Histology-based Average Template of the Marmoset Cortex with Probabilistic Localization of Cytoarchitectural Areas
Presenting author:
The rapid adoption of marmosets in neuroscience has created a demand for three-dimensional (3D) brain atlases to facilitate data integration in a common reference space. We introduce the Nencki–Monash (NM) template, a new, open-access atlas of the marmoset cortex representing a morphological average of brains of 20 young adult individuals, obtained by 3D reconstructions generated from Nissl-stained serial sections.
The resource allows for direct estimates of the most likely spatial coordinates of each cortical area and preserves quantitative information about the laminar structure of the cortex. Further, the method used to generate the template allows us to obtain precise estimates of the cortical thickness as well as the neuronal density profiles of every cortical area. We show how these additional data layers can aid the mapping of the experimental data into the template space. We also demonstrate the practical use of the cytoarchitectonic profiles as a source of reliable training data for artificial intelligence (AI), deep- and machine-learning solutions aimed at identifying cortical areas based on their cytoarchitectonic characteristics. To enable integration with magnetic resonance imaging (MRI) and tracer-based connectivity data we provide spatial transformations between the NM atlas and other available marmoset brain templates.
Overall, the NM template combines the main advantages of histology-based atlases with features more commonly associated with MRI-based templates, such as isotropic nature of the dataset and probabilistic analyses. We are confident the new resource will strongly facilitate detailed anatomical and functional studies of the marmoset cortex, and benefit the neuroscience community.
The resource allows for direct estimates of the most likely spatial coordinates of each cortical area and preserves quantitative information about the laminar structure of the cortex. Further, the method used to generate the template allows us to obtain precise estimates of the cortical thickness as well as the neuronal density profiles of every cortical area. We show how these additional data layers can aid the mapping of the experimental data into the template space. We also demonstrate the practical use of the cytoarchitectonic profiles as a source of reliable training data for artificial intelligence (AI), deep- and machine-learning solutions aimed at identifying cortical areas based on their cytoarchitectonic characteristics. To enable integration with magnetic resonance imaging (MRI) and tracer-based connectivity data we provide spatial transformations between the NM atlas and other available marmoset brain templates.
Overall, the NM template combines the main advantages of histology-based atlases with features more commonly associated with MRI-based templates, such as isotropic nature of the dataset and probabilistic analyses. We are confident the new resource will strongly facilitate detailed anatomical and functional studies of the marmoset cortex, and benefit the neuroscience community.