Reduction of motion-related artifacts in resting state fMRI using aCompCor

By John Muschelli, Mary Beth Nebel, Brian C. Caffo, A. Barber, James J. Pekar, Stewart H. Mostofsky. NeuroImage, 2014

431 citations recorded by the source corpus.

Recent studies have illustrated that motion-related artifacts remain in resting-state fMRI (rs-fMRI) data even after common corrective processing procedures have been applied, but the extent to which head motion distorts the data may be modulated by the corrective approach taken. We compare two different methods for estimating nuisance signals from tissues not expected to exhibit BOLD fMRI signals of neuronal origin: 1) the more commonly used mean signal method and 2) the principal components analysis approach (aCompCor: Behzadi et al., 2007). Further, we investigate the added benefit of “scrubbing” (Power et al., 2012) following both methods. We demonstrate that the use of aCompCor removes motion artifacts more effectively than tissue-mean signal regression. In addition, inclusion of more components from anatomically defined regions of no interest better mitigates motion-related artifacts and improves the specificity of functional connectivity estimates. While scrubbing further attenuates motion-related artifacts when mean signals are used, scrubbing provides no additional benefit in terms of motion artifact reduction or connectivity specificity when using aCompCor.

Subfields: Cognitive Neuroscience, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/04ba86ec33e1f0486e2ba4278529af8a579ddb33

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

1. Olaf Sporns
2. Yong He
3. D. Bassett
4. Alan C. Evans
5. Jinhui Wang
6. E. Bullmore
7. Xi-Nian Zuo
8. Alex Fornito

In Radiology, Nuclear Medicine and Imaging

1. D. Bassett
2. Olaf Sporns
3. Yong He
4. Jinhui Wang
5. Jonathan D. Power
6. Alan C. Evans
7. Alex Fornito
8. Xi-Nian Zuo

Citations in the corpus

• Data Quality Influences Observed Links Between Functional Connectivity and Behavior (2017, methodology): “found that some distance-related motion artifact persisted following CompCor regression, but was removed when scrubbing was performed in addition to CompCor (Muschelli et al. 2014).”
• Recent progress and outstanding issues in motion correction in resting state fMRI (2014, neutral): “Note that, within a single dataset, measured magnitudes of FD FSL are roughly half the magnitudes of FD Power . effects (since different signal disruptions could partially or completely cancel each other in an average signal calculation (Jo et al., 2013; Muschelli et al., 2014)).”
• GRETNA: a graph theoretical network analysis toolbox for imaging connectomics (2015, background): “There are many methodological challenges that remain elucidative, such as head motion correction (Muschelli et al., 2014; Patel et al., 2014), null model construction (Zalesky et al.”
• An evaluation of the efficacy, reliability, and sensitivity of motion correction strategies for resting-state functional MRI (2017, methodology): “4 aCompCor A popular method for modelling noise in BOLD data is to apply temporal principal component analysis (PCA) to putative nuisance signals (Behzadi et al., 2007; Muschelli et al., 2014).”
• The Global Signal in fMRI: Nuisance or Information? (2017, positive): “In comparing these approaches, Muschelli et al. (2014) found that the component-based approach performed better than the mean signal approach, partly reflecting the fact that component-based approaches can preserve information that may otherwise be averaged out when using the mean signal.”
• Benchmarking of participant‐level confound regression strategies for the control of motion artifact in studies of functional connectivity (2017, methodology): “We extracted 5 components for WM and CSF each, yielding 10 compcor components (Muschelli et al., 2014).”
• Evaluation of ICA-AROMA and alternative strategies for motion artifact removal in resting state fMRI (2015, methodology): “As proposed by Muschelli et al. (2014), we constructed a nuisance model comprising the 24 realignment parameters complemented by the principle components that explained 50% of respectively the WM and CSF signals.”