Skip to main content
Fig. 3 | Journal of Neuroinflammation

Fig. 3

From: Cyclic multiplex fluorescent immunohistochemistry and machine learning reveal distinct states of astrocytes and microglia in normal aging and Alzheimer’s disease

Fig. 3

Unsupervised spectral clustering of astrocyte profiles reveals three distinct phenotypes. a Heatmap depicts the unsupervised spectral clustering of 5172 ALDH1L1+ astrocyte cell bodies based on their mean gray intensity (MGI) for the other 7 astrocytic markers. Note that MGI z-scores were scaled from 0 to 100 to facilitate comparison across markers. Three clusters are evident, which we termed “homeostatic,” “intermediate,” and “reactive.” b High-plex images from representative astrocytes of each of these clusters. Scale bar: 5 µm. c Stacked bar graphs show the proportions of astrocytic states by diagnosis and the proportions of diagnoses by astrocytic state. d Box and whisker plots illustrate the distribution (box: median and interquartile range [IQR]; whiskers: 1.5 × IQR) of MGI z-scores for each astrocytic marker across the homeostatic, intermediate, and reactive phenotypes. Statistical comparisons between states are available in Table S3, Additional file 1

Back to article page