Key points are not available for this paper at this time.
Electroencephalography (EEG) is a powerful tool for investigating neural processes underlying cognition and neuropsychiatric disorders. Yet, variability in EEG preprocessing strategies restricts reproducibility and data integration across study sites and laboratories, particularly in larger research consortia. This paper introduces the CLEAN-EEG preprocessing pipeline, designed to standardize data processing and documentation across multiple sites. The CLEAN pipeline is implemented in MATLAB using EEGLAB. It comprises three modular, script-based stages: main preprocessing (including down-sampling, filtering, line noise removal, and channel interpolation), independent component analysis preparation and decomposition with flexible options for artifact rejection or neural component extraction, and component exclusion with support for automated classification and dipole fitting. Emphasis is placed on transparency through comprehensive logging and quality-control plotting, as well as on minimizing rank reduction to preserve data suitability for advanced analyses such as source localization and connectivity modeling. By providing clear, adaptable recommendations while ensuring detailed documentation of every step, the CLEAN pipeline aims to harmonize EEG preprocessing in large-scale, multi-center studies. This open and reproducible approach facilitates high throughput analyses, supports the training of researchers, and enables the rigorous integration of neurophysiological data across study sites, study designs, and populations.
Böttcher et al. (Sat,) studied this question.