Joint correction of Nyquist artifact and minuscule motion-induced aliasing artifact in interleaved diffusion weighted EPI data using a composite two-dimensional phase correction procedure

Abstract

Diffusion-weighted imaging (DWI) obtained with interleaved echo-planar imaging (EPI) pulse sequence has great potential of characterizing brain tissue properties at high spatial-resolution. However, interleaved EPI based DWI data may be corrupted by various types of aliasing artifacts. First, inconsistencies in k-space data obtained with opposite readout gradient polarities result in Nyquist artifact, which is usually reduced with 1D phase correction in post-processing. When there exist eddy current cross terms (e.g., in oblique-plane EPI), 2D phase correction is needed to effectively reduce Nyquist artifact. Second, minuscule motion induced phase inconsistencies in interleaved DWI scans result in image-domain aliasing artifact, which can be removed with reconstruction procedures that take shot-to-shot phase variations into consideration. In existing interleaved DWI reconstruction procedures, Nyquist artifact and minuscule motion-induced aliasing artifact are typically removed subsequently in two stages. Although the two-stage phase correction generally performs well for non-oblique plane EPI data obtained from well-calibrated system, the residual artifacts may still be pronounced in oblique-plane EPI data or when there exist eddy current cross terms. To address this challenge, here we report a new composite 2D phase correction procedure, which effective removes Nyquist artifact and minuscule motion induced aliasing artifact jointly in a single step. Our experimental results demonstrate that the new 2D phase correction method can much more effectively reduce artifacts in interleaved EPI based DWI data as compared with the existing two-stage artifact correction procedures. The new method robustly enables high-resolution DWI, and should prove highly valuable for clinical uses and research studies of DWI.

Navigator-Free EPI Ghost Correction With Structured Low-Rank Matrix Models: New Theory and Methods

“…(8)) variations, and LORAKS with AC-LORAKS constraints as in Eq. (14). For SENSE-based reconstruction, sensitivity profiles were estimated using ESPIRiT [41].…”

“…Similarly, the third terms in Eqs. (13) and (14) are the only terms that introduce coupling between k + tot and k − tot , and the only terms that use the LORAKS S-matrix to introduce constraints on the image phase. The use of phase constraints is useful both for partial Fourier EPI acquisition and for stabilizing the reconstruction of symmetrically-acquired EPI data [24].…”

“…(13), there are also many ways to solve the optimization problem in Eq. (14). These two optimization problems have very similar structure (i.e., the first two terms are least-squares penalties, while the third term encourages low-rank matrix structure), and as a consequence, we have used a minor modification of the algorithm we used for solving Eq.…”

Magnetic resonance fingerprinting for simultaneous renal T1 and T2* mapping in a single breath‐hold

“…Therefore, improving the EPI baseline image quality was shown to improve the MRF map quality. Reduction of ghosting artifacts, Nyquist artifact, and motion correction were recently published, which all have the potential to improve the image quality of the proposed MRF method. Despite advanced shimming, field inhomogeneities disturb the k‐space echo train and therefore lead to geometric distortions .…”

Robust autocalibrated loraks for EPI ghost correction

“…Nyquist ghosts are one of the most common EPI artifacts, and occur because of systematic differences between the interleaved lines of k-space that are acquired with different readout gradient polarities, and/or because of systematic differences between interleaved lines of k-space data that are acquired with different shots in a multi-shot acquisition. Despite substantial efforts over several decades to solve this problem (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20), the widely-deployed modern ghost correction schemes are still prone to incomplete ghost suppression, as illustrated in Supporting Information Fig. S1.…”