This illustration provides a high-level overview of the various steps of our process. Briefly, high-resolution immunofluorescent brain tissue sections stained for β-amyloid, Iba-1, GFAP, and counterstained with DAPI are acquired using a digital whole slide scanner.

We start by measuring the stain density of each channel independently using local thresholding. The DAPI and Iba-1 channels are then used to identify and delineate individual microglia, and a computer vision algorithm is used to classify the activation state of each microglia based on the morphological features of the cell. β-amyloid plaques are segmented from the Amyloid channel. The segmented plaques are expanded iteratively to automatically establish the local and distant plaque microenvironment.

Finally, we analyze the microgliosis and astrogliosis both in the local microenvironment of the plaques and distant to the plaques as a function of disease progression.