A Motion Correction Strategy for Multi-Contrast based 3D parametric imaging: Application to Inhomogeneous Magnetization Transfer (ihMT)

Abstract

PurposeTo propose an efficient retrospective image-based method for motion correction of multi-contrast acquisitions with a low number of available images (MC-MoCo) and evaluate its use in 3D inhomogeneous Magnetization Transfer (ihMT) experiments in the human brain. MethodsA framework for motion correction, including image pre-processing enhancement and rigid registration to an iteratively improved target image, was developed. The proposed method was compared to Motion Correction with FMRIBs Linear Image Registration Tool (MCFLIRT) function in FSL over 13 subjects. Native (pre-correction) and residual (post-correction) motions were evaluated by means of markers positioned at well-defined anatomical regions over each image. ResultsBoth motion correction strategies significantly reduced inter-image misalignment, and the MC-MoCo method yielded significantly better results than MCFLIRT. ConclusionMC-MoCo is a high-performance method for motion correction of multi-contrast volumes as in 3D ihMT imaging.