Patient-specific computational fluid dynamic simulation of cerebrospinal fluid flow in the intracranial space

Evaluating amplified magnetic resonance imaging as an input for computational fluid dynamics models of the cerebrospinal fluid

Computational models that accurately capture cerebrospinal fluid (CSF) dynamics are valuable tools to study neurological disorders and optimize clinical treatments. While CSF dynamics interrelate with deformations of the ventricular volumes, these deformations have been simplified and even discarded in computational models because of the lack of detailed measurements. Amplified magnetic resonance imaging (aMRI) enables visualization of these complex deformations, but this technique has not been used for predicting CSF dynamics. To assess the feasibility of using aMRI as an input for computational fluid dynamics (CFD) models of the CSF, we deduced the amplified deformations of the cerebral ventricles from an aMRI dataset and imposed these deformations in our CFD model. Then, we compared the resulting CSF flow rates with those measured in vivo. The aMRI deformations yielded CSF flow following a pulsatile pattern in line with the flow measurements. The CSF flow rates were, however, subject to noise and increased. As a result, scaling of the deformations with a factor 1/8 was necessary to match the measured flow rates. This is the first application of aMRI for modelling CSF flow, and we demonstrate that incorporating non-uniform deformations can contribute to more detailed predictions and advance our understanding of ventricular CSF dynamics.

Regulation of brain fluid volumes and pressures: basic principles, intracranial hypertension, ventriculomegaly and hydrocephalus

The principles of cerebrospinal fluid (CSF) production, circulation and outflow and regulation of fluid volumes and pressures in the normal brain are summarised. Abnormalities in these aspects in intracranial hypertension, ventriculomegaly and hydrocephalus are discussed. The brain parenchyma has a cellular framework with interstitial fluid (ISF) in the intervening spaces. Framework stress and interstitial fluid pressure (ISFP) combined provide the total stress which, after allowing for gravity, normally equals intracerebral pressure (ICP) with gradients of total stress too small to measure. Fluid pressure may differ from ICP in the parenchyma and collapsed subarachnoid spaces when the parenchyma presses against the meninges. Fluid pressure gradients determine fluid movements. In adults, restricting CSF outflow from subarachnoid spaces produces intracranial hypertension which, when CSF volumes change very little, is called idiopathic intracranial hypertension (iIH). Raised ICP in iIH is accompanied by increased venous sinus pressure, though which is cause and which effect is unclear. In infants with growing skulls, restriction in outflow leads to increased head and CSF volumes. In adults, ventriculomegaly can arise due to cerebral atrophy or, in hydrocephalus, to obstructions to intracranial CSF flow. In non-communicating hydrocephalus, flow through or out of the ventricles is somehow obstructed, whereas in communicating hydrocephalus, the obstruction is somewhere between the cisterna magna and cranial sites of outflow. When normal outflow routes are obstructed, continued CSF production in the ventricles may be partially balanced by outflow through the parenchyma via an oedematous periventricular layer and perivascular spaces. In adults, secondary hydrocephalus with raised ICP results from obvious obstructions to flow. By contrast, with the more subtly obstructed flow seen in normal pressure hydrocephalus (NPH), fluid pressure must be reduced elsewhere, e.g. in some subarachnoid spaces. In idiopathic NPH, where ventriculomegaly is accompanied by gait disturbance, dementia and/or urinary incontinence, the functional deficits can sometimes be reversed by shunting or third ventriculostomy. Parenchymal shrinkage is irreversible in late stage hydrocephalus with cellular framework loss but may not occur in early stages, whether by exclusion of fluid or otherwise. Further studies that are needed to explain the development of hydrocephalus are outlined.

Computational Fluid Dynamics of Cerebrospinal Fluid

Cerebrospinal fluid (CSF) plays a critical role in the healthy function of the brain, yet the mechanics of CSF flow remain poorly understood. Computational fluid dynamics is a powerful tool capable of resolving the spatiotemporal evolution of CSF pressures and velocities, but technical and methodological limitations have limited the clinical use of CFD to date. With improvements in medical imaging, computational power, and machine learning, however, CFD may be on the cusp of breaking through into the medical mainstream. In this chapter, we will review the applications of CFD of CSF, present our methodological recommendations for conducting CFD of CSF, present the results of a novel CFD methodology incorporating patient-specific tissue displacements, and discuss the barriers and pathways to clinically useful CFD simulation.

Computational fluid dynamics model to predict the dynamical behavior of the cerebrospinal fluid through implementation of physiological boundary conditions

Cerebrospinal fluid (CSF) dynamics play an important role in maintaining a stable central nervous system environment and are influenced by different physiological processes. Multiple studies have investigated these processes but the impact of each of them on CSF flow is not well understood. A deeper insight into the CSF dynamics and the processes impacting them is crucial to better understand neurological disorders such as hydrocephalus, Chiari malformation, and intracranial hypertension. This study presents a 3D computational fluid dynamics (CFD) model which incorporates physiological processes as boundary conditions. CSF production and pulsatile arterial and venous volume changes are implemented as inlet boundary conditions. At the outlets, 2-element windkessel models are imposed to simulate CSF compliance and absorption. The total compliance is first tuned using a 0D model to obtain physiological pressure pulsations. Then, simulation results are compared with in vivo flow measurements in the spinal subarachnoid space (SAS) and cerebral aqueduct, and intracranial pressure values reported in the literature. Finally, the impact of the distribution of and total compliance on CSF pressures and velocities is evaluated. Without respiration effects, compliance of 0.17 ml/mmHg yielded pressure pulsations with an amplitude of 5 mmHg and an average value within the physiological range of 7-15 mmHg. Also, model flow rates were found to be in good agreement with reported values. However, when adding respiration effects, similar pressure amplitudes required an increase of compliance value to 0.51 ml/mmHg, which is within the range of 0.4-1.2 ml/mmHg measured in vivo. Moreover, altering the distribution of compliance over the four different outlets impacted the local flow, including the flow through the foramen magnum. The contribution of compliance to each outlet was directly proportional to the outflow at that outlet. Meanwhile, the value of total compliance impacted intracranial pressure. In conclusion, a computational model of the CSF has been developed that can simulate CSF pressures and velocities by incorporating boundary conditions based on physiological processes. By tuning these boundary conditions, we were able to obtain CSF pressures and flows within the physiological range.