resting state fMRI preprocessing with Nipype

This repository hosts the Nipype (Gorgolewski et al., 2011) scripted pipeline used to preprocess resting-state fMRI data for the second trial of psilocybin for Major Depressive Disorder from the Imperial College London Centre for Psychedelic Research. This pipeline incorporates the following packages:

FMRIB Software Library (FSL) (Smith et al., 2004; Woolrich et al., 2009)

AFNI (Cox, 1996; Cox & Hyde, 1997; Gold et al., 1998)

Advanced Normalization Tools (ANTS) (Avants et al., 2014)

Details

• rsfmri_preproc.py defines the primary pipeline for preprocessing this data

• dsp_functions.py defines MATLAB signal processing steps to pass sever

• the preprocessing_notebook.ipynb is used to import these and run the workflow

My other work and publications can be found here: Google Scholar

Tutorial

First install the necessary packages from Nipype. Docker is the easiest way to ensure you have the correct UNIX based tools and Nipype version.

1. If your scans are stored in a BIDS formatted directory, this pipeline should work with no modifications (only specifying your paths). However, if your data is not stored in BIDS format, please reformat the 'anatomical' and 'functional_filename' variables in the jupyter notebook to fit the formatting of your filenames.

The following codeblock defines this path if it's the first subject's first scan: ' /sub-ID01/ses-01/anat/sub-ID01_ses-01.nii.gz '

anatomical_filename = opj('sub-ID{subject_id}', 'ses-{ses}', 'anat', 'sub-ID{subject_id}_ses-{ses}*.nii.gz')

2. The subject and ses lists define in the notebook can also be modified depending on your number of subjects and sessions. These numbers will be substituted into the filename {subject_id} and {ses} pathname locations.

subject_list = ['01', '02', '03']  # I have three subjects I wish to iterate through
ses_list = ['1', '2']  # Each subject has two scans I wish to iterate through

3. Once you specify your inputs and outputs, you can pass these variables to the workflow function. This will write your graphs for each sub-workflow and the total workflow in both simple and detailed format

pipeline = preproc_workflow(input_dir=input_dir,
                             output_dir=experiment_dir,
                             anat_file=anatomical_filename,
                             func_file=functional_filename)