A0705
Title: Modeling the neural response to head movement
Authors: Martin Lindquist - Johns Hopkins University (United States) [presenting]
Abstract: Head motion during functional magnetic resonance imaging (fMRI) studies can introduce both artifacts and motion-evoked neural responses, potentially biasing inferences about brain activity. Most correction methods treat motion purely as noise, overlooking its possible neural effects. Using large-scale meta-analytic brain maps, the spatial and temporal features of head-motion-evoked BOLD responses are characterized during resting-state scans in healthy young adults. A systematic BOLD increase of 3.6 seconds is observed after motion, particularly in motor-related regions of the brain. Based on this finding, FD-regression is proposed, a method that convolves thresholded framewise displacement (FD) measures with a modeled motion-evoked response to create a nuisance regressor for the GLM. FD-regression more effectively reduced motion-related signal in resting-state data compared to standard motion correction techniques. These findings suggest FD-regression is a promising approach for improving fMRI analyses.
