Comprehensive Pipeline for Large-Scale Neuronal Network Construction and Analysis

This repository contains the complete computational pipeline for constructing, simulating, and analyzing large-scale biologically detailed neuronal networks spanning multiple cortical and thalamic regions. The documentation is written in MyST (Markedly Structured Text) format for optimal compatibility with Jupyter Book and NeuroLibre publishing.

Overview

Our methodology integrates morphologically realistic single-cell models with region-specific connectivity rules derived from anatomical data, enabling the generation of networks containing over 145,000 neurons and 581,743 synaptic and electrical connections. The pipeline employs a multi-layered approach combining chemical synapses (AMPA, NMDA, GABA_A, GABA_B) with electrical gap junctions, and incorporates cell-type-specific classification based on transcriptomic and morphological criteria.

Key Features

• Morphologically detailed single-cell models based on experimental reconstructions
• Cell-type-specific connectivity rules derived from anatomical and physiological data
• Comprehensive thalamocortical integration with appropriate driver/modulator distinctions
• Standardized model representation using NeuroML2 for interoperability
• Scalable architecture supporting networks with >145K neurons and >580K connections
• MyST-formatted documentation with full citation support for academic publishing

Repository Structure

├── cell_files/           # Morphologically detailed neuronal models
├── channel_files/        # Ion channel models
├── net_files/            # Generated network files
├── net_params/           # Network parameter configurations
├── analysis_out/         # Analysis output files (JSON format)
├── html/                 # Interactive HTML visualizations
├── plots/                # Static plots and figures
├── *.py                  # Core analysis and simulation scripts
├── *.nml                 # NeuroML2 model files
├── neuronal_network_construction_analysis_paper.md  # Main paper (MyST format)
└── references.bib        # Bibliography for citations

Installation and Setup

Using Conda

# Clone the repository
git clone https://github.com/trernghwhuare/network-analysis-package.git
cd network-analysis-package

# Create environment from environment.yml
conda env create -f environment.yml
conda activate pynml-neurolibre

# Install NEURON separately (not available via conda)
# Follow instructions at: https://neuron.yale.edu/neuron/download

Using Pixi (Recommended)

# Install pixi if not already installed
curl -fsSL https://pixi.sh/install.sh | bash

# Activate the environment
pixi shell

Usage

The pipeline consists of several key steps:

1. Single-cell simulation: Run individual cell simulations using scripts like 00_sim_TCRc.py
2. Network construction: Generate networks using scripts like 11_C2T_max+.py
3. Analysis: Perform network analysis using analysis.py and related scripts
4. Visualization: Generate interactive HTML visualizations using create_interactive_layered_graph.py

For detailed usage instructions, refer to the main paper document.

Reproducibility

This repository is designed for full reproducibility and has been prepared for publication on neuroLibre using MyST format. All dependencies are specified in:

• pixi.toml (primary dependency management)
• environment.yml (conda environment)
• requirements.txt (pip requirements)

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• Blue Brain Project for neuronal reconstructions and standardized cell templates
• NeuroMorpho.Org database for morphological data
• Allen Institute Cell Types Database for transcriptomic classifications
• pyNeuroML and NEURON communities for simulation infrastructure