stefanches7 / gift-bids

GIFT BIDS-App

GIFT under BIDS-Apps using Docker

Table of Contents

1. Introduction
2. Regular GIFT
3. Usage
4. Demo
5. Version

Introduction

This BIDS-App (gift-bids) works after Docker is installed and does not need any other application such as MATLAB since a compiled MATLAB engine is used. The gift-bids app implements multiple algorithms for independent component analysis and blind source separation of group (and single subject) functional magnetic resonance imaging data.

Regular GIFT

If you have MATLAB installed on your computer and you are not interested in running GIFT under a computer cluster you may want to try our regular GIFT found here. Our regular GIFT may be easier to learn as it has an intuitive user interface.

Usage

After Docker is properly installed you can download your gift-bids image from the internet (https://hub.docker.com/u/trends) by following:

$ docker pull trends/gift-bids

To run gift-bids you need:

1. A directory with image files in accordance with the BIDS format (my_data)
2. An empty output directory (gift_out) you may create
3. A configuration directory with your configuration file and change the syntax of ZZZ in accordance with box below and your configuration file.
4. Choose subject(s) you want to process, by replacing XX YY NN with subject number(s) in box below (e.g., 01 02 ... NN) and you may run:

$ docker run -ti --rm \
  -v /tmp:/tmp \
  -v /var/tmp:/var/tmp \
  -v /path/to/local/bids/input/my_data/:/data \
  -v /path/to/local/gift_out/:/output \
  -v /path/to/local/cfg/:/cfg \
  trends/gift-bids \
    /data /output participant --participant_label XX YY NN \
    --config /cfg/config_ZZZ.m

Demo

A demo (gift-bids-demo.sh) is available, using a public fMRI dataset and the fMRIPrep BIDS app to preprocess and smooth the fMRI data. In 6 simple steps you may run the entire demo:

1. Install Docker (https://www.docker.com/products/personal).
2. Open terminal, create and go to working directory of your choice.
3. Clone or unzip the https://github.com/trendscenter/gift-bids repository to the root of your working directory.
4. Download the ds005.tar dataset from https://drive.google.com/drive/folders/0B2JWN60ZLkgkMGlUY3B4MXZIZW8?resourcekey=0-EYVSOlRbxeFKO8NpjWWM3w into the demo directory. This ds005.tar dataset is described at the BIDS App tutorial web page (https://bids-apps.neuroimaging.io/tutorial).
5. If you do not have a free FreeSurfer license you need to apply for the license file by registering according with the link https://surfer.nmr.mgh.harvard.edu/registration.html and then place your FreeSurfer license.txt file into the demo directory.
6. Finally this script (gift-bids-demo.sh) may be run in terminal by:

$ cd demo
$ chmod +x gift-bids-demo.sh
$ ./gift-bids-demo.sh

After this Docker may download images for fmriprep and trends/gift-bids images (if needed)). It will then run fmriprep for a single subject which may take 2h or more. Finally trends/gift-bids will run for 5 min or more. Results are found under demo/ds005-gift-out directory

Version

GIFT-BIDS-Apps-Scripts=1.000 GIFT-Linux-Compile=1.000