Biologically informed simulation of perivascular glymphatic transport within cerebral porous tissue

Sammanfattning

Background and Aim: The glymphatic system plays a critical role in clearing metabolic waste from the brain via perivascular pathways. This degree project aims to develop and compare three biologically informed computational models of perivascular glymphatic transport within cerebral porous tissue. A specific objective is to evaluate whether an advection-only, purely convective mechanism can reproduce the early non-linear depth-dependent tracer distribution patterns observed in human in-vivo MRI studies following intrathecal gadolinium (Gd) injections. Methods: Three computational models with varying degrees of geometric complexity were constructed using MATLAB and COMSOL Multiphysics: (1) a simplified 2D baseline model, (2) a mathematically defined, high-resolution 3D vascular network based on physiological bifurcation rules, and (3) a geometrically simplified 3D network that preserves macroscopic topology. Glymphatic transport was simulated over a 24-hour period under the assumption of bulk-flow-dominated movement (Péclet number approaching infinity). Results: While all three models consistently underestimated the absolute tracer concentrations in deeper cerebral layers, their spatial distributions differed significantly. The simplified 2D representation (Model 1) exhibited an unrealistic linear decay profile. In contrast, the biologically informed 3D models (Models 2 and 3) successfully replicated the qualitative, non-linear spatial decay of Gd perfusion observed clinically. Model 3 emerged as the optimal compromise, preserving the physiological 3D transport pattern while reducing computational solver time by a factor of five compared to Model 2. Conclusions: This study serves as a vital proof-of-concept, demonstrating that fluid flow routed through complex, realistic macroscopic topologies is inherently capable of generating spatial patterns that appear diffusion-like, even without reliance on molecular diffusion. The findings highlight the fundamental role of 3D vascular architecture in shaping cerebral fluid dynamics.