Axon velocity

Graph-based algorithm to reconstruct axonal branches and compute velocities from extracellular multi-electrode array (MEA) recordings in Python.

Installation

To install the axon_velocity package, you can clone the repo and install using pip:

pip install axon_velocity

To install from sources:

git clone https://github.com/alejoe91/axon_velocity.git
cd axon_velocity
python setup.py install (or develop)

Requirements

axon_velocity depends on the following packages, which are automatically installed

• numpy
• matplotlib
• scipy
• networkx
• sklearn
• MEAutility
• probeinterface

For the simulation notebooks in the simulation_notebooks folder, additional requirements are needed:

• NEURON
• LFPy
• neuroplotlib

All additional requirements can be installed with: pip install -r requirements_fill.txt

Usage

The inputs to the tracking algorithm are:

• templates: mean extracellular waveforms (n_channels x n_samples)
• locations: x-y position of electrodes (n_channels x 2)
• fs: sampling frequency (float)

The graph-based method can be run as follows:

import axon_velocity as av

gtr = av.compute_graph_propagation_velocity(template=your_template, locations=your_locations, fs=fs)

To inspect available arguments, you can use av.compute_graph_propagation_velocity?.

The output gtr is an object of a class called GraphAxonTracking. It contains the following fields:

• branches: List of dictionaries containing the following fields:
  • 'selected_channels': selected channels in the path
  • 'velocity': velocity estimate in mm/s (if locations in um and fs in Hz)
  • 'offset': offset of velocity estimate (same units as locations)
  • 'r2': r-squared of the velocoty fit
  • 'error': standard error of the linear fit
  • 'pval': p_value of the fit
  • 'distances': array with distances computed along the branch
  • 'peak_times': array with peak differences with initial channel
  • 'init_channel': channel used as initial channel
• selected_channels: List of selected channels
• graph: NetworkX directed graph

The GraphAxonTracking also implements useful methods for plotting the selected channels (gtr.plot_channel_selection()), plot the underlying graph (gtr.plot_graph()), plot the selected axonal branches (gtr.plot_branches()), and plot the estimated velocities for each branch (gtr.plot_velocities()).

Contribute

Contributions are welcome! Before pushing, make sure to clean up all notebooks with nbconvert:

pip install nbconvert (just once)

jupyter nbconvert --to notebook --ClearOutputPreprocessor.enabled=True --ClearMetadataPreprocessor.enabled=True --inplace **/*.ipynb (before committing)