The motion artifact suppression technique (MAST) in magnetic resonance imaging: clinical results

K-space in the clinic

Magnetic resonance imaging (MRI) sequences are characterized by both radio frequency (RF) pulses and time-varying gradient magnetic fields. The RF pulses manipulate the alignment of the resonant nuclei and thereby generate a measurable signal. The gradient fields spatially encode the signals so that those arising from one location in an excited slice of tissue may be distinguished from those arising in another location. These signals are collected and mapped into an array called k-space that represents the spatial frequency content of the imaged object. Spatial frequencies indicate how rapidly an image feature changes over a given distance. It is the action of the gradient fields that determines where in the k-space array each data point is located, with the order in which k-space points are acquired being described by the k-space trajectory. How signals are mapped into k-space determines much of the spatial, temporal, and contrast resolution of the resulting images and scan duration. The objective of this article is to provide an understanding of k-space as is needed to better understand basic research in MRI and to make well-informed decisions about clinical protocols. Four major classes of trajectories-echo planar imaging (EPI), standard (non-EPI) rectilinear, radial, and spiral-are explained. Parallel imaging techniques SMASH (simultaneous acquisition of spatial harmonics) and SENSE (sensitivity encoding) are also described.

Distortions from curved flow in magnetic resonance imaging

We present an analysis of how vessel curvature can create distortions in magnetic resonance images of flowing blood. Steady flow in curved vessels produces distortions of the vessel shape and intensity variations in the image due to motion during the interval between phase encoding or slice selection and the echo center. Even with steady flow, vessel curvature produces motion moments higher than velocity (acceleration, etc.), but use of a first order oblique flow compensated phase encoding gradient waveform reduced the distortion in the image. Numerical calculations of image distortions based on simple flow models are in good agreement with experimental results in a phantom.

Significance of the point of expansion in interpretation of gradient moments and motion sensitivity

The relationship between magnetic field gradient waveform moments and the motion sensitivity of magnetic resonance imaging was explored analytically and by computer simulation. The analysis and simulations revealed several key points. In general, waveform time moments define sensitivity to the time derivatives of position of moving material only at a single time point: the time about which the moments are computed. A Taylor series description of instantaneous position is expanded about this same time point to compute the phase acquired due to specific derivatives of position. A moment is proportional to phase sensitivity to a particular derivative of position throughout the waveform only when sensitivity to all lower-order derivatives is zero. Under restricted conditions of waveform symmetry and motion characteristics, the phase due to motion may be expressed in terms of the average value of a derivative of position over the duration of the waveform. The choice of the moment center, or point of expansion, adds a degree of freedom that may be used advantageously in the design of motion-compensating and motion phase-encoding gradient waveforms. These results facilitate a more complete understanding of the effects of motion through a magnetic field gradient.