Characterization and Mitigation of a Simultaneous Multi-Slice fMRI Artifact: Multiband Artifact Regression...

Area of Science:

• Neuroimaging
• Magnetic Resonance Imaging (MRI)

Background:

• Simultaneous multi-slice (multiband) acceleration is a common technique in functional MRI (fMRI).
• This acceleration method may introduce novel signal artifacts.
• A previously unreported shared signal artifact between simultaneously acquired slices in multiband fMRI data has been observed.

Purpose of the Study:

• To demonstrate and characterize a novel artifact in multiband fMRI data.
• To propose and validate a new method for detecting and correcting this artifact.
• To evaluate the effectiveness of the proposed method in improving fMRI data quality and subsequent analysis.

Main Methods:

• Investigation of resting-state and task-based multiband fMRI datasets, including Human Connectome Project (HCP) and Adolescent Brain Cognitive Development (ABCD) Study data.
• Development and application of Multiband Artifact Regression in Simultaneous Slices (MARSS), a regression-based artifact detection and correction technique.
• Comparison of MARSS with and in tandem with sICA+FIX denoising methods.

Main Results:

• MARSS successfully mitigates a shared signal artifact present in unprocessed multiband fMRI data.
• The corrected signal exhibits characteristics suggesting it is nonneural, with higher prevalence in neurovasculature than gray matter.
• MARSS reduces residual artifact signal not captured by sICA+FIX, leading to improved signal-to-noise ratio and reduced coefficient of variation.
• MARSS correction mitigates artefactual spatial patterns in task betas and impacts second-level statistical analyses.

Conclusions:

• MARSS is an effective method for detecting and correcting a novel artifact in multiband fMRI.
• The method improves data quality, enhances signal-to-noise ratio, and refines statistical outcomes in fMRI analyses.
• MARSS is recommended for use with multiband fMRI datasets, particularly those with moderate to high acceleration factors, alongside existing denoising techniques.