Into the gray matter: construction and coupling of multiscale highly-detailed arterial networks in the human cerebral cortex

Stroke-related diseases represent 11% of worldwide deaths. State-of-the-art medical imaging describes many aspects of cerebral circulation. Nevertheless, the limited resolution of these techniques hinders the study of the hemodynamics in smaller vessels. Computational models aid in this problem, providing tools to characterize cerebral hemodynamics in clinical scenarios of interest that are otherwise inaccessible. This work aims to build vascular networks in the human cerebral cortex to study hemodynamics across different regions of the gray matter. The method is based on the Constrained Constructive Optimization method (CCO), called PDCCO, which generates vascular networks following a set of anatomical rules. A patient-specific MRI geometry of the cerebral gray matter is used to register an existing vascular model of the major cerebral vessels in the pial surface of the brain, from where the network is expanded. The brain geometry is partitioned into three territories for each large cerebral artery. The vascular generation is conducted independently for each territory and merged into a single network. First, a thin volume of the pial surface is filled with blood vessels down to the scale of penetrating arterioles of 50 μm. The network is further extended downwards to penetrate the gray matter, and deep vascularization is achieved by appending sub-tree networks to each terminal vessel. These sub-trees are generated separately with vascular properties associated with the gray matter. The final network reaches 234000 vascular segments in the pial network, out of which 116977 are terminal vessels. In turn, each terminal vessel gives rise to a network of 50 terminals, yielding 5 million terminals, and 10 million vascular segments, for each hemisphere. Over the pial surface, the diameters vary between 2100 μm and 26 μm, with terminals between 50 μm and 60 μm, and the pressure drops from 100 mmHg to a range between 50 mmHg and 70 mmHg. The network can be coupled with existing blood flow models of the entire cardiovascular system for the simulation of pulsatile blood flow to study pressure heterogeneities along the cerebral cortex in normal and pathological systemic conditions. This novel approach proposes a strategy for the automatic vascularization of the brain, with the aim of understanding microcirculation under different conditions, allowing the study of the risk of stroke, among other mechanisms involved in the onset and progress of degenerative diseases.