MRIQC

MRIQC: Quality control software for multimodal MRI data. MRIQC extracts no-reference IQMs (image quality metrics) from structural (T1w and T2w) and functional MRI (magnetic resonance imaging) data.

The purpose of the MRIQC project is to build a tool for extracting quality measures and fitting a binary (accept/exclude) classifier. This removes the need for subjective decision-making when deciding what scans are 'good enough' to be included in further analyses. You can read more details about the project in the original paper in PLOS ONE.

NSB Lab Pipeline

The lab has developed an adapted QC pipeline that includes MRIQC’s IQMs. This is a comprehensive set of steps for processing your MRI data. The scripts are available in the MRI_QC folder on the repository. You should read each script before running it, making sure you change any parameters (e.g. file locations, number of subjects) as needed. You can read more about MRIQC below.

Running MRIQC

The tool can be run both locally and as a free online service via the OpenNeuro.org portal. MRIQC can be run on any BIDS-compatible dataset. It is run using bash (although the source code is written in Python).

The MRIQC documentation has examples of how to run their pipeline for BIDS-organised data. Have a look here for a comprehensive tutorial. As an example:

• mriqc bids-root/ output-folder/ participant --participant-label S01 S02 S03 runs MRIQC only on the subjects indicated.
• mriqc bids-root/ output-folder/ participant runs MRIQC on all the T1w and BOLD images found under the BIDS-compliant folder bids-root/.
• mriqc bids-root/ output-folder/ group generates the group level results (the group level report and the features CSV table).

MRIQC Outputs

IQMs

Some no-reference IQMs are extracted in the final stage of all processing workflows run by MRIQC. A no-reference IQM is a measurement of some aspect of the actual image which cannot be compared to a reference value for the metric since there is no ground-truth about what this number should be.

The IQMs can be grouped in four broad categories, providing a vector of 56 features per anatomical image.

-- MRIQC Documentation

Visual reports

In order to ease the screening process of individual images, MRIQC generates individual reports with mosaic views of a number of cutting planes and supporting information (for example, segmentation contours). The most straightforward use-case is the visualization of those images flagged as low-quality by the classifier.

After the extraction of IQMs in all the images of our sample, a group report is generated. The group report shows a scatter plot for each of the IQMs, so it is particularly easy to identify the cases that are outliers for each metric. The plots are interactive, such that clicking on any particular sample opens the corresponding individual report of that case. Examples of group and individual reports for the ABIDE dataset are available online at mriqc.org.

-- MRIQC Documentation

Extra references

Here is a review on doing QC with fMRIprep and MRIQC