Max Planck Research Group MR Physics

Functional MRI makes it possible to study how brain-wide functional networks support perception, cognition, and behavior. These measurements are highly informative for studying functional brain organization, but they are also sensitive to head motion, physiological fluctuations, and susceptibility-related signal loss. In this context, we have explored the use of advanced EPI-based acquisition schemes such as multi-echo imaging. Multi-echo EPI can improve BOLD sensitivity, reduce bias from signal dropout, and help disentangle neural signal contributions from non-neural fluctuations during data analysis.

Quantitative MRI techniques such as T₁ mapping and multiparameter mapping (MPM) can provide measurements that are more directly related to underlying tissue properties than conventional anatomical imaging. Achieving this goal requires careful control of the physical factors that influence the measured signal. One important factor is the transmit radiofrequency field (B₁+), which can vary substantially across the brain. Such variation can lead to spatially dependent signal bias and thereby affect quantitative measurements. For this reason, B₁⁺ mapping is routinely incorporated into quantitative MRI protocols, including MP2RAGE-based T1 mapping and MPM. At the MPIB, we are further developing a dedicated phase-based B₁⁺ mapping method that uses composite RF pulses and can be applied, with minor adaptations, to a range of applications, including infant imaging. The method yields informative whole-brain B₁⁺ maps within a short acquisition time while keeping the measurement largely separate from anatomical image contrast. These maps support the correction of spatially varying signal bias in quantitative MRI. This is particularly important when small differences in MRI-derived measures are of interest—for example between age groups, between intervention and control groups, or between repeated time points in longitudinal study designs.

Figure 1:
Left: A near-midsagittal slice of a T₁ map acquired with MP2RAGE at 3 T. Center: corresponding B₁⁺ map acquired with our phase-based method. Right: T₁ map corrected using the B1+ map shown in the center panel. According to the relationship between MP2RAGE signal amplitude and T₁ (J.P. Marques et al., NeuroImage 2010), the correction increases estimated T₁ values in regions with B₁⁺ > 1, whereas it decreases estimated T₁ values in regions with B₁⁺ < 1. This effect is more pronounced for longer T₁ values, such as those of cerebrospinal fluid (CSF). Accordingly, in the corrected T₁ map, central CSF regions appear brighter, whereas CSF in the peripheral regions of the brain shows lower T₁ values.

Studying the locus coeruleus poses particular challenges because this small brainstem nucleus is difficult to delineate with standard imaging protocols due to limited tissue contrast and its proximity to cerebrospinal fluid. In collaboration with the LINE project, the MR Physics Research Group contributes methodological expertise to MRI of the locus coeruleus, which is highly relevant for sympathetic arousal, higher-order cognitive functions, and research on vagus-nerve-related neuromodulation. Our work focuses on optimizing structural contrast for reliable visualization of the locus coeruleus, on the use of multi-echo EPI for functional MRI in this region, and on adapting multiparameter mapping for the locus coeruleus region. Here, methodological care is especially important: meaningful measurements depend on carefully optimized acquisition parameters that balance spatial resolution and signal-to-noise ratio while minimizing error propagation in derived quantitative maps.

Mobile MRI opens new possibilities for studying the brain in settings that are closer to everyday life and for reaching populations that may be difficult to examine in conventional imaging facilities. At the same time, it raises fundamental methodological questions: How stable are measurements after relocation? Which technical and environmental factors influence image quality? And under which conditions do measurements remain repeatable, reliable, and comparable between mobile and stationary 1.5T systems? We conducted and published a systematic comparison study designed to address these questions. Using phantom-based quality assurance measurements as well as repeated measurements in adult participants, we examined whether relocation of the mobile scanner affected magnetic-field conditions, structural imaging, and functional MRI, with a stationary clinical 1.5T system at a collaborating hospital serving as a reference. The results showed a high level of agreement between mobile and stationary systems: in-vivo measurements of local field deviations (ΔB0) and B1+ distributions were equivalent across systems, structural image-intensity distributions and volume estimates of grey and white matter were highly consistent, and relocation of the mobile scanner did not introduce noticeable changes in the distributions of the measured results. Together, these findings show that high-quality neuroimaging is feasible with mobile 1.5T MRI systems. This line of work is being extended through further comparison measurements, including comparisons of image quality measures between 1.5T and 3T systems.

Figure 2: Human Development ROVER (HD-ROVER) mobile 1.5 T MR imaging platform.
(A) External view of the HD-ROVER mobile imaging unit parked at a scanning site. (B) Interior view of the scanning suite, showing the 1.5 T MRI scanner and the integrated patient table.

Diffusion MRI provides insight into brain microstructure by measuring how water moves through tissue. Beyond conventional diffusion tensor imaging, advanced diffusion methods can better describe complex fibre orientations in white matter and provide more specific quantitative markers of tissue properties. This makes diffusion MRI relevant both for studying the organization of white-matter pathways and for characterizing microstructural changes in grey and white matter. Our work focuses on the development, optimization, and evaluation of diffusion MRI protocols. This requires carefully balancing spatial resolution, angular resolution, acquisition time, sensitivity to image distortions, signal-to-noise-ratio, and the reproducibility of derived measures. This is particularly demanding in small structures such as the fornices. In fibre tracking, orientation information from diffusion MRI is used to follow possible axonal pathways. For quantitative microstructural analyses, model fitting is used to estimate tissue-related parameters rather than fibre trajectories. In both cases, controlling partial-volume effects and ensuring that the derived measures remain reproducible are essential.

At MPIB, such diffusion methods are used in studies of early development, cognitive development across the lifespan, healthy aging, learning-related plasticity, and environmental influences. Examples of methods used in this context include constrained spherical deconvolution (CSD) for fibre-orientation modelling and fibre tracking, diffusion kurtosis imaging (DKI) for tissue-complexity-sensitive parameters, and NODDI-like compartment models for neurite-related measures.

Figure 3: Directional information in diffusion MRI of the fornix region.
(A) Sagittal and transverse b = 0 images, i.e. images acquired without diffusion weighting, serve as anatomical reference images; the marked boxes show the regions displayed in panels B and C. (B, C) Tensor-based color-coded orientation maps show the main local diffusion directions in the region of the fornices. Red indicates left–right, green anterior–posterior, and blue superior–inferior orientation. These maps illustrate how diffusion MRI provides directional information that can be used to reconstruct white-matter pathways. In small fibre bundles such as the fornices, this requires high spatial resolution and careful control of partial-volume effects.