Abstract: Many magnetic resonance imaging (MRI) applications requires the acquisition of a time series of images with both spatial and temporal high-resolution as well as with an orthogonal dualchannel. In conventional Fourier transform (FT) based imaging methods, on the one hand, each of these images is independently reconstructed from a frame of spatial encodings, so that temporal resolution is limited by the number of the spatial encodings of each frame collected in the case of given spatial resolution, or one has to sacrifice spatial resolution to obtain temporal resolution. On the other hand, the dual-channel data in Fourier space are independently acquired and reconstructed, then the MR image is obtained by the square modulus, so that the simplicity and efficientibility of computation are affected. In this paper, Hartley transform (HT) based MR imaging technique are proposed to address this problem. This technique makes use of the fact that MR images are real, and the support extent of high-resolution image (morphology) does not change from one image to another in most time-sequential imaging problem, and which helps to improve imaging efficiency and resolutions, and to reduce the complexity of MRI system over the conventional Fourier imaging method by eliminating the repeated encodings of this stationary information. This method should prove useful for a variety of dynamic imaging applications such as dynamic studies of functional brain imaging.

Key words: Fourier transform, Hartley transform, Signal extrapolation, MR imaging