Advancing motion denoising of multiband resting-state functional connectivity fMRI data

Area of Science:

• Neuroimaging
• Functional Magnetic Resonance Imaging (fMRI)
• Brain Connectivity

Background:

• Simultaneous multi-slice (multiband) accelerated fMRI enhances temporal and spatial resolution for resting-state functional connectivity (RSFC) studies.
• Subject motion is a significant confound in RSFC research, reducing data reliability and reproducibility.
• Existing motion denoising evaluation metrics may be problematic for assessing pipeline performance.

Purpose of the Study:

• To comprehensively evaluate existing and novel volume censoring-based motion denoising approaches for multiband fMRI.
• To develop and validate new quantitative metrics for motion denoising evaluation, independent of problematic assumptions.
• To provide optimized, dataset-specific methods for motion denoising parameter selection in RSFC analysis.

Main Methods:

• Evaluation of motion denoising pipelines using the Human Connectome Project (HCP) dataset.
• Development of novel quantitative metrics for evaluating denoising performance, agnostic to quality control-functional connectivity (QC-FC) correlations.
• Validation of methods for determining optimal dataset-specific volume censoring parameters.

Main Results:

• Commonly used metrics (e.g., QC-FC correlations) for evaluating motion denoising pipelines exhibit problematic assumptions.
• Novel quantitative metrics were developed and demonstrated as effective benchmarks for comparing volume censoring methods.
• Quantitative methods for optimizing volume censoring parameters were developed and validated.

Conclusions:

• Existing metrics for evaluating motion denoising in RSFC are potentially inappropriate.
• New, robust quantitative metrics are proposed for reliable assessment of motion denoising pipelines.
• Optimized, dataset-specific motion denoising strategies can enhance the reliability of RSFC findings.