neurocombat modified for:

neurocombat_pyClasse

neurocombat function (Fortin, J. P. et al.) implementation in a python classe in order to be compatible with python fit/transform methods and sklearn Pipelines.

original combat function written in python

neurocombat function used in this repository is originally written by Nick Cullen, extended and currently maintained by JP Fortin and can be found in: https://github.com/Jfortin1/neuroCombat

modifications to neurocombat function:

neurocombat integration in a python classe:

References for development of this work:

• Original ComBat model: Johnson, W. E., Li, C. & Rabinovic, A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8, 118–127 (2007).
• First ComBat adaptation to neuroimaging : Fortin, J. et al. NeuroImage Harmonization of multi-site diffusion tensor imaging data. Neuroimage 161, 149–170 (2017).
• First ComBat adaptation for CT: Fortin, J. P. et al. Harmonization of cortical thickness measurements across scanners and sites. Neuroimage 167, 104–120 (2018).
• Using a standard/reference batch/site to estimate ComBat and harmonize data (M-ComBat): Stein, C. K. et al. Removing batch effects from purified plasma cell gene expression microarrays with modified ComBat. BMC Bioinformatics 16, 1–9 (2015).

Cite this repository and respective work:

• ComBat for Machine learning analysis and internal validation frameworks: I. W. Sampaio et al.: "Comparison of Multi-site Neuroimaging Data Harmonization Techniques for Machine Learning Applications," IEEE EUROCON 2023 - 20th International Conference on Smart Technologies, Torino, Italy, 2023, pp. 307-312, doi: 10.1109/EUROCON56442.2023.10198911.