Detection of motion during projection reconstruction magnetic resonance imaging

We present two new quantitative methods to determine motion intervals in projection reconstruction (PR) magnetic resonance (MR) imaging. Both methods do not use any a priori information concerning the motion and can be implemented easily. In the first section of this paper, the theoretical concept that formed the basis for the development of these methods, is described. It is shown that the phase difference between the complex means of two consecutive 1D PR signals equals zero when no motion occurs. When motion does occur, it becomes large. Similar conclusions can be drawn when the root mean square deviation between two consecutive projections is considered: it remains very small as long as no motion occurs, while it has a large value when rotational motion or translational motion with high-frequency components occur. These conclusions are used to determine the motion intervals in PR data. The motion-detection methods prove to be able to detect motion intervals with an accuracy of one repetition time.