Welcome to MEGPrep’s documentation!#
MEGPrep is a fully automated preprocessing pipeline for MEG (Magnetoencephalography) data, built on the MNE-Python framework and leveraging the power of Nextflow. It is specifically designed to address the challenges of large-scale MEG data processing with a strong emphasis on reproducibility, efficiency, and user-friendliness in various research environments.
Features#
MEGPrep ensures reliable and robust MEG data processing. Standardized environments through containerization, using Docker and Singularity, guarantee consistent results across computational setups. This minimizes variability and ensures reproducibility across different systems, facilitating cross-subject and cross-site studies of MEG data.
MEGPrep is designed with modularity in mind, allowing users to customize their preprocessing workflows easily. It integrates seamlessly with various libraries, including mne-python, enhancing its functionality for processing and analyzing MEG data.
By using the Nextflow framework, MEGPrep dramatically accelerates every step of the preprocessing pipeline. It is optimized for high parallelization, capable of managing heavy workloads and significantly speeding up data processing through concurrent execution of tasks. This capability is essential for conducting large benchmarks and effective comparisons across various tools and methodologies.
MEGPrep includes an interactive reporting feature based on Streamlit, allowing users to visualize quality control metrics at each processing step. These reports provide alerts for any anomalies detected in the data, ensuring that the quality of each stage of processing is maintained.
MEGPrep offers an easy-to-use configuration system, allowing users to specify parameters simply and intuitively. This configurability empowers researchers to adapt the preprocessing pipeline to their unique datasets and experimental needs without complex coding.
MEGPrep enhances automated detection processes, including Automatic Artifacts Rejection, ICA (Independent Component Analysis) Automatic Detection, and auto-coregistration. These automated features streamline the preprocessing steps, reduce manual intervention, and improve accuracy and efficiency in artifact handling and data integration.