Model-based approaches to characterize cerebral arterial stiffness and CSF transport with MRI

Cerebral small vessel disease (cSVD) is prevalent in the aging population and is believed to be an important contributor to cognitive decline, dementia, and stroke. The underlying mechanisms of cSVD remain largely unknown but are potentially linked to cerebral arterial stiffening. With age and vascular risk factors, the arteries lose their elasticity, facilitating transmission of pulsatile blood flow to the brain which potentially harms the microvasculature through processes involving blood-brain barrier (BBB) disruption. However, the association between cerebral arterial stiffness and cSVD is understudied, likely due to the lack of measurement techniques. Another potential pathway through which brain health can be affected in aging is via its waste clearance system. It entails flow of cerebrospinal fluid (CSF) from the subarachnoid space (SAS) through the brain via perivascular pathways, enabling clearance of interstitial solutes along the way. Here the CSF production, as well as the cyclic motion of the arterial walls are thought to drive the fluid flow. In line with this, studies have demonstrated that injected contrast agents propagate along the major cerebral arteries, although the separate contributions from diffusion and bulk flow are still to be determined. The aim of this thesis was to propose methods to assess key parameters believed to influence brain health in the ageing population, specifically stiffness of, and CSF transport along, the major cerebral arteries. Furthermore, the aim was to employ the proposed techniques in relevant cohorts to study physiological and pathological processes. Using whole-brain 4D flow MRI and leveraging the stiffness-dependent time-delays between blood flow waveforms sampled at increasing depths in the cerebral arterial tree, allowed the quantification of a global cerebral pulse wave velocity (gcPWV). We demonstrated that challenges introduced by low temporal resolution could be handled by utilizing the vast number of potential measurement points along the extent of the cerebrovascular tree (Paper I). We also showed that gcPWV did not critically depend on the included vascular depth (Paper II), or the inclusion of specific arterial branches, and that it demonstrated robustness to large reductions in the amount of input data, as well as the expected sensitivity to age (Paper I). In a population-based cohort, higher gcPWV was associated with white matter hyperintensity (WMH) volume, the most frequently recognized fe