Combination of prospective and retrospective motion correction for multi-channel MRI

INTRODUCTION For high resolution MRI of human brain, subject motion is a major problem which degrades the image quality. Real-time prospective correction (1) is a very promising solution. Based on the motion data, this method updates the coordinates of the imaging volume such that the position and orientation of the subject with respect to the imaging volume remain unchanged for every sequence repetition. Due to noise and latency in an external tracking system, the measured motion data were proposed to be filtered and predicted (2) using a Kalman filter. The method improves the motion data. However, “residual motion” which is the difference of the predicted data and the true subject motion possibly presents and leads to an error of the prospective correction. In this study, motion artifacts after a prospective correction in multi-channel MRI (with spatially inhomogeneous sensitivity maps) were investigated by simulations for different levels of residual motion and different subject motion types. Using both subject motion and residual motion data, a retrospective correction to improve the image quality after the prospective correction is proposed and tested.