Unsupervised inference of arbor morphology progression for microglia from confocal microscope images

Microglia are Central Nervous System (CNS) cells that are similar to macrophages. They undergo characteristic changes in cell arbor morphology in response to tissue perturbation. Ensembles of microglia exhibit a progression of arbor morphologies. Our goal is to discover these progressions and determine the underlying arbor features from 3-D multi-channel fluorescence confocal microscope images of rat brain tissue multiplex stained with Hoechst to reveal cell nuclei, and immunolabeled for IBA-1 to reveal microglia. The microglia are automatically traced, and a set of 131 arbor features are computed. An agglomerative clustering algorithm based on Pearson´s correlation is used to derive coherent modules of features. A k-NNG structural similarity analysis of feature modules enables us to construct a global similarity matrix, from which a global multi-level k-NNG is constructed to derive an interactive progression chart through a modified Fruchterman-Reingold algorithm that clearly reveals a progression from highly ramified microglia to round cells proximal to the injury site of an implanted neural recording device.