RELIEF

REmoval of Latent Inter-scanner Effects through Factorization

R package to apply RELIEF harmonization to multi-site or multi-scanner neuroimaging studies. This method has been previously called UNIFAC harmonization.

References

Zhang, R., Oliver, L. D., Voineskos, A. N., Park, J. Y. (2023). RELIEF: a structured multivariate approach for removal of latent inter-scanner effects. Imaging Neuroscience. link

Update logs

April 27, 2023: RELIEF method implemented in the package.

Contents

1. Background
2. Installation
3. Usage

---

Background

RELIEF supports harmonizing neuroimaging data collected from multiple sites and/or scanners. It is equivalent to the batch effect correction in genomic studies. Compared to existing approaches such as ComBat (Johnson et al, 2007), RELIEF supports the removal of latent scanner-specific patterns, which results in better data quality after harmonization.

---

Installation

To install the latest development builds directly from GitHub, please run the followings:

if (!require("devtools"))
  install.packages("devtools")
devtools::install_github("junjypark/RELIEF")

Note: If you see an error message, please update your devtools package first.

update.packages("devtools")

After installation, the package can be loaded directly in R.

library(RELIEF)

---

Usage

relief() is the main function that takes the same input names as neuroCombat. Specifically, the following need to be provided.

• dat: a (p x n) data matrix, where p is the number of features and n is the number of subjects (required)
• batch: Batch variable for the scanner id (required)
• mod: a (n x q) matrix containing biological covariates (optional). This can be obtained by using model.matrix() function in R. However, when you use mod in RELIEF, ensure your covariate of interest (for hypothesis testing) is excluded. In practice, RELIEF preserves covariate effects well even though it is not specified as an input.

Now, the relief can be applied to obtain the harmonized data.

relief.harmonized = relief(dat=dat, batch=batch, mod=mod)

---