A Multi-Scale Neuron Morphometry Dataset from Peta-voxel Mouse Whole-Brain Images

Area of Science:

• Neuroscience
• Computational Biology
• Bioinformatics

Background:

• Neuron morphology and sub-neuronal patterns are crucial for understanding cell types and brain network organization.
• The BRAIN Initiative Cell Census Network (BICCN) has generated extensive whole-brain imaging data.
• Reconstructing multi-scale neuron morphometry at a whole-brain scale presents significant computational and workflow challenges.

Purpose of the Study:

• To develop a scalable, cloud-based platform for processing peta-scale imaging data.
• To generate the largest multi-scale neuron morphometry dataset to date.
• To provide a comprehensive resource for whole-brain morphological studies.

Main Methods:

• Development of a cloud-based, collaborative platform for handling peta-scale imaging data.
• Integration of diverse hardware, tools, and algorithms for neuron reconstruction.
• Application of the platform to hundreds of sparsely labeled mouse brains.

Main Results:

• Generation of the largest multi-scale morphometry dataset from mouse brains.
• Annotation of 182,497 cell bodies, 15,441 local morphologies, and 1,876 full reconstructions.
• Identification of sub-neuronal arborizations, axonal tracts, and 2.63 million putative boutons.
• Registration of all morphometric data to the Allen Common Coordinate Framework (CCF) atlas.

Conclusions:

• The developed platform effectively addresses challenges in whole-brain neuron morphometry reconstruction.
• The generated dataset is a valuable resource for cross-scale morphological studies in the mouse brain.
• This work advances our ability to analyze complex neural structures and network organization.