Robust-ComBat: Mitigating Outlier Effects in Diffusion MRI Data Harmonization
2026-03-18 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors studied ways to adjust brain scan data from different hospitals to remove site-related differences. They found that the common method, ComBat, assumes normal data but struggles when patients have brain disorders causing unusual scan results. Including these unusual cases leads to errors in harmonization. The authors tested many methods and discovered that using a simple neural network (MLP) works better at handling these outliers, keeping important disease information intact. Their new approach, Robust-ComBat, improved accuracy in data adjustment even when most subjects had neurological disorders.
Diffusion MRI (dMRI)ComBat harmonizationSite-effect biasOutlier rejectionNeurological disordersMultilayer Perceptron (MLP)Data harmonizationPathological outliersNormative referenceMulti-site cohort
Authors
Yoan David, Pierre-Marc Jodoin, Alzheimer's Disease Neuroimaging Initiative, The TRACK-TBI Investigators
Abstract
Harmonization methods such as ComBat and its variants are widely used to mitigate diffusion MRI (dMRI) site-specific biases. However, ComBat assumes that subject distributions exhibit a Gaussian profile. In practice, patients with neurological disorders often present diffusion metrics that deviate markedly from those of healthy controls, introducing pathological outliers that distort site-effect estimation. This problem is particularly challenging in clinical practice as most patients undergoing brain imaging have an underlying and yet undiagnosed condition, making it difficult to exclude them from harmonization cohorts, as their scans were precisely prescribed to establish a diagnosis. In this paper, we show that harmonizing data to a normative reference population with ComBat while including pathological cases induces significant distortions. Across 7 neurological conditions, we evaluated 10 outlier rejection methods with 4 ComBat variants over a wide range of scenarios, revealing that many filtering strategies fail in the presence of pathology. In contrast, a simple MLP provides robust outlier compensation enabling reliable harmonization while preserving disease-related signal. Experiments on both control and real multi-site cohorts, comprising up to 80% of subjects with neurological disorders, demonstrate that Robust-ComBat consistently outperforms conventional statistical baselines with lower harmonization error across all ComBat variants.