Effects of motion in sparsely sampled acquisitions

This chapter discusses the effects of motion in sparsely sampled acquisitions and constrained reconstruction. The chapter begins by discussing motion artifacts due to the choice of k-space versus time sampling when coupled with linear sliding window reconstruction. The benefits of non-Cartesian sampling for superior motion encoding are highlighted. Practical challenges associated with non-Cartesian sampling such as the need for trajectory error correction and off-resonance correction are detailed. The chapter then focuses on various constrained reconstruction models, with focus on the assumptions being made to model the dynamic object. These models include (a) parallel MRI, (b) generalized spatiotemporal models, (c) explicit motion estimation and compensated reconstruction models, and (d) implicit motion-resolved models such as manifold regularization, and extra-dimensional regularization models. Detailed examples and discussions are provided for a range of dynamic MRI applications.