HCP PIPELINES

This repository contains pipelines used to analyse HCP fMRI data with subject-level analytic pipelines using FSL and SPM and different parameters. It also contains scripts to perform group-level analysis with one sample t-test (within group).

Table of contents

• How to cite?
• Contents overview
• Installing environment
• Download necessary data
• Reproducing subject-level analyses
• Reproducing group-level analyses
• Reproducing technical validation
• Reproducing figures

How to cite?

Contents overview

src

This directory contains scripts and notebooks used to launch the analysis of raw data with pipelines.

data

This directory is made to contain data that will be used by scripts/notebooks stored in the src directory and to contain results of those scripts. For details, check here.

results

This directory contains notebooks and scripts that were used to analyze the results of the experiments. These notebooks were used to evaluate the quality of data obtained from different pipelines.

figures

This directory contains figures and csv files obtained when running the notebooks in the results directory.

Installing environment

To use the notebooks and launch the pipelines, you need to install the NiPype Python package but also the original software package used in the pipeline (SPM, FSL, AFNI...).

To facilitate this step, we created a Docker container based on Neurodocker that contains the necessary Python packages and software packages. To install the Docker image, two options are available.

docker pull elodiegermani/open_pipeline:latest

Download necessary data

Raw fMRI data used in this project are obtained from the Human Connectome Project. All imaging data and most of the behavioral data are Open Access Data, which means that these are available to those who register an account at ConnectomeDB and agree to the Open Access Data Use Terms. This includes agreement to comply with institutional rules and regulations.

After logging in, chose the S1200-Release, click on "Download Images" and select "Subjects with 3T MR Session Data". On the left, chose "Unprocessed" for the Processing Level and keep only "Structural" and "Task" modalities. Select the chosen packages and download.

The original data must be stored in the /data/original repository.

Reproducing subject-level analyses

Subject-level preprocessing and statistical analyses can be launched using the src/run_pipeline.py script. How to use:

python3 run_pipeline.py -e /srv/tempdd/egermani/hcp_pipelines/data/original -r /srv/tempdd/egermani/hcp_pipelines/data/derived -s '["100206"]' -o '["l1"]' -S 'SPM' -t '["MOTOR"]' -c '["rh"]' -f 8 -p 0 -h 'derivatives'

This will perform the l1 analysis of subject 100206 with SPM, for contrast 'rh' (right hand) of MOTOR task. Parameters of the pipelines are: fwhm 8mm, no motion regressors and use of HRF derivatives.

Reproducing group-level analyses

Within-group analysis can be used to obtain group statistic maps of different pipelines and to compare these maps at a higher level. These can be done using the src/run_group_analysis.py script. How to use:

python3 run_group_analysis.py -e /srv/tempdd/egermani/hcp_pipelines/data/derived/subject_level/"$dataset_name"/original -r /srv/tempdd/egermani/hcp_pipelines/data/derived/group_analysis/"$dataset_name" -s '["100206","100307","100410",...]' -c '["rh"]' -n 1000 -i 3

This will perform the within-group analysis for 1000 groups formed from the list of subjects given with i=3 subjects in each group.

Reproducing technical validation

Technical validation is performed to verify that statistic maps produced by the different pipelines are representative of the task of interest.

To do it, you first need to reorganize the data and rename it using results/run_renaming.py. How to use: modify the paths to the input and output dirs and lauch:

python3 run_renaming.py

Technical validation can be done using the results/run_technical_validation.py script. How to use: modify the parameters in the script and launch:

python3 run_technical_validation.py

This will output csv files that can be used to reproduce figures.

Reproducing figures

Figures of technical validation can be reproduce using the Notebook results/technical_validation.ipynb. Do not forget to modify the path to the data.