Fig. 6

Iron heat maps can be used to predict GFAP-positive cell heat maps. A Example iron heat map for a section from the temporal lobe. B Estimated GFAP heat map based on the best coefficient combination for the iron densities in rings extending out from each pixel. C Actual GFAP heat map of the same section. D Schematic illustrating the 500 μm-thick rings extending outward from each iron pixel. E Residuals (comparing models with different numbers of rings surrounding the pixel) and F coefficients (from the model with the inner pixel pixel and all three rings) from the predicted GFAP heat maps that best matched the real heat maps. Each point represents the mean of the residuals or coefficients for sections in a brain. Repeated measures ANOVA test, n = 19 cases. Post-hoc pairwise comparisons were performed using Tukey’s multiple comparisons test. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. E The density of GFAP-positive cells in a pixel is influenced by the density of iron deposits up to 1000 μm away (p = 0.0003). F The predictive effect of iron deposit density on GFAP-positive cell density decreases with distance (p < 0.0001)