In vivo Diffusion Tensor Magnetic Resonance Tractography of the Sheep Brain: An Atlas of the Ovine White Matter Fiber Bundles

Abstract

Diffusion Tensor Magnetic Resonance Imaging (DTI) allows to decode the mobility of water molecules in cerebral tissue, which is highly directional along myelinated fibers. By integrating the direction of highest water diffusion through the tissue, DTI Tractography enables a non-invasive dissection of brain fiber bundles. As such, this technique is a unique probe for in vivo characterization of white matter architecture. Unraveling the principal brain texture features of preclinical models that are advantageously exploited in experimental neuroscience is crucial to correctly evaluate investigational findings and to correlate them with real clinical scenarios. Although structurally similar to the human brain, the gyrencephalic ovine model has not yet been characterized by a systematic DTI study. Here we present the first in vivo sheep (ovis aries) tractography atlas, where the course of the main white matter fiber bundles of the ovine brain has been reconstructed. In the context of the EU's Horizon EDEN2020 project, in vivo brain MRI protocol for ovine animal models was optimized on a 1.5T scanner. High resolution conventional MRI scans and DTI sequences (b-value = 1,000 s/mm², 15 directions) were acquired on ten anesthetized sheep o. aries, in order to define the diffusion features of normal adult ovine brain tissue. Topography of the ovine cortex was studied and DTI maps were derived, to perform DTI tractography reconstruction of the corticospinal tract, corpus callosum, fornix, visual pathway, and occipitofrontal fascicle, bilaterally for all the animals. Binary masks of the tracts were then coregistered and reported in the space of a standard stereotaxic ovine reference system, to demonstrate the consistency of the fiber bundles and the minimal inter-subject variability in a unique tractography atlas. Our results determine the feasibility of a protocol to perform in vivo DTI tractography of the sheep, providing a reliable reconstruction and 3D rendering of major ovine fiber tracts underlying different neurological functions. Estimation of fiber directions and interactions would lead to a more comprehensive understanding of the sheep's brain anatomy, potentially exploitable in preclinical experiments, thus representing a precious tool for veterinaries and researchers.

Keywords: DTI tractography; atlas; brain; diffusion tensor imaging; sheep.

Figure 1

DTI-derived maps for a representative ovine brain. Computed maps of diffusion (mm²/s × 10⁻³) are shown in gray scale (upper row) and color LUT (Look Up Tables) display (lower row). (A) Axial Diffusivity (AD) map. (B) Radial Diffusivity (RD) map; notably, low values in correspondence of the ovine internal capsule indicate the dominance of the principal direction in that region. (C) Mean Diffusivity (MD) map. (D) Fractional Anisotropy (FA) map. Color-coded FA maps (lower row) display fibers with rostral-caudal (R-C) direction in blue, fibers with dorsal-ventral (D-V) direction in green, and fibers with medial-lateral (M-L) in red.

Figure 2

Location of the ROIs for tractography reconstructions. Seed regions-of-interest (ROIs) were contoured in precise anatomical positions selected to encompass the tract cross-section, according to the location of different fiber tracts. (A) ROIs for CST tractography reconstruction were placed on the transverse plane. The first ROI was placed in the genu of the internal capsule, that it is known to comprehend the ovine pyramidal tract and is found between the caudate nucleus and the putamen. The other ROI included the ventral mesencephalic plane, where the pons is located. (B) ROIs for CC tractography reconstruction were placed on the sagittal plane. ROIs were carefully positioned on two slices to contain a section of the fiber tract of interest in 3D space. In particular, the corpus callosum at the level of septum pellucidum was contoured by a ROI extending from the infrasplenial sulcus to the genual sulcus. (C) ROIs for VP tractography reconstruction were placed on the transverse plane. The ROI including the chiasm was set between the rostral encephalic longitudinal fissure and the posterior tuber cinereum. The optic radiations are known to run in the retrolentiform portion of the internal capsule and to diverge caudally before reaching the occipital cortex. Thus, a ROI has been placed just above the caudal portion of the lateral ventricle. (D) ROIs for FX tractography reconstruction were placed on the transverse plane. The first ROI encompassed the white matter medial to the genu of the internal capsule, rostral to the anterior horns of the lateral ventricles. The other ROI has been contoured posteriorly to the optic chiasm, external to the rostral arm of the internal capsule. (E) ROIs for OF tractography reconstruction were placed on the transverse plane. The rostral ROI was placed in the orbital gyrus, just posterior to the olfactory bulbs. The caudal ROI has been placed just above the caudal portion of the lateral ventricle, exactly as for the optic radiations that are known to be closely surrounded by the OF fibers. (F) Example of raw CC and FX tracts resulting from the first tracking procedures, that were successively refined by inserting exclusion ROIs that allowed the removal of spurious fibers.

Figure 3

Reproducibility of CST tractography reconstruction in all the 10 sheep. The anatomical course of the corticospinal tract (CST, blue) is shown as three-dimensional rendering in a sagittal (A) and frontal (B) view for each of the 10 sheep. At a qualitative, visual assessment the course of tracts appears very similar across the 10 animals, thus highlighting the reproducibility of the tractography pipeline.

Figure 4

Details of the course of the CST in a representative sheep. (A) CSTs are displayed as volumes on the corresponding T1-weighted anatomical image. (B) CSTs are overlaid as binary masks onto the T1-weighted volumetric images.

Figure 5

Details of the course of the main ovine white matter fiber bundles in a representative sheep. Tracts are displayed as volumes onto high-resolution T1 anatomical images and showed in sagittal, lateral, coronal and transverse views. (A) Corpus callosum (CC, orange). (B) Visual pathway (VP, yellow). (C) Fornix (FX, pink). (D) Occipitofrontal fasciculus (OF, green).

FIGURE 6

Representative images from the ovine tractographic atlas. The tractography atlas of the five ovine tracts is visualized by means of FSLeyes in sagittal view for CST (A) and CC (B), transverse view for FX (C), and coronal view for VP (D) and OF (E), overlaid onto the publicly available stereotaxic T1-weighted ovine brain atlas 0.5 mm. The origin of the reference system with the xyz-values (0;0;0) is a vertical line perpendicularly intersecting the superior aspect of the rostral commissure. All coordinates are given in millimeters (mm). Values of the x-axis increase from left to right, values of the y-axis increase from rostral to caudal, while values of the z-axis increase in the dorsal direction (31). (A) sagittal view for CST: on x axis, from −12.9 to 5.1 mm (slice spacing: 1.2 mm). (B) sagittal view for CC: on x axis, from −14.6 to 3.4 mm (slice spacing: 1.2 mm). (C) transverse view for FX: on y axis, from 3.65 to −17.35 mm (slice spacing: 1.4 mm). (D) coronal view for VP: on z axis, from −11.1 to 15.9 mm (slice spacing: 1.8 mm). (E) coronal view for OF: on z axis, from −6.225 to 17.025 mm (slice spacing: 1.55 mm). By fixing the cutoff value at 4, only the fibers that are consistent in at least 4 sheep out of 10 are shown.

FIGURE 7

Scatterplots displaying FA and MD in the five reconstructed fiber bundles. Color-coding of the scatterplots corresponds to Figures 4, 5. Blue=CST; orange=CC; yellow=VP; fuchsia=FX; green=OF). Black bars show the mean and SD between 10 CCs and 20 CSTs, VPs, FXs, OFs (10 right and 10 left), each one represented by a single dot.