Motion correction using the k-space phase difference of orthogonal acquisitions

OPERA: a novel method to reduce ghost and aliasing artifacts

OBJECTIVE

A method for Orthogonal Phase Encoding Reduction of Artifact (OPERA) was developed and tested.

MATERIALS AND METHODS

Because the position of ghosts and aliasing artifacts is predictable along columns or rows, OPERA combines the intensity values of two images acquired using the same parameters, but with swapped phase-encoding directions, to correct the artifacts. Simulations and phantom experiments were conducted to define the efficacy, robustness, and reproducibility. Clinical validation was performed on a total of 1003 images by comparing the OPERA-corrected images and the corresponding image standard in terms of Signal-to-Noise Ratio (SNR) and Contrast-to-Noise Ratio (CNR). The method efficacy was also rated using a Likert-type scale response by two experienced independent radiologists using a single-blinded procedure.

RESULTS

Simulations and phantom experiments demonstrated the robustness and effectiveness of OPERA in reducing artifacts strength. OPERA application did not significantly change the SNR [+ 4.16%; inter-quartile range (IQR): 2.72-5.01%] and CNR (+ 4.30%; IQR: 2.86-6.04%) values. The two radiologists observed a total of 893 original images with artifacts (89.03% of the total images), a reduction in the perceived artifacts of 82.0% and 83.9% (p < 0.0001), and an improvement in the perceived SNR (82.8% and 88.5%; K = 0.714) and perceived CNR (86.9-88.9%; K = 0.722).

DISCUSSION

The study demonstrated that OPERA reduces MR artifacts and improves the perceived image quality.

MRI restoration using edge-guided adversarial learning

Magnetic resonance imaging (MRI) images acquired as multislice two-dimensional (2D) images present challenges when reformatted in orthogonal planes due to sparser sampling in the through-plane direction. Restoring the "missing" through-plane slices, or regions of an MRI image damaged by acquisition artifacts can be modeled as an image imputation task. In this work, we consider the damaged image data or missing through-plane slices as image masks and proposed an edge-guided generative adversarial network to restore brain MRI images. Inspired by the procedure of image inpainting, our proposed method decouples image repair into two stages: edge connection and contrast completion, both of which used general adversarial networks (GAN). We trained and tested on a dataset from the Human Connectome Project to test the application of our method for thick slice imputation, while we tested the artifact correction on clinical data and simulated datasets. Our Edge-Guided GAN had superior PSNR, SSIM, conspicuity and signal texture compared to traditional imputation tools, the Context Encoder and the Densely Connected Super Resolution Network with GAN (DCSRN-GAN). The proposed network may improve utilization of clinical 2D scans for 3D atlas generation and big-data comparative studies of brain morphometry.

Automated Detection of Motion Artefacts in MR Imaging Using Decision Forests

The acquisition of a Magnetic Resonance (MR) scan usually takes longer than subjects can remain still. Movement of the subject such as bulk patient motion or respiratory motion degrades the image quality and its diagnostic value by producing image artefacts like ghosting, blurring, and smearing. This work focuses on the effect of motion on the reconstructed slices and the detection of motion artefacts in the reconstruction by using a supervised learning approach based on random decision forests. Both the effects of bulk patient motion occurring at various time points in the acquisition on head scans and the effects of respiratory motion on cardiac scans are studied. Evaluation is performed on synthetic images where motion artefacts have been introduced by altering the k-space data according to a motion trajectory, using the three common k-space sampling patterns: Cartesian, radial, and spiral. The results suggest that a machine learning approach is well capable of learning the characteristics of motion artefacts and subsequently detecting motion artefacts with a confidence that depends on the sampling pattern.

Motion correction of magnetic resonance imaging data by using adaptive moving least squares method

Image artifacts caused by subject motion during the imaging sequence are one of the most common problems in magnetic resonance imaging (MRI) and often degrade the image quality. In this study, we develop a motion correction algorithm for the interleaved-MR acquisition. An advantage of the proposed method is that it does not require either additional equipment or redundant over-sampling. The general framework of this study is similar to that of Rohlfing et al. [1], except for the introduction of the following fundamental modification. The three-dimensional (3-D) scattered data approximation method is used to correct the artifacted data as a post-processing step. In order to obtain a better match to the local structures of the given image, we use the data-adapted moving least squares (MLS) method that can improve the performance of the classical method. Numerical results are provided to demonstrate the advantages of the proposed algorithm.

