Fig. 4

Quantification of capillary width a, Representative images of COL4 immunostained capillaries in the cortex and WM used to determine capillary width. Scale bar represents 25 μm. b, Correlation of mean capillary widths assessed by COL4 and GLUT-1 immunostaining in ageing controls. Pearson’s analysis revealed that mean capillary width in both the WM and the cortex was positively correlated (r = 0.64, P = 0.001). c, Histogram showing mean capillary width in the WM and cortex in controls and dementia groups. In the WM, mean capillary width was consistently larger by 19–45% compared to cortex in all control and dementia groups (^§P < 0.01). In all the dementias, capillary width in the WM was consistently greater by < 20% compared to ageing and young controls (**P < 0.01). In the cortex, mean capillary widths in dementia subjects were not significantly wider compared with ageing and young controls, but only AD subjects showed wider capillaries compared with ageing and young controls (^†P < 0.01)