Prospective Motion Correction for Functional MRI Using Sparsity and Kalman Filtering

摘要

We propose a novel algorithm to adaptively correct head motion during functional magnetic resonance imaging scans. Our method combines a Kalman-filter-like motion tracker and a registration cost function based on a sparse residual image model. Using simulated data, we compare a time series correlation analysis of our prospectively corrected reconstruction against the same analysis using post-scan motion correction provided by standard software. Our experiments demonstrate our prospective correction method is capable of mitigating motion effects and improving the sensitivity and specificity of the correlation analysis, without relying on costly external tracking hardware or separate navigational data that would take extra time to acquire during each time frame.