Quality Assurance Checks

Contributors: Nathan TM Huneke

Maintainers: Nathan TM Huneke

It is important to check the quality of your fMRI data before embarking on analyses to ensure there are no systematic issues with your data.

Standard deviation and Signal to Noise ratio maps

Looking at these maps derived from your functional EPI runs can be helpful for ‘eyeballing’ any clear artefacts or systematic issues with the acquisition of your data.

Todo

Add examples of artefacts on standard deviation and SNR images

Creating standard deviation and tSNR maps

Some scanners can produce these maps for you, however, you might want to create them offline. This can be achieved easily with fslmaths commands. You will need FSL to be installed.

We can therefore use the following commands to create the maps:

# create mean map
fslmaths <input_bold.nii.gz> -Tmean <output_mean.nii.gz>

# create standard deviation map
fslmaths <input_bold.nii.gz> -Tstd <output_std.nii.gz>

# calculate temporal SNR (tSNR)
fslmaths <output_mean.nii.gz> -div <ouput_std.nii.gz> <output_tSNR.nii.gz>

Replace the filenames in < > above with the paths to your input images and chosen output paths.

Using MRIQC for Quality Assurance

MRIQC is a BIDS app that extracts quality metrics from structural and functional MRI data.

It is best to run and install MRIQC using Docker.

Note

If you are using a university machine then you will need to be given permission to use Docker. Please speak to the admin for the machine.

Install MRIQC with Docker

Type the following to download the latest MRIQC image:

docker run -it poldracklab/mriqc:latest --version

Run MRIQC with Docker

Participant-level

To prevent your machine’s RAM from being overloaded, it is best to run 1 participant at a time. Use the following command:

docker run -it --rm -v \
    <path/to/bids/dataset>:/data:ro -v \
    <desired/path/to/outputs>:/out \
    nipreps/mriqc:latest \
    /data /out participant \
    --participant-label 01 \
    --user $( id -u )

Warning

The paths above must be absolute paths.

If you see potential artifacts on the reports, you can re-run the participant and output extra information:

docker run -it --rm -v \
    <path/to/bids/dataset>:/data:ro -v \
    <desired/path/to/outputs>:/out \
    nipreps/mriqc:latest \
    /data /out participant \
    --participant-label 01 \
    --user $( id -u ) \
    --ica --verbose-reports

This will provide a more detailed report as well as the results of independent component analysis.

Group-level

MRIQC can also generate group-level reports to help you identify any outlying participants. Use the following command to run a group-level QC check:

docker run -it --rm -v \
<path/to/bids/dataset>:/data:ro -v \
<desired/path/to/outputs>:/out \
nipreps/mriqc:latest \
/data /out group \
--user $( id -u )