SourceLoc

This is still under construction. Once finished it should become a software pipeline to create and maintain a BIDS dataset of clinical MEG resting state measurements and for electromagnetic source imaging in python with free software.

Under the hood it mainly uses

• mne-python
• mne-bids
• nipype
• nilearn

Installation

• clone the git repository (git clone https://github.com/RuKrei/SourceLoc.git)
• install mne (https://mne.tools/dev/install/index.html)
• Activate mne-environment (conda activate name_of_your_mne_environment)
• pip install -r requirements.txt

Configuration

• freesurfer (https://surfer.nmr.mgh.harvard.edu/fswiki/DownloadAndInstall) should be installed on your system for use of anatomical processing steps. The pipeline checks, if SUBJECTS_DIR is set.

• BIDSROOT (root directory for output of the pipeline) and INPUTFOLDER (where MRI and MEEG-data are stored) need to be set. This can be done either when launching a process as arguments, or in form of environment variables.

• OPENMP (the number of processes to use) defaults to 1 and should be set to meet your needs.

• SRCSPACING (https://mne.tools/dev/overview/cookbook.html#setting-up-source-space) should be set on your system, it defaults to "oct6".

Running the pipeline

• put MEEG-data in INPUT_FOLDER, filenames MUST contain the name of the subject
• put MRI-data in a folder named with subjects name in INPUT_FOLDER
• cd to directory of pipeline on harddisk
• python run_pipeline.py
• • if no arguments are given you will be prompted for subject name to process
• • specify INPUT_FOLDER and BIDS_ROOT via one of the options below

BIDS_ROOT and INPUT_FOLDER (mandatory) +

OPENMP and SRCSPACING (optional)

permanent solution --> set shell variables (shown for bash or zsh)

• export BIDS_ROOT=/path/to/store/results/in
• export INPUT_FOLDER=/path/to/input/folder

temporary solution --> provide as arguments

• python run_pipeline.py --bidsroot /path/to/store/results/in --inputfolder /path/to/input/folder --openmp 8 --srcspacing ico4