Advancements in neurological imaging modalities have the potential to enable new understanding in nervous system organization. One recently developed example, diffusible iodine-based contrast-enhanced computed tomography (diceCT), is an iodine staining and X-ray μCT-imaging technique that allows for the differentiation of myelinated and unmyelinated nervous tissues at finer scales of spatial resolution than standard 3D brain-imaging tools permit. DiceCT is versatile, capable of imaging specimens across several orders of magnitude in size, and enables large structures such as human brains to be studied alongside those of much smaller organisms like fish, frogs, and birds. The staining agent can also be removed, allowing the application of histological methods and analyses following diceCT imaging. However, because of its relative newness, ground-truthing diceCT by determining its reliability for distinguishing precise boundaries between neuroanatomical structures is critical for its potentially broad application by anatomists, neuroscientists, and organismal biologists. For example, some tissue-level neuroanatomical features are poorly differentiated due to broad tissue similarities among adjacent structures or obscure boundaries (such as thalamic nuclei, which are separated by many small white-matter tracts). To address this, we compared 3D, orientation-matched image stacks of diceCT adolescent Sprague Dawley rat brains to structural and region-outline diagrams from the well-established, Paxinos and Watson 2D rat brain atlas (7th edition). We scored the visibility and consistency by which atlas-identifiable neuroanatomical features can be documented in diceCT images. We attempted to identify more than 1,000 neuroanatomical structures, spanning seventeen 2D atlas sections and all brain regions. The relative brightness (e.g., “medium gray”, “high white”) of diceCT structures was documented qualitatively, and results were exclusively scored as: (i) visibly distinct and replicable, (ii) visible but with ambiguous boundaries; (iii) visibly distinct and replicable after image enhancement, (iv) visibly distinct but non-replicable (including asymmetrically present), or (v) visibly non-distinct and non-replicable. This scheme allowed us to reliably distinguish neuroanatomical structures in diceCT images and estimate what proportion can be visualized. Preliminary results scored approximately 60% of white-matter (= 172) and 20% of gray-matter (= 157) structures as (i) visibly distinct and replicable in diceCT brains, which contrasted with approximately 20% (= 57) and 55% (= 432), respectively, as (v) visibly non-distinct and non-replicable. These results are generally better for structure differentiation than atlas-provided cresyl violet brain sections. Notably, few structures (< 5%) were (iv) visibly distinct but non-replicable (e.g., asymmetrically present), and most of those were concentrated in the forebrain. Altogether, our findings point towards a potential future for diceCT as a consistently reliable neuroanatomical imaging standard.
Read full abstract