RF excitation encoding: a fast imaging technique for dynamic studies

Introduction: For dynamic studies such as functional MRI (fMRI), to achieve reasonable temporal resolution, spatial resolution and spatial coverage, it is necessary to use an under-sampling technique to speed up data acquisition. There are two effective ways to remove aliasing artifacts associated with under-sampling. One is parallel imaging (1, 2) and the other is the UNFOLD technique (3). Parallel imaging utilizes non-redundant sensitivity information provided by multi-channel coils to obtain alias-free images, whereas the UNFOLD technique utilizes temporal frequency information to filter out aliasing artifacts. Both approaches attack the problem from the acquisition side. They can be combined to enhance the effectiveness in removing aliasing artifacts (4). However, if we try to combine them to achieve a higher reduction factor, significant complication can be observed. In this work, a new technique called RF excitation encoding (REE) will be introduced. It approaches the problem from the excitation side. Slices will first be grouped into clusters with each cluster containing the same number of slices. For each time frame, clusters are excited with unique combinations of all slices. The excitation pulse is selected from a pool of RF pulses and looped through the pool repeatedly. The pool of RF pulses is designed so that each slice in the cluster corresponds to a specific frequency in the spectrum of temporal signal at any give k-space location. Then k-space signal for each slice can be easily separated by using an appropriate filter. Since slices in each cluster are excited together for each time frame, the volume TR can be reduced by a factor which equals the number of slices in each cluster.