This paper addresses the challenge of harmonizing multi-site diffusion tensor imaging (DTI) data to ensure comparability across different scanners and sites. DTI is a widely used MRI technique for studying white matter microstructure, but it is susceptible to technical variations that can affect the reliability of measurements such as fractional anisotropy (FA) and mean diffusivity (MD). The authors analyze data from two different scanners, showing that DTI measurements are highly site-specific and require harmonization to improve statistical inference. The study evaluates several harmonization methods, including global scaling, functional normalization, RAVEL, Surrogate Variable Analysis (SVA), and ComBat. ComBat, a popular batch-effect correction tool from genomics, is found to be the most effective in removing unwanted site effects while preserving biological variability. The evaluation framework assesses the performance of each method by calculating t-tests to detect site effects and measuring the replicability of voxels associated with age. The results show that ComBat successfully reduces inter-site variability in FA and MD maps, preserving biological variability at each site. It also improves the replicability of voxels associated with age across independent datasets, even in the presence of confounding between age and site. Additionally, ComBat recovers unbiased effect sizes for both signal and null silver-standards, demonstrating its robustness and effectiveness in correcting for site effects. Overall, the study highlights the importance of harmonization in multi-site DTI studies and recommends ComBat as a promising method for improving the reliability and comparability of DTI data.This paper addresses the challenge of harmonizing multi-site diffusion tensor imaging (DTI) data to ensure comparability across different scanners and sites. DTI is a widely used MRI technique for studying white matter microstructure, but it is susceptible to technical variations that can affect the reliability of measurements such as fractional anisotropy (FA) and mean diffusivity (MD). The authors analyze data from two different scanners, showing that DTI measurements are highly site-specific and require harmonization to improve statistical inference. The study evaluates several harmonization methods, including global scaling, functional normalization, RAVEL, Surrogate Variable Analysis (SVA), and ComBat. ComBat, a popular batch-effect correction tool from genomics, is found to be the most effective in removing unwanted site effects while preserving biological variability. The evaluation framework assesses the performance of each method by calculating t-tests to detect site effects and measuring the replicability of voxels associated with age. The results show that ComBat successfully reduces inter-site variability in FA and MD maps, preserving biological variability at each site. It also improves the replicability of voxels associated with age across independent datasets, even in the presence of confounding between age and site. Additionally, ComBat recovers unbiased effect sizes for both signal and null silver-standards, demonstrating its robustness and effectiveness in correcting for site effects. Overall, the study highlights the importance of harmonization in multi-site DTI studies and recommends ComBat as a promising method for improving the reliability and comparability of DTI data.