The present study describes methodological aspects of image analysis for angiographic image data with long-term two-photon microscopy acquired for the investigation of dynamic changes in the three-dimensional (3D) network structure of the capillaries (less than 8 μm in diameter) in the mouse cerebral cortex. Volume images of the identical capillaries over different periods of days up to 32 days were compared for adaptation under either chronic hypoxia (8-9% O2) or hyperoxia (40-50% O2). We observed that the median diameters of measured capillaries were 5.8, 8.4, 9.0, and 8.4 μm at 0, 1, 2, and 3 weeks during exposure to hypoxia, respectively (N = 1, n = 2193 pairs at day 0), and 5.4, 5.7, 5.4, 6.0, and 6.1 μm measured weekly up to 32 days under hyperoxia (N = 1, n = 1025 pairs at day 0). In accordance with these changes in capillary diameters, tissue space was also observed to change in a depth-dependent manner under hypoxia, but not hyperoxia. The present methods provide us with a method to quantitatively determine three-dimensional vascular and tissue morphology with the aid of a computer-assisted graphical user interface, which facilitates morphometric analysis of the cerebral microvasculature and its correlation with the adaptation of brain cells imaged simultaneously with the microvasculature.
Keywords: 3D Image analysis; Animal experiments; Brain microcirculation; Machine learning; Optical imaging.