Blood flow in the human cerebral cortex: Large-scale pial vascularization and 1D simulation

Understanding cerebral circulation is crucial for early diagnosis and patient-oriented therapies for brain conditions. However, blood flow simulations at the organ scale have been limited. This work introduces a framework for modeling extensive vascular networks in the human cerebral cortex and conducting pulsatile blood flow simulations. Using a patient-specific cerebral geometry, we applied a parallelized adaptive constrained constructive optimization algorithm to create a comprehensive pial vascular network in the left hemisphere, starting from the main cerebral arteries. The resulting network included over 75 000, 103 000, and 55 000 vessels for the anterior, middle, and posterior territories, respectively. Pial vessel diameters featured a median [interquartile range, IQR] value of 62.8 [49.3, 89.3] µm. We integrated the pial vascular network model with the Anatomically-Detailed Arterial Network (ADAN) model to conduct one-dimensional (1D) blood flow simulations under normotensive and hypertensive conditions. Viscoelastic dissipation proved to be a key ingredient in the characterization of the hemodynamic environments in the pial circulation. In the normotensive scenario, mean blood pressure in the pial vessels resulted in a median [IQR] value of 56.7 [49.8, 63.5] mmHg. The flow pulsatility index and its corresponding damping factor were effective descriptors of the hypertensive state. The median [IQR] pulsatility index in the normotensive state was 0.39 [0.38, 0.40], and in hypertension it increased up to 0.84 [0.83, 0.85], while its corresponding damping factor in the normotensive state was 2.07 [1.78, 2.48], and in the hypertensive state it was reduced to 1.20 [1.16, 1.39]. We observed large regional pressure gradients in terminal vessels, with pressure levels ranging from 50 mmHg in normotension to 70 mmHg in hypertension. Additionally, the pulsatility index at terminal vessels increased with distance from the Circle of Willis in the hypertensive case, contrasting with the decreasing pattern seen in normotension. This approach provides a unique characterization of hemodynamics in the pial vascular network of the human cerebral cortex, paving the way for research into microcirculatory environments, the link between hemodynamics and neural function, and their roles in conditions like stroke and dementia.