Ultra High Field Magnetic Resonance

Exploring the BOLD-Effect at Ultra-High Field

Our group focuses mainly on the development of MR techniques for observing the way the brain works at ultra high field. For this aim, we investigate the characteristics and advantages of ultra high magnetic fields for brain research, develop new techniques to improve the information that can be gained with MRI, and implement novel, multimodal approaches to better understand the BOLD effect. With a 9.4 T MR scanner for humans and a 14.1 T instrument for small animals, our institute has a unique instrumentation that makes it a pioneer in the investigation of ultrahigh field imaging for brain research. The main goals of our group are to develop techniques to make it possible to fully take advantage of the high magnetic field.

Human venogram acquired at 9.4 T (left) and image of the rat brain from 16.4 T.

The high SNR [1] and enhanced contrast [2], as well as the novel types of information that can be obtained at ultra-high field allows for novel insights into the operation of the brain. However, this requires a deeper understanding of the way the functional MR signal arises and of the interaction of physiological processes that are involved in brain function. We aim to develop novel techniques for both animal and human experiments that aid in understanding the dynamics of brain activity and to improve the information that can be gained in neuroscientific studies.

To take advantage of the ultra-high field for neuroscientific studies, we develop techniques to improve the significance of the data. This involves at first to optimize the data acquisition to obtain high-quality, undistorted functional and anatomical data with highest spatial resolution. Among other things, this includes implementing motion correction techniques, SNR efficient acquisition schemes and advanced reconstruction algorithms. For a deeper understanding of the data, multimodal experiments on rats are designed that use invasive techniques for a more detailed and quantitative insight into the physiology of brain activity.

We aim to develop the technology that make it possible to fully take advantage of the expected gain at ultra high field:

  • Evaluation of the potential advantages and limitations imposed by using high magnetic fields,
  • Development of optimized measuring techniques with the aim of taking full advantage of the improved imaging characteristics,
  • Combine MRI with other technologies to enhance the significance of neuroscientific studies.
  • Support of other groups within the department by developing and optimizing imaging techniques adapted to their applications.

The main projects currently are: 

Ultra-high resolution anatomical and functional imaging in the human brain at 9.4 T

In neuroscientific studies on humans, new analysis techniques in combination with modern ultra-high field MRI data allow for novel insights into brain function and connectivity [3]. This requires anatomical and functional imaging data with high spatial and temporal resolution. However, motion, SNR limitations and distortions can affect the quality of high resolution images. By implementing a camera-based real-time motion correction system in our sequences, using an SNR-optimized, k-space weighted acquisition scheme for the anatomical, and a zoomed and distortion-corrected technique for functional images, combined with an SNR-optimized reconstruction scheme, we try to optimize the quality of images with very high resolutions. Using the prospective motion correction system, we were able to obtain T2*-weighted images with a spatial resolution of 0.108 × 0.134 × 0.523 mm3 (7.6 nl voxel size, based on PSF width) with sufficient SNR within the acceptable scan time of 37 min. T1-weighted images using our own MP2RAGE sequence reached voxel sizes of (0.35 mm)3. Distortion-free, zoomed functional EPI images with spatial resolutions down to (0.65 mm)3 show great detail and clear functional activation (Fig. 1). These techniques represent the basis to potentially detect cortical subunits such as cortical layers or columns. Investigations using 31P-spectroscopy and perfusion measurements will help us to gain further information on the physiology of the brain.

Functional Imaging

In functional imaging, isotropic spatial resolutions of 1 mm for humans and 150 μm in rats were reached with high signal variations due to brain activation. Using a special EPI sequence with correction of image distortions due to field inhomogeneities, it was possible to accurately overlay phase information, which distinctly shows the positions of veins, and functional activation maps with the aim of comparing the main BOLD-signal contributions with the venous architecture. For fMRI based on gradient-echo EPI, it could be shown that the major signal contributions arise from the draining veins as expected from theoretical predictions. It will have to be shown whether the expectation of improved weighting of the small vasculature in comparison to larger veins for spin-echo based imaging realizes the promises for improved spatial specificity at high field strengths.


In addition to the high sensitivity, in vivo MR spectroscopy profits from the increased spectral dispersion, causing the expectation of an especially high gain at ultra high field. Thus, proton spectroscopy in the rat brain at 16.4 T was implemented and optimized acquisition and quantification techniques were developed. A localized STEAM sequence with short echo times and improved water suppression was established and a novel, improved algorithm to take the signals of macromolecules into account was developed. With these techniques, the high sensitivity and spectral dispersion made it possible to quantify up to 20 metabolites, including glycine and acetate (Fig. 3). Additional studies verified the high precision and accuracy of these measurements even in short experiments or from difficult deep brain regions, like the medulla oblongata or the cerebellum [3]. Applications in different rat strains or brain regions showed the capability of detecting even small variations in the concentration of single metabolites [4].

Perfusion measurements

Quantifying the perfusion in the brain or other organs without the need to use contrast agent is challenging due to the low amplitude of the signal obtained with Arterial Spin Labeling techniques. We concentrate on increasing the sensitivity of ASL measurements by developing and applying highly sensitive techniques like continuous ASL or a specially developed technique for quantifying low perfusion, as in muscle or in white matter brain tissue. In addition, using ultra-high field ASL will yield improved perfusion images due to the high intrinsic sensitivity at high field and the increased longitudinal relaxation time.

Go to Editor View