The cytoarchitectonic landscape revealed by deep learning method facilitated precise positioning in mouse neocortex.

Area of Science:

• Neuroscience
• Computational Biology
• Brain Imaging

Background:

• The neocortex's intricate structure presents challenges in precisely locating cortical regions due to a lack of clear landmarks.
• Accurate anatomical referencing is crucial for understanding brain function and development.

Purpose of the Study:

• To develop and validate a cytoarchitectonic landmark identification pipeline for the mouse neocortex.
• To enable detailed 3D analysis of neuronal distribution and morphology.
• To investigate age-related structural changes in cortical regions.

Main Methods:

• Utilized fluorescence micro-optical sectioning tomography for whole-brain imaging of stained mouse brains.
• Employed a fast 3D convolution network for segmenting neuronal somas across the entire neocortex.
• Analyzed 3D cytoarchitectonic profiles and neuronal morphology, independent of sectioning angles.

Main Results:

• Generated distribution maps visualizing neuronal numbers and morphological types, defining the cytoarchitectonic landscape and identifying landmarks like the barrel cortex.
• Successfully aligned cortical regions across different ages, revealing structural alterations in the barrel cortex during aging.
• Observed spatiotemporally gradient distributions of spindly neurons, particularly in the primary visual area's deep layers, with decreasing proportions over time.

Conclusions:

• The developed pipeline effectively identifies neocortical cytoarchitectonic landmarks, improving structural understanding.
• This method facilitates the study of neuronal organization, age-related changes, and regional specializations.
• The findings provide a foundation for further neuroscientific exploration using advanced imaging and computational analysis.