Model-based attenuation of movement artifacts in fMRI

▶ Motor fMRI is invariably affected by movement artifacts. Most algorithms in literature were designed for slow changes of head position over several brain scans. ▶ This paper introduces a simple model-based method to remove motion artifacts during short duration movements. ▶ The proposed algorithm can account for head movement and field deformations due to movement, within and outside of the scanner's field of view. ▶ It uses information from the experimental design and subject kinematics to focus the artifact attenuation in time and space and minimize the loss of uncorrupted data. ▶ Applications of the algorithm on arm reaching experimental data obtained with blocked and event-related designs demonstrate attenuation of motion artifacts with minimal effect on the brain activations.