Motion correction for fetal functional magnetic resonance imaging

Abstract

We present a novel motion correction algorithm designed specifically for fetal functional magnetic imaging resonance (fMRI). Fetal motion is a main limiting factor of fetal fMRI and standard algorithm cannot correct for fetal motion. The goals in designing the algorithm were: (i) the ability to correct for both large and small motion, (ii) the preferential weighting of fetal tissue, (iii) the development of a framework robust to artifacts, and (iv) the automatic censoring of low quality frames. The key feature of the algorithm is the use of 2nd order edge features instead of raw intensity or 1st order edge features. We demonstrate that our algorithm significantly out performs competing approaches.