Brain Cell-Type Density Analysis

Pipeline for registaring, analyzing and visualizing neuroanatomical point cloud data from whole brain imaging. The registration pipeline uses affine and b-spline transformations using intrinsic anatomical features with Elastix. The density-based analysis pipeline uses DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to identify clusters of neuronal cells and visualize their spatial distribution in 3D.

This repository provides a modular and reproducible pipeline for:

1 - Whole-brain 3D registration

2 - Whole-brain 3D transformation at high resolution

3 - 3D point cloud clustering using DBSCAN

4 - Visualization and analysis of clustered cellular data

Citation:

A Comprehensive Atlas of Cell Type Density Patterns and Their Role in Brain Organization

Rodrigo Muñoz-Castañeda#, Ramesh Palaniswamy, Jason Palmer, Rhonda Drewes, Corey Elowsky, Karla E. Hirokawa, Nicholas Cain, Kannan Umadevi Venkataraju, Hong-Wei Dong, Julie A. Harris, Zhuhao Wu, Pavel Osten#

#Corresponding Author

bioRxiv doi: 10.1101/2024.10.02.615922

The datasets used in this paper have been deposited for access at: http://download.brainimagelibrary.org/biccn/osten/cellcounting/

High-Resolution Image Registration (MATLAB)

File	Purpose
registerWholeBrain.m	Aligns whole-brain volume to reference space
registerBrainSlice.m	Optional slice-level refinement
run_FullResolutionReg.sh	Bash script to automate full registration pipeline

To run:

bash scripts/run_FullResolutionReg.sh

---

Requirements

MATLAB

• R2021a+ with Image Processing Toolbox
• Elastix (if required by registration pipeline)

Python

• numpy
• scipy
• open3d
• scikit-image
• matplotlib
• natsort

---