Functional connectivity MRI quality control procedures in CONN.

Area of Science:

• Neuroimaging
• Functional Connectivity Magnetic Resonance Imaging (FC-MRI)
• Data Quality Control

Background:

• Quality control (QC) is crucial for the validity and replicability of functional connectivity magnetic resonance imaging (FC-MRI) studies.
• Noise and artifacts in MRI data can introduce biases, compromising the interpretation of functional connectivity measures.
• Current QC practices in FC-MRI are often underutilized and lack systematic reporting.

Purpose of the Study:

• To describe and demonstrate a comprehensive quality control (QC) pipeline for evaluating MRI data within the CONN toolbox.
• To implement both visual and automated QC procedures for resting-state fMRI data.
• To establish best practices for QC reporting in FC-MRI studies.

Main Methods:

• Utilized publicly available resting-state fMRI data from the FMRI Open QC Project (N=139).
• Applied standard preprocessing steps including realignment, normalization, segmentation, and smoothing.
• Implemented denoising strategies combining scrubbing, motion regression, and aCompCor, alongside participant-level and dataset-level QC metrics.

Main Results:

• Developed and demonstrated a QC pipeline integrating visual inspection and automated metrics (e.g., framewise displacement, global signal change, connectivity strength).
• Evaluated inter-subject variability in functional connectivity and assessed associations between connectivity and noise indicators.
• Provided visualizations of QC procedures and results on a reference dataset.

Conclusions:

• The described QC pipeline enhances the systematic evaluation of FC-MRI data quality.
• Standardized QC procedures are vital for improving the validity, replicability, and interpretability of neuroimaging research.
• This work aims to promote the dissemination and standardization of QC reporting in the scientific community.