Research topics
Areas of Research
Development of New Imaging Methods
MRI is also a key tool for research in biomedical science. In neuroscience applications, functional and diffusion weighted MR imaging and MR spectroscopy provide new insight into the living human brain since the function of brain parts and their mutual interplay, their connections via nerve fibres, and even the underlying metabolic processes can be investigated. The MR techniques can contribute to a better understanding of structure and function of the brain, in the healtyh as well as in the pathological state as for example in neurological or psychiatric disorders.
Echo planar imaging (EPI) is one of the most widely used techniques in brain imaging. It allows very short measurement times (less than 200 ms for a slice image) at high signal-to-noise ratio. However, it is susceptible to image artefacts such as regions of signal attenuation and geometric distortions (see Figs. 1 and 2). For this reason imaging some parts of the brain is virtually impossible using EPI. In this situation one can restrict the imaged area to a smaller region by the application of so-called two-dimensional excitation pulses. This reduces the image artefacts significantly (Fig. 1). As an alternative new techniques like line scan imaging were developed which can also produce full images of a slice without being prone to geometric distortions. Although line scan images require long acquisition times (approximately 400 to 500 ms) and exhibit a lower signal-to-noise ratio, they yield a geometrically exact mapping of the brain.
Fig. 1: Slice image of a healthy volunteer (left, acquisition time approximately 2 min), echo planar image (middle, approximately 180 ms) with strong distortions in the frontal and medial brain parts, and echo planar image of a small field-of-view defined by a two-dimensional excitation pulse (right, approximately 130 ms). The arrow points to a region containing the amygdala.
Fig. 2: Slice image of a healthy volunteer (left, acquisition time approximately 2 min), echo planar image (middle, approx. 180 ms) with strong distortions in the frontal part of the brain and in the vicinity of the eyes, and line scan image (right, approx. 380 ms).
Functional Brain Imaging
Functional imaging of the brain has become a major tool for brain research. The technique which is most widely used for this purpose is based upon the fact that the activation of a brain area leads to a local increase in blood supply which occurs a few seconds after the start of the activation. This increase causes a localized change in the oxygen content of the blood, which can be detected as a change in image intensity when using an appropriate imaging method. This phenomenon is called blood oxygenation-level dependent (BOLD) contrast. As the related changes are very small (a few percent) and can easily be masked by signal intensity changes of other origin (e.g. due to motion), often a reliable conclusion on the state of activation can only be drawn by repeating activation and measurement a number of times, and by comparison with images which were acquired during a control condition, i.e. without activation of the relevant brain region.
Fig. 3 provides an example for functional brain imaging. It shows the upper part of the brain of a healthy volunteer. The brighter regions mainly represent cerebrospinal fluid, the grey areas contain what is known as the grey matter. The white matter appears dark on these images. The motor cortex can be seen approximately in the centre of the image. This grey matter structure is related to motion of the right hand. During the 8-min scan which comprised 480 images of the motor cortex the volunteer moved the fingers of the right hand every 72 s for a duration of 24 s. A computer program then searched within the set of acquired images for brain areas where the image intensity alternated along with the alternating sequence of rest and motion. In the figure, these areas are marked with colours. They cover the region which is related to the motion of the right hand. In the right part of Fig. 3 the signal time course in the coloured image region is shown: it is clearly visible that the image intensity is increased during the periods of motion (marked by the boxes at the bottom of the figure) and drops off in the periods in between.
Fig. 3: Image of the brain of a healthy volunteer (left). The coloured region marks the brain part where the image intensity followed the alternating sequence of rest and motion (right). It covers accurately the area known to be linked to right hand motion. The remaining, uncoloured regions in the image did not exhibit a significant correlation between motion and image intensity.
One of the main problems of this technique is its susceptibility to image artefacts, such as signal attenuation in regions close to the sinuses. These mean that investigating certain brain regions faces great difficulty or is even impossible. In addition, the method is a rather indirect technique because it does not detect the electrical neuronal activity itself but rather its effects on the oxygen content of the blood. The coupling between eletrical activity and blood oxygen concentration can, however, be impaired by disease or by the influence of medication, such that brain activity might not cause a signal change any more. At present, it is a topic of intensive research how the image artefacts in functional imaging can be reduced in order to broaden the range of successful application. On the other hand, new approaches are being followed to achieve a more direct method or at least an alternative method for mapping human brain activity, e.g. by exploiting the effects of neuronal currents on the MR signal.
