POCS-enhanced correction of motion artifacts in parallel MRI

Abstract

A new method for correction of MRI motion artifacts induced by corrupted k-space data, acquired by multiple receiver coils such as phased arrays, is presented. In our approach, a projections onto convex sets (POCS)-based method for reconstruction of sensitivity encoded MRI data (POCSENSE) is employed to identify corrupted k-space samples. After the erroneous data are discarded from the dataset, the artifact-free images are restored from the remaining data using coil sensitivity profiles. The error detection and data restoration are based on informational redundancy of phased-array data and may be applied to full and reduced datasets. An important advantage of the new POCS-based method is that, in addition to multicoil data redundancy, it can use a priori known properties about the imaged object for improved MR image artifact correction. The use of such information was shown to improve significantly k-space error detection and image artifact correction. The method was validated on data corrupted by simulated and real motion such as head motion and pulsatile flow.

Figure 1

Block-diagram of the proposed motion correction algorithm.

Figure 2

Magnitude (a) and phase (b) of an MR image corrupted by motion.

Figure 3

Comparison of detection plots for different motion correction approaches (x-axis represents phase encode direction). Detection plots are formed using method of multiple k-space copies regeneration (13) and the proposed method without constraining, with object support, with phase constraint, and both constraints. Arrows point to unresolved k-space errors. Inclusion of more information gradually improves identification of corrupted lines. Note excellent resolution of continuous k-space errors when detection is attempted with phase constraint.

Figure 4

Correction of brain data corrupted by bulk motion. a: Corrupted image, b: Corrected image. c: Absolute value of difference between (a) and (b) (x4).

Figure 5

Correction of neck images corrupted by pulsatile motion of cerebrospinal fluid (CSF). Corrupted image (a) was corrected by the proposed method both without (b) and with support and phase constraints (c). The corresponding magnified ROIs are given in d, e, f, respectively. Note the increase in CSF intensity and its homogeneity around spinal cord from d to e to f. The mean values within the CSF ring are 0.901, 0.986, and 1.019 for uncorrected (d), POCSENSE-corrected (e), POCSENSE+support+phase corrected (f) cases, respectively. The differences between corrupted and corrected images (b, c) (x4) illustrating the removed artifacts are given in (g) and (h), respectively.