Research / Cerebrovascular and Brain Mechanics / 02 / 02

Cerebrovascular Networks and Collateral Circulation

The cerebrovascular system distributes blood through a hierarchical network that extends from major cerebral arteries to arterioles and capillaries. Alternative collateral pathways can redistribute flow when a primary arterial route is compromised. We develop image-based and synthetic vascular-network models to quantify how vascular connectivity, geometry, and collateral configuration affect cerebral pressure, flow, and tissue perfusion. These models support controlled evaluation of anatomical variability and compensatory flow under stenotic and ischemic conditions.

The Circle of Willis opening out along the major arteries into synthetic arteriolar trees, a stenosis on one artery, the territory behind it losing perfusion, and the collateral pathways that carry flow back to it
Vascular connectivity, the alternative pathways it leaves open, and the redistribution of perfusion that follows. Vessel diameters are enlarged for visualization.

Image-Based and Synthetic Cerebrovascular Networks

Major cerebral arteries can be reconstructed from clinical vascular imaging, but smaller vessels remain below the available image resolution. We extend the image-derived arteries with synthetic vascular trees that represent the unresolved arteriolar network. The tree-generation algorithm accounts for local tissue demand, proximity between vessels and supplied tissue, and the mechanical cost of maintaining blood flow.

In the tissue-growth-based approach, vascular and tissue domains develop concurrently. An increase in the supplied tissue volume and its metabolic demand triggers the addition and adaptation of vascular branches. The resulting network connects nearby tissue regions while satisfying prescribed flow demand and minimum-work constraints.

MRA → segmented major arteries → tissue domain → synthetic tree growth → the complete multiscale network, labelled as computational network generation
Computational network generation, not biological growth: branches are added where tissue demand, proximity and the mechanical cost of carrying flow call for them.

Hemodynamic Coupling across Vascular Scales

We solve one-dimensional blood-flow equations in the explicitly represented arteries, arterioles, and collateral vessels and couple the terminal branches with a continuum representation of tissue perfusion. Terminal vascular pressures provide boundary information to the tissue domain, while tissue flow demand is transferred back to the vascular network. The iterative coupling maintains consistency between network hemodynamics and regional perfusion.

The explicit vascular network coloured by pressure, the terminal vessel–tissue coupling, and the tissue perfusion field — with arterial pressure and branch flow distribution beneath
Terminal vessel pressures inform the tissue domain and tissue demand returns to the network, iterated until the two agree.

Circle of Willis Anatomy and Flow Redistribution

The Circle of Willis provides major communicating pathways among the anterior and posterior cerebral circulations. Its anatomy varies substantially among individuals, and incomplete or asymmetric configurations can alter the redistribution of flow when cerebral inflow is reduced. We construct patient-specific three-dimensional models with physiologically informed boundary conditions to examine how anatomical variants and autoregulatory responses affect cerebral pressure and flow.

Current research evaluates the coupled effects of Circle of Willis anatomy, arterial stenosis, and autoregulatory capacity on patient-specific cerebral hemodynamics. The objective is to determine how vascular connectivity modifies pressure and flow redistribution under impaired inflow conditions.

A complete Circle of Willis beside an incomplete or asymmetric one, each marked with inlet flow, communicating arteries, pressure distribution, flow direction and the vessel with reduced inflow — labelled as a study configuration
Study configuration for the ongoing analysis: the same reduced-inflow condition imposed on complete and incomplete Circle of Willis anatomies.

Computational Modeling of Collateral Circulation

Collateral vessels provide alternative connections between vascular territories and may redirect blood toward regions affected by arterial stenosis or occlusion. We generate collateral pathways between adequately supplied and hypoperfused branches and systematically vary their caliber, density, and spatial distribution. This approach allows the effects of individual collateral-network properties to be evaluated under controlled conditions.

Four panels on one colour scale — no collateral pathways, small-caliber collaterals, increased collateral density, larger-caliber collaterals — beside a sketch contrasting few large vessels with many small ones
Collateral pathways generated between adequately supplied and hypoperfused branches, with caliber, density and connection location varied one at a time.

Effects of Collateral Caliber and Density

Under the investigated conditions, collateral caliber has a stronger influence on perfusion recovery than collateral number alone. Small-caliber collateral networks provide limited additional flow even when their density is increased, whereas larger vessels form lower-resistance pathways and produce a greater hemodynamic effect. The relative contribution of collateral circulation also increases as primary arterial stenosis becomes more severe.

Recovered perfusion against collateral radius ratio, with collateral density distinguished by line colour and stenosis severity separating the panels
Collateral effectiveness depends more strongly on vessel caliber than on vessel number alone, under the stenosis severities and network configurations analysed here.

Network Response in Ischemic Tissue

We combine subject-specific arterial geometry, synthetic downstream networks, collateral pathways, and tissue-perfusion models to examine blood-flow redistribution under ischemic conditions. Comparisons among baseline, stenotic, collateral, and vasodilated configurations allow the individual and combined contributions of structural and active compensatory mechanisms to be assessed. The arteriolar side of that response is treated in Cerebral Perfusion and Autoregulation; here the emphasis is on network connectivity and collateral morphology.

Baseline, then stenosis, then an alternative collateral configuration, then the recomputed pressure and perfusion — presented as a comparison of configurations rather than as vessels growing in
Baseline, stenotic, collateral and vasodilated configurations compared, so that structural and active compensation can be separated.

Model Evaluation and Scope

We evaluate synthetic vascular morphology through comparisons with high-resolution imaging and anatomical measurements. Simulated flow and perfusion patterns are compared with physiological ranges, literature data, and arterial spin-labeling MRI where available. These comparisons assess the plausibility of the reconstructed networks but do not establish direct clinical diagnostic accuracy.

The current collateral model represents intraparenchymal arterial and arteriolar networks. Pial arterial networks, pial-to-pial collaterals, and collateral pathways involving the external carotid circulation are not explicitly represented.

Selected work

2023

Tissue-growth based synthetic tree generation and perfusion simulation

Biomechanics and Modeling in Mechanobiology 22:1095–1112

To be submitted

Computational analysis of cerebral hemodynamics: effects of autoregulation and Circle of Willis anatomy

Computers in Biology and Medicine

Cerebrovascular and Brain Mechanics 2 topics in this area
Cerebral Perfusion and AutoregulationCerebrovascular Networks and Collateral Circulation