A Simple Intro for fMRIPrep

Liang, Zhang

2024-01-21

BIDS

Brain Imaging Data Structure (BIDS) is a simple and intuitive way to organize and describe your neuroimaging and behavioral data (official site: https://bids.neuroimaging.io/). Why BIDS? (Source: https://rpubs.com/sarenseeley/bids-fmriprep-mriqc)

Heterogeneity in how complex data are organized can lead to confusion (including within-lab as well as between-lab), and unnecessary manual metadata input.
Researchers can take advantage of the numerous and ever-expanding library of “BIDS apps”, or software packages that are written to take valid BIDS datasets as input.
Avoids the need for highly study- or lab-specific pipelines, improves reproducibility for the field as a whole, we can be more confident in our results.
The ability to automatically validate a dataset allows you to spot issues (files missing or in the wrong place, inconsistent naming, etc.) and makes curation easier and faster.
Having a standardized format facilitates data reuse/sharing (benefits for cost-effectiveness of research $).

a heuristic-centric DICOM converter

Essentially, it requires us prepare a so-called heuristic file, which is a python script that tells the tool how to convert the DICOM files to BIDS format.
Typically, we could let the tool generate a template heuristic file for us, and then we can modify it to fit our needs.
It should be very easy, but the details can be tricky. Read these tutorials.

Use docker container (convert to singularity if on server)
Three-step process:
- Do a dry-run of HeuDiConv on a sample subject to generate heuristic file and data sequence info files
- Modify heuristic file based on the DICOM characteristics listed in .heudiconv/dicominfo.tsv file
- Run HeuDiConv using the revised heuristic file
The heuristic file requires some python programming skills
Plan the protocol names when you collect the data

Overview of the whole picture

A BIDS derivatives dataset of the form:

<output_dir>/
  logs/
  sub-<label>/
  sub-<label>.html
  dataset_description.json

Derivatives data is in sub-<label>/ folder, including anatomical, freesurfer ¹ and functional derivatives
Visual report is genrated as sub-<label>.html, here is a sample report
dataset_description.json is a a metadata file in which fMRIPrep records metadata recommended by the BIDS standard
The confounds of each BOLD results are also stored in derivatives data, but note fMRIPrep only generates them, not uses them to denoise data

Use the newest stable version because it is constantly updated
Download TemplateFlow data beforehand to avoid network issues
Run each subject separately and run all in parallel
Possible issue: now --use-aroma is removed, ICA is not done (maybe not so necessary)