Universität Tübingen
1H MR Spectrscopy and Chemical Shift Imaging of the In Vivo Brain at 7 Tesla
Abstract
Neurotransmitters and compounds of the brain’s energy metabolism are directly linked to brain function. In vivo magnetic resonance (MR) spectroscopy is a useful method to quantify the neurochemical profiles of localized brain regions noninvasively and to provide valuable information for medical treatment or for the study of underlying mechanisms. The concentrations of substances in the brain detectable by MR, however, are at best in the range of millimols and the sensitivity of MR methods that are based on the Boltzmann spin polarization is inherently low. Most anatomical or functional structures in the brain of humans and monkeys are in the millimeter range or below. The spatial resolution that would be necessary to account for the small target dimensions could not be realized so far in humans and with current MR systems. Reasonable signal strength could only be achieved at the cost of limited spatial specificity. The spectral quality of in vivo MR spectroscopy and the amount of information available from it strongly depend on the existence and strength of potential experimental artifacts like frequency drifts, localization errors or spectral contaminations. In addition, the homogeneity of the magnetic field within the sensitive probe volume is a critical issue. The goal of the thesis was to establish different methods for MR spectroscopy and chemical shift imaging at maximum spatial specificity and spectral quality. The methods were to be adopted for investigations of the non-human primate visual cortex, an area of intense study of the brain’s physiology and function. To maximize the sensitivity and the quality of the MR spectra, a number of methods have been implemented and optimized. Artifact sources were analyzed and methodological corrections were proposed to overcome them. For instance, strong limitations of the MR scanner’s built-in capacity to compensate actively for inhomogeneities of the magnetic field by electrical coils were overcome by newly designed ferromagnetic permalloy assemblies. A combined modular passive (ferromagnetic) and active (electrical) shimming method was developed that provides very strong and highly accurate correction fields for the homogenization of magnetic field distributions in the in vivo brain with reasonably low experimental effort. MR spectroscopy of the macaque primary visual cortex (V1) is challenging, because of its position adjacent to the skull bone and the cortical thickness of only 1.7–2.0 mm. A dedicated 7 Tesla high field MR setup, an anesthetized monkey preparation and optimized MR methods enabled single voxel MR spectroscopy from 40 microliter volumes of V1. Due to the varying magnetic susceptibility conditions around macaque V1 this region is prone for susceptibility induced field distortions. To achieve optimal field homogeneity, field distributions were analyzed, and an appropriate shimming strategy was developed. With the established methods the successful sampling from brain regions entirely confined within V1 gray matter is now possible. Chemical shift imaging methods were implemented and optimized to permit high resolution spatial mapping of metabolite distributions from the macaque visual cortex. The feasibility of chemical shift imaging along planar slices through the brain with a spatial resolution of 1–2 mm has been shown using conventional phase encoding. In addition, Hamming acquisition weighting was provided to improve the reliability of the metabolic mapping based on its favorable imaging properties. Applying the techniques developed in this thesis the achieved resolution limits for 1H MR spectroscopy and chemical shift imaging in monkeys were 2–3 orders of magnitude better compared to previous studies in humans. The spatial specificity now reaches the level of cortical dimensions which is the basis for the investigation of physiology and function in the primate visual system.
Author and committee
dc:creator, dc:contributor.*- Author
-
- Juchem, Christoph
Identifiers
dc:identifier.*- Identifier
- hdl:10900/48984