ComBat Harmonization: Empirical Bayes versus Fully Bayes Approaches

Abstract

Studying small effects or subtle neuroanatomical variation requires large-scale sample size data. As a result, combining neuroimaging data from multiple datasets is necessary. Variation in acquisition protocols, magnetic field strength, scanner build, and many other non-biologically related factors can introduce undesirable bias into studies. Hence, harmonization is required to remove the bias-inducing factors from the data. ComBat, introduced by (Johnson et al., 2007), is one of the most common methods applied to features from structural images. ComBat models the data using a hierarchical Bayesian model and uses the empirical Bayes approach to infer the distribution of the unknown factors. The empirical Bayes harmonization method is computationally efficient and provides valid point estimates. However, it tends to underestimate uncertainty. This paper investigates a new approach, fully Bayesian ComBat, where Monte Carlo Sampling is used for statistical inference. Our experiments show that our new fully Bayesian approach offers more accurate harmonization, unconstrained posterior distributions, and representative uncertainty quantification at the expense of higher computation costs for the inference. This fully Bayesian approach generates a rich posterior distribution, which is also useful for generating simulated imaging features for improving classifier performance in a limited data setting. We show the generative capacity of our model for augmenting and improving the detection of patients with Alzheimer's disease. Posterior distributions for harmonized imaging measures can also be used for brain-wide uncertainty comparison and more principled downstream statistical analysis. Code for our new fully Bayesian ComBat extension is available at https://github.com/batmanlab/BayesComBat.

Fed-ComBat: A Generalized Federated Framework for Batch Effect Harmonization in Collaborative Studies

“…ii) when covariate distributions slightly overlap but follow different patterns, and when there are differences in sample sizes (mid-non-IID), and iii) when there is no overlap between covariate distributions between cohorts and there are evident differences in sample sizes (non-IID). We simulated ComBat parameters following the methodology outlined in Reynolds et al (2022), using the graphical model depicted in Figure 2. The primary difference between our approach and that of Reynolds et al (2022) was the model used to relate the covariates and phenotypes ϕ.…”

“…We simulated ComBat parameters following the methodology outlined in Reynolds et al (2022), using the graphical model depicted in Figure 2. The primary difference between our approach and that of Reynolds et al (2022) was the model used to relate the covariates and phenotypes ϕ. We used a nonlinear function in this step, which was employed as the unbiased target function to be learned after the harmonization process for the synthetically biased phenotypes.…”

MISPEL: A supervised deep learning harmonization method for multi-scanner neuroimaging data

“…Harmonization can be applied to, two broad categories: (1) harmonization of image-derived measures, and (2) harmonization of images. The methods of the first category can be described as ComBat ( Johnson et al, 2007 ) and its extensions ( Beer et al, 2020 ; Chen et al, 2020 ; Pomponio et al, 2020 ; Reynolds et al, 2022 ). ComBat is a location and scale adjustment method used in neuroimaging for harmonizing image-derived measures and has been applied to images of different modalities: DTI ( Fortin et al, 2017 ), MRI ( Fortin et al, 2018 ), and fMRI ( Nielson et al, 2018 ).…”

MISPEL: A supervised deep learning harmonization method for multi-scanner neuroimaging data

“…Depending on the type of data that harmonization can be applied to, it could also fall into two broad categories: (1) harmonization of image-derived measures, and (2) harmonization of images. The methods of the first category can be described as ComBat (Johnson et al, 2007) and its extensions (Beer et al, 2020; Chen et al, 2020b; Pomponio et al, 2020; Reynolds et al, 2022). ComBat is a location and scale adjustment method used in neuroimaging for harmonizing image-derived measures and has been applied to images of different modalities: DTI (Fortin et al, 2017), MRI (Fortin et al, 2018), and fMRI (Nielson et al, 2018).…”