Challenges and best practices when using ComBAT to harmonize diffusion MRI data

Over the years, ComBAT has become the standard method for harmonizing MRI-derived measurements, with its ability to compensate for site-related additive and multiplicative biases while preserving biological variability. However, ComBAT relies on a set of assumptions that, when violated, can result in flawed harmonization. In this paper, we thoroughly review ComBAT's mathematical foundation, outlining these assumptions, and exploring their implications for the demographic composition necessary for optimal results. Through a series of experiments involving a slightly modified version of ComBAT called Pairwise-ComBAT tailored for normative modeling applications, we assess the impact of various population characteristics, including population size, age distribution, the absence of certain covariates, and the magnitude of additive and multiplicative factors. Based on these experiments, we present five essential recommendations that should be carefully considered to enhance consistency and supporting reproducibility, two essential factors for open science, collaborative research, and real-life clinical deployment.

Authors

Institutions

Publication Details

Journal
Scientific Reports
Published
2025-11-24
DOI
https://doi.org/10.1038/s41598-025-25400-x
Citations
2
Primary Topic
Advanced Neuroimaging Techniques and Applications
Type
article
Field-Weighted Citation Impact
2.72

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Challenges and best practices when using ComBAT to harmonize diffusion MRI data

Donna Masterman, Konstantinos Chiotis, Evan Fletcher, Emily A. Largent et al.
2 citations
Scientific Reports
Advanced Neuroimaging Techniques and Applications
2.72
article

Challenges and best practices when using ComBAT to harmonize diffusion MRI data

Donna Masterman, Konstantinos Chiotis, Evan Fletcher, Emily A. Largent, Winnie Kwang, Joe Strong, Clifford R. Jack, Payam Mahboubi, Catherine Conti, Magdalena Korecka, Alexander Knaack, Juliet Fockler, Duygu Tosun, Jason Karlawish, Nina Silverberg, Renaud La Joie, Sophia I. Thomopoulos, Julien Lagarde, Shannon L. Risacher, Nick C. Fox, Neda Jahanshad, Karen Crawford, Lisa Thomas, Miriam T. Ashford, Andrea Fidell, John Q. Trojanowki, Andrew J. Saykin, Diana Truran‐Sacrey, Erin Franklin, Elizabeth Shaffer, Cecily Jenkins, Arvin Forghanian-Arani, Michael Donohue, Tom Montine, Erin Drake, Yuliana Cabrera, Kristin Harkins, Liana G. Apostolova, John Hsiao, Caitie Hedberg, Danielle Harvey, Charles DeCarli, Ian G. Malone, Perminder Singh Sachdev, Jeremy Pizzola, Naomi Saito, Zhiyun Lin, Prashanthi Vemuri, Jonathan Jackson, Sheila Ogwang, Maxime Descoteaux, Manon Edde, Joel P. Felmlee, Sandhitsu R. Das, Kaci Lacy, Magdalena Brylska, Mark Choe, Matthieu Dumont, Taeho Jo, Matthew L. Senjem, Caileigh Zimmerman, Gil Dan Rabinovici, Taylor Clanton, Bret Borowski, Derek Flenniken, Zaven Kachaturian, Victor L. Villemagne, Kaori Kubo Germano, Laurel Beckett, Joshua D. Grill, Paul A. Yushkevich, Christopher G. Schwarz, Chengshi Jin, Olusegun Adegoke, Stephanie Ruth Rainey‐Smith, Edward B. Lee, David N. Soleimani‐Meigooni, Gabriel Girard, John C. Morris, Sarah Walter, Virginia M.‐Y. Lee, Rema Raman, Brian J. Lopresti, Maison Abu Raya, Kelly Nudelman, Neil Buckholtz, Nigel J. Cairns, Guillaume Theaud, Héctor Alfredo Baptista González, Eliezer Masliah, Robert W. Reid, Talia M. Nir, Agathe Vrillon, Charles Windon, Duygu Tosun, Cat Conti, Denise Reyes, Ronald Petersen, Vanessa Guzmán, Melanie J. Miller, Lindsey Hergesheimer, Rachel L. Nosheny, Félix Dumais, Robert A. Koeppe, Adrienne Kormos, Peter J. Snyder, Pierre‐Marc Jodoin, Kwangsik Nho, Calvin Reyes, Stephanie Rossi Chen, Hannah Craft, Li Shen, Paul M. Thompson, Yang Wan, Robert C. Green, Justin Robison, Michael Walter Weiner, Sandra Talavera, Laurie Ryan, Marı́a C. Carrillo, Arthur W. Toga, Virginia Boatwright, Joseph Di Benedetto, Hannatu Amaza, Omobolanle Ayo, Trinity Weisensel, John Moore-Weiss, Leonard Matoush, Shaniya Parkins, Matt Glittenberg, Paul Aisen, William Z. Potter, William Jagust, Carole Ho, Jean-Christophe Houde, Scott Neu, Tatiana Foroud, Susan Molchan, Fabiola Magana, Alice Hills, R.T. Frank, Richard J. Perrin, Rima Kaddurah-Daouk, Michael Rafii, Isabella Hoang, Adeyinka Ajayi, Mai Seng Thao, Jennifer Salazar, Lisa Taylor‐Reinwald, Chad Ward, Jeff Gunter, Caitlin Webb, Kedir Adem Hussen, Kejal Kantarci, Gustavo Jiménez, Leslie M. Shaw, Kelley Faber, Susan M. Landau, Yoan David, ZoA Potter
article en
2 citations

