailed brain models, which require numerous three-dimensional neuron morphologies exhibiting realistic cell-to-cell variability. However, the limited availability of experimental reconstructions restricts parameter estimation for many morphology synthesis algorithms, which typically rely on extensive datasets. Here, we propose enhancing our branching-and-annihilating random walk method by incorporating a set of mathematical equations that estimate branching and annihilation probabilities directly from Sholl plots and branch point counts. Because these morphological metrics are commonly reported in the literature, our approach facilitates the generation of realistic three-dimensional morphologies even in the absence of experimental reconstructions.