Diffusion Weighted Imaging of the Central Nervous System
By means of additional magnetic fields, which are also used to achiev spatial resolution in MR imaging, it is possible to investigate the diffusion motion of water molecules in living tissue. This already plays a key role in clinical diagnostics of stroke: on quantitative maps of the diffusion coefficient the area affected by a stroke can be detected earlier and more accurately than with conventional methods. On the basis of such measurements the physicians can decide on the initiation of thrombolytic therapy (dissolution of blood clots by medication) at an early stage. The method is thought to rely on the phenomenon that the cells in the brain tissue start swelling within minutes after disruption of blood supply. The cell swelling reduces the gaps between the cells such that the diffusing water molecules in the extracellular space face more obstacles during their motion around the cells, leading to a reduction in the measured diffusion coefficient.
For a number of years researchers have been trying to retrieve more information on the structure of white matter than is available from conventional imaging methods.
Diffusion Tensor Imaging (DTI) relies on the dependence of the diffusion coefficient on the direction of motion considered (anisotropy). This relatively new technique can be used to visualize the orientation of the fibre bundles which represent the main constituent of white matter. It exploits that water molecules in the tissue move faster along the fibre direction than perpendicular to them. The degree of diffusion anisotropy depends on the thickness and constitution of the myelin sheath which covers the individual neuronal fibres (axons). Thus, it can be detected when the white matter is impaired by pathological processes which occur for example in multiple sclerosis or in motor neuron disease (amyotrophic lateral sclerosis). Using this approach, it was also possible to detect anatomical differences in the brain between control persons and dyslexics, or volunteers with persistent developmental stuttering, respectively.
Fig. 4: Fibre orientation maps based on a measurement of the diffusion tensor (DTI). In both images, the dark long structure at the top left corner represents a slice through the central sulcus. Left: direction (red lines) of the largest diffusion coefficient, projected onto the image plane, on top of a conventional T1 weighted MR image. Fibre orientations in volume elements with low anisotropy (fractional anisotropy < 0.2) are suppressed, e.g. in the fibre crossing region marked by the arrow. Right: the same region as in the left image, superimposed with the orientations corresponding to the lowest diffusion coefficient (blue lines). The anisotropy threshold is FA = 0.1 here. The image in the background is a greyscale map of the anisotropy values, FA.
A DTI measurement of the human brain can be used to produce a fibre orientation map which shows the main fibre orientation in every volume element of approximately 3 mm size, say. On such a map the prominent fibre bundles connecting different parts of the cortex can be seen. The automatic reconstruction of the complete bundle pathways
(fibre tracking) requires computer programs which can be designed in many different ways. Algorithms for fibre tracking are being developed and tested by many research groups.
Fig. 5: Probabilistic fibre tracking on the basis of diffusion tensor data. In this case, a Monte Carlo algortihm was used to visualize the main bundles originating from the poisition marked by the arrow.
An important limitation of DTI is apparent where crossings between fibre bundles occur. In a volume element containing, for instance, three perpendicular bundles in equal shares, diffusion appears isotropic. Hence, it is not possible to determine the orientation of the fibre bundles in the volume element. However, new approaches have been proposed to measure the fibre orientation in such a situation. One of them is to measure the "displacement probability density function" (DPDF) in a direct way (
q-space imaging). This function states the probability with which the water molecules have moved in the different directions of space, during some given time.
Fig. 6: Isosurfaces (green) of the DPDF as measured with q-space imaging, in the brain of a healthy volunteer. All points of the isosurface have been moved to the origin of the frame of reference by the distance of the point closest to origin. The corresponding high spatial resolution MR image (T1 weighted) is seen in the background. In the centre of the image the crossing region between corticospinal tract, superior longitudinal fascicle, and right-left fibres of the corpus callosum is clearly visible.
Methods for MR Spectroscopy
Magnetic resonance imaging relies on the detection of water and fat which are present at high concentration in the human body. However, some metabolites can also be detected with MR methods, although with lower spatial resolution.
