Multidimensional k-space model for analysis of flow-related phenomena in MR imaging

The multidimensional partition model, which combines the k-space descriptio n of magnetic resonance imaging pulse sequences with a partition concept, h as been developed to model flow-related effects. A partition represents the magnetization created due to excitation by a given radiofrequency pulse. I n the present model, it is visualized as a set of parameters rather than a vector sum taken over a collection of spins. The phase dispersion due to ve locity and higher orders of motion is described by adding new dimensions to the three spatial k-space dimensions. The computer implementation of the m odel is capable of simulating the outcome from pulse sequences used in clin ical routines and it does not need assumptions regarding pulse sequence par ameters. Furthermore, flow-induced effects due to phase dispersion, caused by movements in a magnetic field gradient, as well as the inflow of unsatur ated spins, are handled by the model. Constant or varying (pulsatile) flow velocity may be assigned together with different velocity profiles (e.g., p lug flow, laminar flow) over a simulated vessel. The model makes it possibl e to simulate images of an object containing vessels with pulsatile laminar flow pattern, scanned using a three-dimensional angiography pulse sequence , in <10 min on a Pentium-based PC system. (C) 1999 John Wiley & Sons, Inc.