fMRI Preprocessing Pipelines

Table of contents

1. Overview of Pipelines

Several reproducible processing pipelines are available to for quality control and preprocessing of fMRI and MRI data. Once you have your data in BIDS format, these pipelines can be used to easily prepare your data for analysis using best practices appropriate for most data. The pipelines below are design with fMRI preprocessing below, but usually include anatomical processing (e.g. bias correction, segmentation, FreeSurfer reconstruction) as well.

Quick start guides are provided for:

• mriqc, for the initial quality control of functional and anatomical data.
• fmriprep, for preprocessing functional and anatomical data. fmriprep can be used to prepare functional data for volumetric and/or surface analysis. Optional preprocessing steps include distortion correction with or without fieldmaps, ICA-AROMA denoising, and optimal combination of multi-echo data.
• ciftify, an fmriprep-based pipeline that additionally prepares Human Connectome Project style outputs and quality control of Freesurfer reconstructions. ciftify can be used by itself to generate Human Connectome Project style outputs from BIDS data without a T2w anatomical.
• hcp-bids, the minimal preprocessing pipeline for the Human Connectome Project implemented as a BIDS app. Currently, the PreFreeSurfer, FreeSurfer, PostFreeSurfer, fMRIVolume, and fMRISurface stages are available.

Users may also be interested in

• XCP Imaging Pipeline, for functional BOLD or ASL preprocessing and analysis, particularly resting state connectomics.
• C-PAC, for functional BOLD, particularly resting state connectomics.
• Nipype, for constructing custom pipelines in Python
  • Giraffe tools and PORCUPINE provide graphical tools for creating Nipype pipelines

Overview of Pipelines

---

Table of contents

• MRIQC
• fmriprep
• ciftify
• HCP