Related Experiment Video
Updated: Jun 15, 2025
A Standardized Pipeline for Examining Human Cerebellar Grey Matter Morphometry using Structural Magnetic Resonance Imaging
Published on: February 4, 2022
A novel approach to cytoarchitectonics: developing an objective framework for the morphological analysis of the
Matija Vid Prkačin1,2, Zdravko Petanjek1,2, Ivan Banovac1,2
1Department of Anatomy and Clinical Anatomy, University of Zagreb School of Medicine, Zagreb, Croatia.
Introduction:
The cytoarchitectonic boundaries between cortical regions and layers are usually defined by the presence or absence of certain cell types. However, these cell types are often not clearly defined and determining the exact boundaries of regions and layers can be challenging. Therefore, in our research, we attempted to define cortical regions and layers based on clear quantitative criteria.
Methods:
We performed immunofluorescent anti-NeuN labelling on five adult human brains in three cortical regions-Brodmann areas (BA) 9, 14r, and 24. We reconstructed the cell bodies of 90,723 NeuN-positive cells and analyzed their morphometric characteristics by cortical region and layer. We used a supervised neural network prediction algorithm to classify the reconstructions into morphological cell types. We used the results of the prediction algorithm to determine the proportions of different cell types in BA9, BA14r and BA24.
Results:
Our analysis revealed that the cytoarchitectonic descriptions of BA9, BA14r and BA24 were reflected in the morphometric measures and cell classifications obtained by the prediction algorithm. BA9 was characterized by the abundance of large pyramidal cells in layer III, BA14r was characterized by relatively smaller and more elongated cells compared to BA9, and BA24 was characterized by the presence of extremely elongated cells in layer V as well as relatively higher proportions of irregularly shaped cells.
Discussion:
The results of the prediction model agreed well with the qualitative expected cytoarchitectonic descriptions. This suggests that supervised machine learning could aid in defining the morphological characteristics of the cerebral cortex.
