catchaMouse16: CAnonical Time-series CHaracteristics for fMRI (trained on mouse data)

This is a collection of 16 time series features contained in the hctsa toolbox coded in C that can be run from Python, MATLAB and C, licensed under the GNU GPL v3 license (or later). Features were selected from the total set of hctsa features by their classification performance across a collection of mouse fMRI time-series classification problems. See the pipeline and paper specific repo for details.

📒📘📗catchaMouse16 documentation

There is comprehensive documentation for catchaMouse16, including:

• Installation instructions (across C, python, and Matlab)
• Information about the theory behind and behavior of each of the features,
• A list of publications that have used the catchaMouse16 feature set or specific time series pipeline
• And more 😋

Acknowledgement 👍

If you use this software, please read and cite this open-access article:

• 📗 Canonical time-series features for characterizing biologically informative dynamical patterns in fMRI, bioRxiv (2024).

Performance Summary

Summary of the performance of the catch22 feature set across 93 classification problems, and a comparison to the hctsa feature set (cf. Fig. 4 from [our paper][LINK TO BE ADDED]:

Contribution

Please feel free to ask any query in the Issue section. Any suggestions or improvements are welcomed through Pull Requests.