Functional ultrasound imaging has recently joined the neuroimaging realm providing a brain-wide, high spatiotemporal resolution strategy to monitor local and global brain hemodynamic changes. However, analysis is mostly driven by the local signal change, regional average, or single-voxel extraction while rarely considering cerebrovascular specificities including vessel density and arteriovenous pattern. In this paper, we present the digitization and the skeletonization of the rat brain vasculature based on the signal segmentation of the ultrasound image without the need of contrast agent. We have performed cortex-wide extraction and quantification of vessel features including number, length, and vein/artery discrimination. Then, we evaluated our approach in a pathological context of cortical stroke with which we have identified the ischemic core and quantified an average drop of ~50% in detected vessels. This result is encouraging in that it indicates the feasibility of a strategy for the longitudinal follow-up of patient-specific cerebrovascular neuro-pathologies