Improved parallel imaging with GRAPPA with large virtual coils arrays for time-resolved applications

Introduction: Parallel imaging [1,2] is based on the simultaneous measurement of the MR signal with multiple receiver coils. The different sensitivity profiles of the coils modulate the signal of the object in different ways. It is thus possible to accelerate the measurement by compensating a reduction of the number of phase encodings by using spatial information derived from the coils sensitivity profiles without loss of information. The motion of an object can also be seen as a modulation of the object and of its MRsignal. Therefore, a time resolved measurement of a moving object can be considered as an acquisition with multiple (virtual) receiver coils along the temporal dimension. Assuming a N-element receive array and M time frames, parallel imaging reconstruction can then be performed by treating each element and time frame as an independent virtual coil resulting in a large virtual N x M coil array. In this work we investigated this hypothesis for time-resolved (CINE) phantom and an in-vivo cardiac measurement and compared the results to standard GRAPPA [2] reconstructions.