[MRIQC 1] MRIQC: Magnetic Resonance Imaging Quality Control
19 May 2024 #bio #brainImaging
To advance research on MRI images and enhance quality, it is essential to check the condition of the image data and secure high-quality data. However, assessing MRI quality is challenging due to several factors. There are many types of artifacts that can occur during MRI scans, people evaluate image quality differently, and some artifacts are difficult for humans to detect. In this context, an objective MRI quality control (QC) system can be helpful in the early stages of MRI quality assessment. Additionally, the recent trend of acquiring very large image data samples from multiple scanning sites increases the need for fully automated and minimally biased QC protocols.
Magnetic Resonance Imaging Quality Control (MRIQC)
MRIQC (Magnetic Resonance Imaging Quality Control) can be used as an automated tool for assessing MRI quality. MRIQC is an open-source tool designed to evaluate the quality of structural(anatomical) and functional MRI images. MRIQC extracts image quality metrics (IQMs) solely from the input images themselves, without referencing any target images. Additionally, it provides a standardized method for evaluating and comparing MRI scans from various sources or sessions.
Priciples
- Modular and Integrable: MRIQC uses a modular workflow built on the Nipype framework, integrating various third-party software toolboxes such as ANTs and AFNI.
- Minimal Preprocessing: It focuses on minimal preprocessing to estimate IQMs from the original or minimally processed data, ensuring that the quality metrics reflect the raw image data as closely as possible.
- Interoperability and Standards: MRIQC adheres to the Brain Imaging Data Structure (BIDS) standard, promoting interoperability and facilitating integration into various neuroimaging workflows.
- Reliability and Robustness: The tool is rigorously tested for robustness against data variability, ensuring consistent performance across different datasets and acquisition parameters.
- Visual Reports: MRIQC generates detailed visual reports for both individual images and group analyses. These reports include mosaic views and segmentation contours for individual images, and scatter plots for group analyses to identify outliers.
Image Quality Metrics (IQMs)
MRIQC computes a range of IQMs categorized into four main groups:
- Noise-related metrics: Evaluate the impact and characteristics of noise within the images.
- Information theory-based metrics: Assess the spatial distribution of information using prescribed masks.
- Artifact detection metrics: Identify and measure the impact of specific artifacts, such as inhomogeneity and motion-related signal leakage.
- Statistical and morphological metrics: Characterize the statistical properties of tissue distributions and the sharpness/blurriness of images.
Paper
Esteban O, Birman D, Schaer M, Koyejo OO, Poldrack RA, Gorgolewski KJ (2017) MRIQC: Advancing the automatic prediction of image quality in MRI from unseen sites. PLoS ONE 12(9): e0184661. https://doi.org/10.1371/journal.pone.0184661
How to run MRIQC
Interested in running MRIQC? Check out this post for detailed instructions.