comet-toolbox 0.3.7

Dynamic functional connectivity toolbox for multiverse analysis

Comet - A dynamic functional connectivity toolbox for multiverse analysis

Important notes:

• This package is at an early stage of development, with frequent changes being made. If you intend to use this package at this stage, I kindly ask that you contact me via the email address in the pyproject.toml file.
• Many features are not yet tested, so there will be bugs (the question is just how many). A comprehensive testing suite and documentation will be added in the near future

Current Features

Installation

It is recommended to use a dedicated Anaconda or Miniconda environment to mitigate the risk of potential version conflicts:

conda create -n comet python==3.11
conda activate comet

Installation is then possible through the Python Package Index (PyPI) with the pip or pip3 command, depending on your system:

pip install comet-toolbox

Installation from the source code of this repository is also possible:

1. Download/clone the repository
2. Open a terminal in the folder which contains the pyproject.toml file
3. Install the package via pip (installing in editable mode (-e) is a helpful approach if you intend to modify the source code):

pip install -e .

Usage

General

The toolbox is designed in a modular way, which means you can use the individual parts in combination with others, but also by themselves.

• continuous and static dFC measures require 2D time series data (n_timepoints x n_regions) as input
• state-based dFC methods require a TIME_SERIES object (as used in the pydfc toolbox) containing data for multiple subjects as input
• Graph measures need 2D adjacency/connectivity matrices as input
• Multiverse analysis needs decision/option pairs of any kind to create forking paths in the analysis as well as a template script for the analysis

GUI

After installation, you can use the graphical user interface through the terminal by typing:

comet-gui

If you want to explore the toolbox with example data, you can load data included in the tutorials/example_data/ folder:

• simulation.txt contains simulated BOLD data for 10 brain regions with 2 changing brain states (usable for continuous and static dFC measures)
• abide_50088.txt contains parcellated BOLD data for a single subject from the ABIDE data set (usable for continuous and static dFC measures)
• aomic_multi.pkl contains parcellated BOLD data for 5 subjects from the AOMIC data set (usable for state-based dFC measures)

Scripting

If you intend to use the toolbox in a standard python script, demo scripts are provided as a starting point:

• Demo script for calculating dFC: click here
• Demo script for performing multiverse analysis: click here
• Demo script for the multiverse analysis as presented preprint (+ additional visualizations): click here
• Demo script for graph analysis: click here

Feedback and contribution

If you have any wishes, suggestions, feedback, or encounter any bugs, please don't hesitate to contact me via email or create an issue here. Contributions or future collaboration are also welcome.