RamiGlyph Captures Microglial Morphological Diversity and Predicts Functional States in Ischemia Reperfusion and Amyloid Pathology

Abstract

Microglia are resident immune cells of the central nervous system, whose ramified processes rapidly remodel in response to injury. However, how to capture subtle morphological changes and whether functional state can be predicted from morphology remain open questions. Here, we present RamiGlyph, a contrastive learning framework integrating topological and structural features, trained on more than 20,000 reconstructed microglia. RamiGlyph not only distinguishes physiological and pathological states of microglia but also generalizes to neuronal cell type classification. Projection of microglia morphological embeddings revealed a continuum rather than discrete classes, from which a morphology score was derived to quantify dynamic process remodeling. To link morphology with function, Gromov Wasserstein optimal transport was used to align unpaired morphological and functional data across stages of ischemia reperfusion injury and amyloid pathology. These alignment results enable prediction of microglial functional states using RamiGlyph embeddings alone, with prediction reliability increasing upon cell aggregation. In summary, RamiGlyph provides a robust framework for resolving continuous microglial morphological variation and linking morphology to functional states across acute and chronic neuropathological contexts.