Time Varying Filter Bank Approach to Parallel K-T MRI

Dynamic magnetic resonance imaging has become a widely accepted clinical diagnosis method, which is capable of capturing time variation or motion of a certain part in patients. One crucial characteristic is temporal resolution, which depends on the data acquisition rate of each time frame. In the last decade, it is proved that parallel MRI (pMRI) has the advantages of accelerating the data acquisition rate for each time frame. Therefore, a number of methods have been proposed to employ pMRI to improve the temporal resolution of dynamic imaging. However, the use of pMRI highly relies on the knowledge of spatial sensitivity map of each coil, which is not static over a whole imaging process. The general idea of this article is to devise a dynamic model of the time-variant sensitivity map based on certain pre-information and to apply this model to estimate the true sensitivity during imaging process. The experimental results show that the proposed method can effectively suppress the artifacts caused by the variation of sensitivity map over time.