Abstract

Over the years, ComBAT has become the standard method for harmonizing MRI-derived measurements, with its ability to compensate for site-related additive and multiplicative biases while preserving biological variability. However, ComBAT relies on a set of assumptions that, when violated, can result in flawed harmonization. In this paper, we thoroughly review ComBAT's mathematical foundation, outlining these assumptions, and exploring their implications for the demographic composition necessary for optimal results. Through a series of experiments involving a slightly modified version of ComBAT called Pairwise-ComBAT tailored for normative modeling applications, we assess the impact of various population characteristics, including population size, age distribution, the absence of certain covariates, and the magnitude of additive and multiplicative factors. Based on these experiments, we present five essential recommendations that should be carefully considered to enhance consistency and supporting reproducibility, two essential factors for open science, collaborative research, and real-life clinical deployment.

Scientific ReportsVol. 15(1)
University of Southern California (US), Alzheimer's Association (US), Biogen (United States) (US), Mount Sinai Hospital (US), General Electric (United States) (US), Université de Sherbrooke (CA), University of Wisconsin–Madison (US), University of Connecticut (US), University of California, Los Angeles (US), Fordham University (US), University of Pittsburgh (US), Duke University (US), University of California, San Francisco (US), University of Washington (US), University of California, Irvine (US), Washington University in St. Louis (US), University of Michigan (US), Douglas Mental Health University Institute (CA), Harvard University Press (US), University of California San Diego (US), Massachusetts General Hospital (US), Mount Sinai Medical Center (US), Denali Therapeutics (United States) (US), PC Krause & Associates (United States) (US), Q & T Research (CA), National Institute on Aging (US), Mayo Clinic in Florida (US), National Institute of Mental Health (US), Northern California Institute for Research and Education (US), Indiana University School of Medicine, University College London (GB), Rush University (US), Indiana University – Purdue University Indianapolis (US), Eisai (Japan) (JP), University of Pennsylvania (US), University of California, Davis (US), Philadelphia University (US), University of California, Berkeley (US)
U.S. Department of Defense, Alzheimer's Association, Bristol-Myers Squibb, Eli Lilly and Company, Pfizer, Biogen, University of Southern California, Meso Scale Diagnostics, Alzheimer's Disease Neuroimaging Initiative, BioClinica, Novartis Pharmaceuticals Corporation, Northern California Institute for Research and Education, Université de Sherbrooke, Québec Consortium for Drug Discovery, Eisai, Mitacs, Servier, H. Lundbeck A/S, IXICO, National Institutes of Health, Genentech, Canadian Institutes of Health Research, Natural Sciences and Engineering Research Council of Canada
Openalex Percentile: Top 8%
Advanced Neuroimaging Techniques and Applications
2.72
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.