A New Computational Method for Improving Brain-Behavior Correlation Associations
posted October 22, 2025
Researchers studying brain function are interested in functional connectivity (FC), which is the correlation between brain regions that share functional properties. Structural and functional MRI (fMRI) are extremely useful in these studies, but a patient’s head motion can cause artifacts resulting in erroneous conclusions. Even tiny, involuntary movements can yield these artifacts, and post-processing to remove them is difficult.
Several MSI PIs and researchers, Research Computing staff members, and colleagues at the Washington University School of Medicine (St. Louis, Missouri), the National Institute of Mental Health (Bethesda, Maryland), the University of California San Diego (San Diego, California), and Oregon Health and Science University (Portland, Oregon) have developed the Split Half Analysis of Motion Associated Networks (SHAMAN), a computational method that assigns a motion-impact score to specific trait-FC relationships. Using SHAMAN allows researchers to differentiate between motion that causes over- or under-estimation of trait-FC effects. The researchers used data from over 7,000 participants the Adolescent Brain Cognitive Development (ABCD) Study, and high-performance computing resources at both the Minnesota Supercomputing Institute and the Pittsburgh Supercomputing Center.
MSI-associated participants in this study include:
MSI PIs:
- Brenden Tervo-Clemmens (Psychiatry and Behavioral Sciences; Masonic Institute for the Developing Brain; Institute for Translational Neuroscience)
- Eric Feczko (Pediatrics, Masonic Institute for the Developing Brain)
- Anita Randolph (Pediatrics, Masonic Institute for the Developing Brain)
- Óscar Miranda-Domínguez (Pediatrics, Masonic Institute for the Developing Brain)
- Stephen M. Malone (Psychology)
- Damien A. Fair (Pediatrics; Co-Director, Masonic Institute for the Developing Brain; Institute of Child Development)
RC Staff:
- Benjamin J. Lynch (Director, Minnesota Supercomputing Institute)
- James C. Wilgenbusch (Director, Research Computing)
- Thomas Pengo (co-Director, Research Informatics, Minnesota Supercomputing Institute)
The paper can be read on the journal website: Kay, Benjamin P., et al., Motion impact score for detecting spurious brain-behavior associations. Nature Communications 16: 8614 (2025). doi: 10.1038/s41467-025-63661-2.
Image description: Trait-specific impact of motion on functional connectivity (FC). Top: Parcel-wise FC, computed as the root mean square (RMS) of connectivity values for each parcel/node in the trait-FC effect matrix for (a) body mass index (BMI) and (b) WISC-V matrix reasoning score. Bottom: Motion impact score (omnibus Stouffer’s Z, higher = more motion) for (c) BMI and (d) WISC-V. Motion over-estimation scores are labeled “Over” in orange and motion underestimation scores are labeled “Under” in blue. All measures were computed after motion processing with ABCD-BIDS and without frame censoring (n = 7270). Image and description: Kay, Benjamin P., et al. NatComms 16: 8614 (2025). doi: 10.1038/s41467-025-63661-2.
