Quantitative MRI of physiological brain tissue deformation in humans

Abstract

The objective of the work presented in this thesis was to develop non-invasive imaging techniques of brain pulsations in response to the beating heart and respiration, in order to pave the way to non-invasive, in-vivo assessment of the impact of disease and physiological stressors on the properties of the brain tissue and microvasculature. Physiological brain tissue deformation is driven, among others, by variations in blood pressure induced by cardiac and respiratory cycles. The pulsations of brain tissue form a valuable source of information. The tissue deformation is not only driven by variations in blood volume from the microvasculature, but also reflects differences in tissue properties like stiffness.

In this work, we started from two MRI motion mapping methods that provide these motion field maps. These MRI methods deploy the MRI phase signal to encode the respective velocity or displacement into the MRI signal. In Chapter 2 we compared the performance of PC-MRI with DENSE through computer simulations and found that DENSE outperforms PC-MRI for small deformations in the human brain tissue, such as induced by the heartbeat.

Cardiac and respiration-induced contributions to brain tissue deformation were disentangled in Chapter 3. Yet, the single-shot approach limited the acquired volume to 2D images only. We acquired two orthogonal slices and performed a 3D analysis of tissue strain along the intersection line. Here we observed, for the first time, the Poisson effect reflected in the tissue deformation, where longitudinal tissue stretch was accompanied by transverse shrinkage of tissue. Furthermore, the results showed that cardiac-induced tissue deformation is dominating respiration contributions by approximately a factor of five.

In Chapter 4, we extended the single-shot DENSE method by combining the approach with a simultaneous multi-slice (SMS) acquisition. The work was the first to report the full cardiac-induced strain tensor of brain tissue deformation with complete brain coverage, for which we found well-defined strain patterns that are consistent between subjects. We called this novel approach Strain Tensor Imaging (STI).

The potential of the STI technique to detect abnormalities in disease, was explored in a case study patient that was treated with a craniectomy. At the time of the MRI acquisitions, the cranial opening – 12 cm in diameter – had not yet been closed by a reconstructed skull part. We compared the strain maps from the patient with the strain maps obtained in healthy subjects, and showed distinct differences between these maps in Chapter 5. This ‘Angelo Mosso experiment in modern days’ shows that the MRI technique is sensitive enough to detect abnormalities in brain tissue deformation.

The DENSE sequence can simultaneously provide both strain data and diffusion data in the brain. We used this property in Chapter 6 to investigate to what extent observed ADC variations in the brain over the cardiac cycle can be explained by measurement errors induced by variations in tissue strain. We found that observed ADC variations are at least a factor of 2 larger than could be explained by variations in the tissue strain.