Background: Huntington's disease is a progressive neurodegenerative disorder. Brain atrophy, as measured by volumetric magnetic resonance imaging (MRI), is a downstream consequence of neurodegeneration, but microstructural changes within brain tissue are expected to precede this volumetric decline. The tissue microstructure can be assayed non-invasively using diffusion MRI, which also allows a tractographic analysis of brain connectivity.

Methods: We here used ex vivo diffusion MRI (11.7T) to measure microstructural changes in different brain regions of end‐stage (14 weeks of age) wild type and R6/2 mice (male and female) modeling Huntington's disease. To probe the microstructure of different brain regions, reduce partial volume effects and measure connectivity between different regions, a 100 μm isotropic voxel resolution was acquired.

Results: Although fractional anisotropy did not reveal any difference between wild‐type controls and R6/2 mice, mean, axial, and radial diffusivity were increased in female R6/2 mice and decreased in male R6/2 mice. Whole brain streamlines were only reduced in male R6/2 mice, but streamline density was increased. Region‐to‐region tractography indicated reductions in connectivity between the cortex, hippocampus, and thalamus with the striatum, as well as within the basal ganglia (striatum—globus pallidus—subthalamic nucleus—substantia nigra—thalamus).

Conclusions: Biological sex and left/right hemisphere affected tractographic results, potentially reflecting different stages of disease progression. This proof‐of‐principle study indicates that diffusion MRI and tractography potentially provide novel biomarkers that connect volumetric changes across different brain regions. In a translation setting, these measurements constitute a novel tool to assess the therapeutic impact of interventions such as neuroprotective agents in transgenic models, as well as patients with Huntington's disease.

2.2 | MRI

For MR imaging, mouse heads were immersed into proton‐free FluorInert (Sigma) in a syringe to immobilize the head for long scanning times at a high resolution, where minimal movement between gradient directions can affect image quality. Images were acquired with a Bruker AV3 HD 11.7T/89mm vertical bore microimaging system, with a Micro 2.5 gradient insert (capable of up to 150 G/cm) in a 25 mm quadrature birdcage coil (Bruker). Sample temperature was maintained at 21 ± 0.1°C with a Bruker SmartCooler BCU‐1 40/50 air chiller and probe heater with a thermocouple feedback loop.39 Scanning parameters were defined in ParaVision 6.0. A 3D spin‐echo diffusion MRI scan was acquired (repetition time, TR = 700 ms, echo time, TE = 24 ms, diffusion time = 12 ms, diffusion encoding duration 5 ms, b value = 1401.57 s/mm², with 12 non-collinear diffusion directions and one A₀ image, number of averages, NA = 1, matrix 196 × 128 × 128, 100 μm isotropic resolution, and 31 h 3 min acquisition time), followed by a coregistered 3D T₂‐weighted spin‐echo scan at the same resolution (TR = 3500, TE = 40, NA = 1, and 15 h 55 min acquisition time). We used a 3D spin‐echo diffusion sequence to achieve high isotropic spatial resolution for investigating connectivity between small anatomical structures in the mouse. This approach is time‐consuming, which limits the diffusion paradigm, but we chose this method over an EPI approach to avoid potential image distortion from susceptibility artifacts at high field.

2.3 | MR image processing

Diffusion MR images were processed using Diffusion Spectrum Imaging (DSI) studio (available at: http://dsi-studio.labsolver.org/dsi-studio-download). The brain was masked, and the rest of the head was removed as background before processing. Reconstruction of the DTI was achieved by performing an Eigenvector analysis on the calculated tensor. Mean (MD), radial (RD), and axial diffusion (AD) maps were computed for measurement of scalar indices, as was fractional anisotropy (FA). Streamlines reflecting connectivity were mapped for the whole brain using a multidirectional deterministic fiber tracking algorithm using 10 seeds/voxel with random subvoxel positioning of seeds. Trilinear interpolation in all orientations with a step size of 0.05 mm (1/2 voxel length) was performed with an FA of 0.02 serving as a tracing endpoint of streamlines at the end of the sample, as fiber tracing was not restricted to only white matter. A streamline length of 0.2 mm (2× voxel length) was considered the minimum length of streamlines, with a maximum length of 10 mm (15 mm sample sagittal length) and an angular threshold of 60° being used to end connectivity tracing. Total number of streamlines for whole brain and individual ROIs were recorded to calculate a streamline density (streamlines/mm³) for group comparisons. Streamline length was also recorded, as there is some evidence that it is associated with regional atrophy in premanifest HD patients.