ComBat-MLE: A Novel Perspective on Developing and Solving ComBat Harmonization Method

This paper provides a novel perspective on the model development of the ComBat method. Two sets of parameter constraints were introduced to help identify parameters during the parameter estimation process. And Maximum Likelihood Estimation (MLE) was employed to solve the re-formulated ComBat using each set of constraints. The simpler and faster solution was selected as the proposed ComBat-MLE harmonization method. A series of experiments were conducted to evaluate the effectiveness of various ComBat variants, both quantitatively and qualitatively. The experiments validated the effectiveness of the proposed ComBat-Mle method in removing non-biological scanner effects while at the same time preserving intrinsic biological information. Although the proposed ComBat-MLE method did not significantly outperform existing ComBat variants, it aids in understanding the ComBat method and offers valuable insights into the development of new harmonization methods.