Poster E100, Monday, March 27, 2:30 – 4:30 pm, Pacific Concourse
Residual relationships between motion and BOLD activity remain after preprocessing and can inflate functional connectivity estimates
Lisa Byrge1, Daniel P. Kennedy1; 1Indiana University
Head motion is problematic for the BOLD signal – particularly for techniques such as functional connectivity MRI (fcMRI), where findings can be spuriously influenced by movement differences. Yet current preprocessing practices might not fully eliminate residual motion influences. Here we present a new method for assessing residual movement-linked BOLD artifact, by quantifying the relationship between movement severity and subsequent BOLD activity. We analyzed resting-state fMRI scans from Indiana University (2 16min scans, TR=813ms, N=53) and the Human Connectome Project (2 14min scans, TR=720ms, N=75). Datasets were preprocessed nearly identically using state-of-the-art cleanup methods (ICA-FIX); alternative preprocessing methods (including/excluding global signal regression) were also examined. We found that movements were systematically linked with structured and prolonged changes in the BOLD signal that depend on the severity of the preceding motion. Nearly all motions (including remarkably small movements below typical censoring thresholds) were associated with structured BOLD changes extending 30s later. Effect sizes of motion-linked BOLD changes were largest approximately 6s and 20s following motion. These patterns were replicated in 4 independent sessions across scanners and preprocessing methods (but not in four different null models). These interactions are not yet addressed by state-of-the-art preprocessing methods, as they persist much later than typical censoring/“scrubbing”. They produce artifactually increased functional connectivity estimates in the epochs following movements, even in strictly censored data. Our results suggest caution in interpreting different patterns of functional connectivity between individuals or groups whose head motion differs, until these interactions between motion and BOLD are fully understood and addressed.
Topic Area: METHODS: Neuroimaging