Ferdinand Schweser1,2, Claire M Modica1,3, Nicola Bertolino1, Paul Polak1, Marilena Preda1,2, and Robert Zivadinov1,2
1Buffalo Neuroimaging Analysis Center, Department of Neurology, Jacobs School of Medicine and Biomedical Sciences, The State University of New York at Buffalo, Buffalo, NY, United States, 2MRI Molecular and Translational Research Center, Jacobs School of Medicine and Biomedical Sciences, The State University of New York at Buffalo, Buffalo, NY, United States, 3Neuroscience Program, Jacobs School of Medicine and Biomedical Sciences, The State University of New York at Buffalo, Buffalo, NY, United States
Synopsis
The technique of MR microscopy (μMRI) has evolved into an important tool for morphologic phenotyping and computational neuroanatomy research. However, due to the inherent challenges of mouse imaging, μMRI of mouse brain has so far mostly been limited to post mortem tissue, often relying on perfusion with a mixture of saline and a T1-shortening constant agent, which increases the MRI sampling efficiency.
In this work, we demonstrate the feasibility of in vivo μMRI of mouse brain at 9.4 Tesla with a resolution of 50μm using a cryogenic brain coil and an optimized imaging sequence.
Introduction
Pre-clinical magnetic resonance imaging (MRI)-based studies have traditionally focused on rat models, because anatomical structures are intrinsically more difficult to resolve in the much smaller mouse brain. Small-animal MRI systems with magnetic field strengths of 7 Tesla and above allow mouse imaging, but the obtained image quality is generally much lower compared to that in rats.
The technique of MR microscopy (μMRI), MRI with ultra-high resolution below 100μm, has evolved into an important tool for morphologic phenotyping and computational neuroanatomy research. However, due to the inherent challenges of mouse imaging, μMRI of mouse brain has so far been limited mostly to post mortem tissue, often relying on perfusion with a mixture of saline and a T1-shortening constant agent, which increases the MRI sampling efficiency1,2. This way, μMRI can achieve an isotropic voxel size as low as 50μm or 42μm at 9.4T1,2. In vivo, μMRI of mice has recently been demonstrated with 100μm isotropic voxel size (at 7T)3.
In this work, we demonstrate the feasibility of in vivo μMRI of mouse brain at 9.4 Tesla with a resolution of 50um using a cryogenic brain coil and an optimized imaging sequence.
Methods
Experimental setup and sequence: Experiments were performed on a 9.4 Tesla Bruker BioSpec 94/20 USR equipped with a 440 mT/m gradient system and a cryogenically cooled dual-element transmit-receive 1H surface coil. To achieve optimal readout efficiency, we used a 3D multi-echo gradient-echo (MGE) sequence with the following parameters: TE1=3.5ms, ΔTE=4ms, 16 echoes, TR=180ms, FA=25 degrees, matrix 315x250x160, FOV 19x12.5x8mm3, BW=85kHz, TA=120min, resulting in a voxel size of 60x50x50μm3. The high number of echoes ensured an efficient sampling of the MRI signal decay.
Animal: The protocol was demonstrated in a Cre transgenic mouse on a C57BL/6 background (28 weeks old), which was anesthetized using 1-3% isoflurane under monitoring of respiration rate and body temperature. The experiment was approved by our Institutional Animal Care & Use Committee (IACUC).
Image reconstruction: Images were reconstructed off-line from raw k-space data on a high-performance computation server (384 GB RAM) using in-house developed software in MATLAB (R2013b). To increase the signal-to-noise ratio, we averaged all 16 magnitude images. Coil-related signal inhomogeneities were corrected for by applying ANTs N4 bias-field correction.
Results
Figures 1 and 2 show exemplary μMRI images (slice thickness 50µm) spaced 250µm apart (ventral to dorsal from top-left to bottom-right). The images showed exquisite anatomical contrast with a detail so far known only from histology stains, with numerous small brain structures discernible (see labels). Figure 3 depicts enlarged views of selected anatomical regions.
A 3D visualization of the ventricular system and arterial vessels (first echo; time-of-flight mechanism) could be obtained via maximum intensity projections (MIPs). A minimum intensity projection (mIP) provided an overview of the venous vasculature through the blood oxygenation dependency (BOLD) effect. (projection images not shown here, because of poor visualization in 2D)
Only minor image artifacts were discernible, primarily in the ventral and caudal parts of the brain. These artifacts may be explained by respiration-related changes of the magnetic field homogeneities (in vivo) during the acquisition.
Discussion and Conclusion
In this work we developed an optimized imaging protocol for in vivo anatomical μMRI of the mouse brain with a resolution so far known only from post mortem studies.
The voxel size demonstrated here is about seven times higher than that recently presented in vivo by Lin et al.3 at 7T (100μm). This ultra-high resolution imaging protocol allows studying in vivo small-scale morphological, structural, vascular, and neuronal changes in mouse models. In particular, it opens the door to extremely accurate longitudinal examinations of morphological changes.
The (almost) isotropic voxel-size allows an arbitrary reformatting of the images into different planes with minimal loss of image detail. A combination with interpolation strategies, such as k-space zero-filling, may even further improve the visualization of small-scale structures below 100μm.
Acknowledgements
No acknowledgement found.References
[1] Liu C, Li W & Johnson GA. MR Microscopy of Brain Cytoarchitecture by Quantitative Mapping of Magnetic Susceptibility. ISMRM 2011. p36.
[2] Johnson GA, Cofer GP, Gewalt SL & Hedlund LW. Morphologic phenotyping with MR microscopy: the visible mouse. Radiology 2002, 222(3), 789–93.
[3] Lin L, Wu S, Bin G & Yang C. Intensity inhomogeneity correction using N3 on mouse brain magnetic resonance microscopy.
J Neuroimaging 2013, 23(4), 502–7.