A Nonlinear GRAPPA Method for Improving SNR

INTRODUCTION GRAPPA [1] reconstructs the missing k-space data by a linear combination of the acquired data using a set of weights obtained through calibrations. Several methods have been proposed in recent years to improve GRAPPA using localized coil calibration and variable density sampling [2], multicolumn multiline interpolation [3], regularization [4,5], reweighted least square [6], high-pass filtering [7], cross validation [8], iterative optimization [9], multi-slice weighting [10], etc. In this abstract, a nonlinear GRAPPA method is proposed to address the poor SNR of GRAPPA at high reduction factors. The method is motivated by the fact that nonlinear filtering usually outperforms linear ones in denoising [11]. For example, TV regularization as a nonlinear method is superior to the linear Tikhonov regularization for SENSE reconstruction [12]. The proposed method uses a nonlinear combination of the acquired k-space data to estimate the missing data. The experimental results demonstrate that the proposed method is able to improve the SNR of GRAPPA at high reduction factors.