Skip to content

Correcting Covariance Batch Effects (CovBat): Harmonization of mean and covariance for multi-site data

Notifications You must be signed in to change notification settings

brainspinner/CovBat_Harmonization_Python_Release

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

CovBat_Harmonization

Correcting Covariance Batch Effects (CovBat): Harmonization of mean and covariance for multi-site data


Fork Maintainer: Candace Makeda Moore, [email protected]

Original Library Author: Andrew Chen, [email protected]

License: Artistic License 2.0

Table of content

1. Fork background

There was an R package, but to get it go back to the library this was forked from. We made this fork to run Python. We therefore chaned the directories of some files. The goal was reproducible code.

2. Science Background

According to the original library readme: "Current harmonization methods often focus on addressing scanner differences in the mean and variance of features. However, machine learning methods employed in multivariate pattern analysis (MVPA) are known to leverage additional properties of the data, including covariance. In our recent paper, we show that ComBat, a state-of-the-art method designed to harmonize mean and variance, is unable to fully prevent detection of scanner manufacturer through MVPA in the Alzheimer's Disease Neuroimaging Initiative data. We design CovBat to harmonize the covariance of multivariate features and show that it can almost fully prevent detection of scanner properties.

CovBat is meant to be applied after initial preprocessing of the images to obtain a set of features and before statistical analyses. The application of CovBat is not limited to neuroimaging data; however, it has yet to be tested in other types of data."

3. Software

The R implementation of CovBat is based on the ComBat package. The Python implementation of CovBat is a modification of the ComBat package for Python here. This is a fork of covbat created for reproducibility.

4. Citation

If you are using CovBat, cite the following article:

Chen, A. A., Beer, J. C., Tustison, N. J., Cook, P. A., Shinohara, R. T., Shou, H., & Initiative, T. A. D. N. (2022). Mitigating site effects in covariance for machine learning in neuroimaging data. Human Brain Mapping, 43(4), 1179–1195. https://doi.org/10.1002/hbm.25688)

About

Correcting Covariance Batch Effects (CovBat): Harmonization of mean and covariance for multi-site data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • R 56.7%
  • Python 43.3%