Welcome to PyReliMRI’s documentation!
Python-based Reliability in MRI (PyReliMRI) is an open-source Python tool to calculate multiple group- and individual-level reliability metrics. This package is designed for researchers using MRI data to easily report reliability estimates in their manuscripts, particularly for multi-run and/or multi-session data.
Several packages exist to address different aspects covered in this package. Specifically, tools are available for calculating similarity coefficients, intraclass correlations (e.g., 3dICC in AFNI), or both (e.g., in Python-based nipype or Matlab’s fmreli). Alternatively, Ting Xu offers a Shiny App for calculating univariate and multivariate ICCs from .csv data. However, some flexibility may be limited or certain features unavailable when working with preprocessed MRI data. For example, ICC_rep_anova is restricted to ICC(3,1), and fmreli requires a Matlab license and does not support tetrachoric correlation, pairwise comparisons across images, or atlas-based reliability estimates.
Our goal is to integrate various functions (see Figure 1) into a single package that can be easily downloaded and imported into Python for universal use.
Figure 1. Functions within the PyReliMRI Library
PyReliMRI
Module Documentation
Usage Examples
Examples:
Citing PyReliMRI
Demidenko, M., Mumford, J., & Poldrack, R. (2024). PyReliMRI: An Open-source Python tool for Estimates of Reliability in MRI Data (2.1.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.12522260