Zebrafish Hindbrain Functional Connectomics

This repository hosts collaborative analyses on the structure–function relationships in the zebrafish hindbrain, integrating synaptic connectivity, morphology-based predictions, and network modeling.

The dataset can be visualized with Neuroglancer here. You need to log in with a Gmail account. For more details regarding proofreading and programatic access to the dataset please visit : https://jboulanger91.github.io/fish1.5-release/

---

Repository Structure

1. Downloading neuronal morphologies and metadata

Pipeline for reconstructing and organizing neuronal morphologies, synapse positions, and connected segments from the clem_zfish1 dataset. This includes:

• Downloading and organizing meshes (nucleus, soma, axon, dendrites)
• Exporting Neuroglancer-resolution synapse tables (8×8×30 nm)
• Generating per-neuron metadata (IDs, reconstruction status, functional labels if available)
• Optional extraction of ΔF/F dynamics for functionally imaged neurons

Outputs are per-neuron/axon folders containing:

• *_metadata.txt (metadata)
• *_axon.obj, *_dendrite.obj, *_soma.obj, *.obj (meshes)
• *_presynapses.csv, *_postsynapses.csv (synapse tables)
• , *_dynamics.pdf, *_dynamics.hdf5 (optional functional data)

Main script: download_axons_neurons_pipeline.py
Helper module: download_axons_neurons_helper.py
Environment file: env_clem_zfish1_global.yaml

---

2. Reference brain registration

Pipeline for mapping neuronal morphologies and synapse coordinates from the clem_zfish1 dataset into a standardized zebrafish reference brain. This step applies ANTs deformation fields to warp neuron meshes and synapse locations into a shared anatomical space, then generates skeletonized neuron reconstructions.

This includes:

• Mapping soma, axon, and dendrite meshes into reference space
• Warping presynaptic and postsynaptic coordinates
• Generating mapped OBJ meshes and synapse OBJ spheres
• Skeletonizing neurons or axons using TEASAR
• Healing skeleton gaps and embedding synapse node labels
• Producing aligned SWC skeletons for further analysis

Outputs are per-neuron folders containing:

• *_axon_mapped.obj, *_dendrite_mapped.obj, *_soma_mapped.obj, *_mapped.obj (mapped meshes)
• *_presynapses_mapped.csv, *_postsynapses_mapped.csv (mapped synapse tables)
• *_presynapses_mapped.obj, *_postsynapses_mapped.obj (synapse OBJ sphere files)
• *_mapped.swc, *.swc (skeletonized neurons with synapse annotations)

Main script: register_and_skeletonize.py
Helper module: register_and_skeletonize_cells_helpers.py
Environment file: env_register_and_skeletonize.yaml

---

3. Connectivity Matrices and Network Diagram Generation

This folder contains the pipelines used to compute synaptic connectivity matrices and to generate two-layer network diagrams from the clem_zfish1 zebrafish hindbrain connectome. These analyses integrate CAVE‑derived synapse tables, registered neuron meshes, and functional classifications.

Both pipelines require the mapped-neuron outputs generated in Step 2 (Reference brain registration).

3.1 Connectivity Matrices

This pipeline constructs directional pre→post synaptic connectivity matrices between functional/morphological neuron and axon classes.

Outputs:

• Pooled connectivity matrix (PDF)
• Left/right–split connectivity matrix (PDF)
• Left/right–split excitatory/inhibitory matrix (PDF)

Main script: make_connectivity_matrices.py
Helper module: connectivity_matrices_helper.py
Environment: env_clem_zfish1_global.yaml

3.2 Two‑Layer Network Diagrams

This pipeline generates compact, population‑level schematic connectivity diagrams for:

• cMI — contralateral motion integrators
• MON — motion onset neurons
• MC/SMI — slow‑motion integrators
• iMI+ — excitatory ipsilateral motion integrators
• iMI− — inhibitory ipsilateral motion integrators
• iMI_all — pooled ipsilateral motion integrators

These diagrams require registered neuron meshes, because hemisphere assignment and cross‑side vs. same‑side connectivity depend on mapped spatial coordinates.

Outputs:

• One four‑panel PDF per population (ipsi/contra × inputs/outputs)
• A detailed text connectivity table per population

Main script: make_connectivity_diagrams.py
Helper module: connectivity_diagrams_helper.py
Environment: env_clem_zfish1_global.yaml

---

4. Morphology-based prediction of neuronal functional types

Includes scripts to predict functional properties (e.g., motion integrator, motion onset neurons) from morphology.

Environment file: env.yaml

---

5. Connectome-constrained network modeling

Computational models that simulate network dynamics under realistic connectome constraints.

Environment file: env.yaml

---