Breaking technical barriers of MRI and MRS to explore human brain functions

Abstract

Our brain plays an important role in our everyday life. Nearly all of our common activities require our brain. Even though brain research has advanced tremendously over the recent years, there are still unanswered questions. These are both fundamental questions about our consciousness and how our brain works, as well as clinical questions in case of neurological diseases like Alzheimer’s and Parkinson’s disease. Magnetic resonance imaging (MRI) has proven itself to be a valuable instrument to study the human brain, allowing structures and anatomy to be visualized in-vivo at high detail. However, in order to gain more knowledge about the mechanics of the brain, static imaging of anatomy is not enough. To take the next step, it is essential to study dynamic in-vivo brain functions, including (neuronally triggered) hemodynamics and metabolism. This thesis focuses on two MR techniques that measure brain activity via changes in hemodynamics and metabolism, specifically blood-oxygen-level dependent functional MRI (BOLD fMRI) and magnetic resonance spectroscopy (MRS). In general, MRI measurements are a tradeoff between the signal-to-noise ratio (SNR), resolution, (dynamic) scan time and imaging volume. With MRI it is therefore possible to acquire high resolution images, but at the cost of a long acquisition time. This is not desirable when aiming to monitor fast and detailed physiological processes. For example for BOLD fMRI, the combination of both a high temporal resolution (< 1 sec) and a high spatial resolution (< 1 mm) is rarely seen. For edited MRS, the commonly reported (dynamic) scan times to measure the main inhibitory neurotransmitter γ-aminobutyric acid (GABA), range between 6 – 30 minutes. This is long when looking at the time scale of brain functions. Therefore, the aim of this thesis is to maximize the SNR, temporal resolution and spatial resolution of BOLD fMRI and MRS techniques to be able to gain further insight into the hemodynamics and metabolism of the human brain. To achieve this aim five approaches (Chapters 2-6) are pursued. First, the dynamic scan time of BOLD fMRI scans was reduced (Chapter 2) by a combination of high density receive arrays, and an advanced under sampling strategy for 3D fMRI scans (2D CAIPIRINHA for 3D EPI scans). Second, simulations were performed (Chapter 3) showing the potential of a 256 channel receive array for whole brain imaging to reduce dynamic scan time of fMRI scans by an order of magnitude further. Third, the idea of ultrasonic switching gradients was presented (Chapter 4), which might hold even higher accelerations for the future, as described in a patent. Fourth, the neurotransmitter GABA could be measured dynamically with GABA edited MRS techniques that have a reduced scan time of only 1:23 min (Chapter 5) by using a specially built half volume coil setup. Fifth, 31P MRS was found as potential new technique to explore brain function further (Chapter 6). This work brings the ability to follow human brain processes in-vivo with high spatial and temporal resolution a step closer. Fundaments upon which both engineering, neuroscience and clinical studies can further build.