Glial morphology is closely linked to cellular function, providing important insights into physiological and pathological processes in the central nervous system. Despite this, quantitative analysis of glial morphology remains challenging due to the time-intensive nature of manual segmentation, variability between annotators, the need for coding expertise, and reliance on costly proprietary software. 

We developed Microcount, a fully integrated, open-source MATLAB-based pipeline for automated segmentation and morphological analysis of glial cells in microscopy images. Microcount enables atlas-based anatomical region-of-interest (ROI) selection using the Allen Mouse Brain Atlas with a custom segmentation strategy that enables physiologically accurate reconstruction of glial somas and processes in two-dimensional images. The pipeline provides quantitative measures of cell density, soma size, branching complexity, convexity, Sholl index, and co-marker co-localisation. Importantly, it is implemented through an intuitive user interface (UI), which removes the need for coding expertise.

We validated Microcount using mouse and human datasets spanning different glial populations, staining methods, and imaging modalities, including fluorescent and DAB-labelled tissue stained for ionised calcium-binding adaptor molecule 1 (Iba1), cluster of differentiation 68 (CD68), glial fibrillary acidic protein (GFAP), S100 calcium-binding protein B (S100B), and purinergic receptor P2Y12 (P2RY12). We also performed a comparison with manual segmentation on 16 fluorescence microscopy images (1000 × 1000 pixels) segmented independently by five trained annotators. Microcount’s segmentations showed strong agreement with consensus manual annotations (mean F1 score = 0.68, accuracy = 0.87). Furthermore, morphological and activation-related measurements generated by Microcount correlated well with those generated from manual segmentations (Pearson’s r range = 0.72-0.98, p < 0.05), whilst requiring orders of magnitude less time to compute. Validation of Microcount in a chronic mouse inflammation model using adult C57BL/6 mice treated with lipopolysaccharide (0.75 mg/kg, intraperitoneal injection, once daily for 7 days; n = 4 per group) detected significant increases in glial density, soma size, and CD68 expression (two-way ANOVA with Šidák’s correction, p < 0.05) compared to saline control. Together, these findings establish Microcount as a robust, accessible, and scalable tool for high-throughput analysis of glial morphology and marker expression in health and disease.