Abstract
Alzheimer's disease is a progressive degenerative condition that has various levels of effect on one's memory. It is thought to be caused by a buildup of protein in small fluid-filled spaces in the brain called perivascular spaces (PVS). The PVS often takes on the form of an annular region around arteries and is used as a protein-clearing system for the brain. To analyze the modes of mass transfer in the PVS, a digitized scan of a mouse brain PVS segment was meshed and used for computational fluid dynamics (CFD) studies. Tandem analyses were then carried out and compared between the mouse PVS section and a cylinder with commensurate dimensionless parameters and hydraulic resistance. The geometry pair was used to first validate the CFD model and then assess mass transfer in various advection states: no flow, constant flow, sinusoidal flow, sinusoidal flow with zero net solvent flux, and an anatomically correct asymmetrical periodic flow. Two mass transfer situations were considered, one being a protein build-up and the other being a protein blend-down. The results showed that, for all the flows in question and both protein scenarios, there was nearly no protein clearance benefit by the anatomically correct PVS geometry. Surprisingly, solute blend-down was diffusion-dominated even for high Peclet numbers. Findings herein lead to the conclusion that minimal risk would be incurred by incorporating PVS geometry into existing reduce-order modeling arterial networks.