White matter and deep gray matter hemodynamic changes in multiple sclerosis patients with clinically isolated syndrome

The dynamic susceptibility contrast magnetic resonance imaging perfusion technique was used to investigate possible hemodynamic changes in normal appearing white matter and deep gray matter (DGM) of 30 patients with clinically isolated syndrome (CIS) and 30 patients with relapsing-remitting multiple sclerosis. Thirty normal volunteers were studied as controls. Cerebral blood volume, cerebral blood flow (CBF), and mean transit time values were estimated. Normalization was achieved for each subject with respect to average values of CBF and mean transit time of the hippocampi's dentate gyrus. Measurements concerned three regions of normal white matter of normal volunteers, normal appearing white matter of CIS and patients with relapsing-remitting multiple sclerosis, and DGM regions, bilaterally. All measured normal appearing white matter and DGM regions of the patients with CIS had significantly higher cerebral blood volume and mean transit time values, while averaged DGM regions had significantly lower CBF values, compared to those of normal volunteers (P < 0.001). Regarding patients with relapsing-remitting multiple sclerosis, all measured normal appearing white matter and DGM regions showed lower CBF values than those of normal volunteers and lower cerebral blood volume and CBF values compared to patients with CIS (P < 0.001). These data provide strong evidence that hemodynamic changes--affecting both white and DGM--may occur even at the earliest stage of multiple sclerosis, with CIS patients being significantly different than relapsing-remitting multiple sclerosis patients.

Distortion correction of high b-valued and high angular resolution diffusion images using iterative simulated images

High b-valued diffusion-weighted images (DWI), which were designed to solve fiber-crossing problems, are susceptible to many artifacts and distortions. Since DWIs with different diffusion gradients produce dissimilar intensity contrasts, and since the distortion is nonlinear when multiple artifactual sources are intermixed, the mutual information-based affine registration may not be adequate for precise correction of distortions in DWIs, especially for images acquired with high b-values. To overcome these problems, we proposed an iterative image registration technique through which simulated DWIs are generated, driven from a diffusion tensor estimate, as targets for measured DWIs in the registration. Since simulated DWIs have similar intensity profiles to those of measured DWIs and the same geometric profiles as b(0)-images, an iterative procedure enables intensity-based nonlinear registration. As a pre-processing step, we also proposed a motion detection and sub-volume utilization for interleaved volumes. Performance evaluation with high b-valued DWIs for high angular resolution diffusion imaging and diffusion kurtosis imaging showed that the proposed method had a superior advantage over the conventional registration technique.

Cross-sampled GRAPPA for parallel MRI

As one widely-used parallel-imaging method, Generalized Auto-calibrating Partially Parallel Acquisitions (GRAPPA) technique reconstructs the missing k-space data by a linear combination of the acquired data using a set of weights. These weights are usually derived from auto-calibration signal (ACS) lines that are acquired in parallel to the reduced lines. In this paper, a cross sampling method is proposed to acquire the ACS lines orthogonal to the reduced lines. This cross sampling method increases the amount of calibration data along the direction that the k-space is undersampled and thus improves the calibration accuracy, especially when a small number of ACS lines are acquired. Both phantom and in vivo experiments demonstrate that the proposed method, named cross-sampled GRAPPA (CS-GRAPPA), can effectively reduce the aliasing artifacts of GRAPPA when high acceleration is desired.

Intrinsic detection of motion in segmented sequences

While many motion correction techniques for MRI have been proposed, their use is often limited by increased patient preparation, decreased patient comfort, additional scan time, or the use of specialized sequences not available on many commercial scanners. For this reason, we propose a simple self-navigating technique designed to detect motion in segmented sequences. We demonstrate that comparing two segments containing adjacent sets of k-space lines results in an aliased error function. A global shift of the aliased error function indicates the presence of in-plane rigid-body translation, while other types of motion are evident in the dispersion or breadth of the error function. Since segmented sequences commonly acquire data in sets of adjacent k-space lines, this method provides these sequences with an inherent method of detecting object motion. Motion corrupted data can then be reacquired proactively or in some cases corrected or removed retrospectively.

Motion correction properties of the shells k-space trajectory

The feasibility of a k-space trajectory that samples data on a set of 3D shells is demonstrated with phantom and volunteer experiments. Details of an interleaved multi-shot, helical spiral pulse sequence and a gridding reconstruction algorithm that uses Voronoi diagrams are provided. The motion-correction properties of the shells k-space trajectory are described. It is shown that when used in conjunction with three point markers, k-space data acquired with the shells trajectory provide a generalization of the RINGLET method, allowing for correction of arbitrary rigid-body motion with six degrees of freedom. Use of dedicated navigator echoes or redundant acquisitions of k-space data are not required. Retrospective motion correction is demonstrated with controlled phantom experiments and with seven healthy human volunteers. The motion correction is shown to improve the images, both qualitatively and quantitatively with a metric calculated from image entropy. Advantages and challenges of the shells trajectory are discussed, with particular attention to acquisition efficiency.

Improvement of the image quality of T1-weighted anatomical brain scans

T1-weighted anatomical brain scans are routinely used in neuroimaging studies, for example, as anatomical reference for functional data and in brain morphometry studies. Subject motion can degrade the quality of these images. An additional problem is the occurrence of signal dropouts in the case of long echo times and low receiver bandwidths. These problems are addressed in two different studies. In the first study, it is shown that the high scalp signal, which results from the low T1 value of fat, may cause a typical ringing artefact in the presence of head motion. This problem may be enhanced if phased array coils are used for signal reception due to their increased sensitivity in the peripheral head regions. It is shown that this artefact can be avoided by combining certain fat suppression techniques that reduce the scalp signal. In the second study, it is shown that signal dropout affects mainly the orbitofrontal cortex and the temporal lobes, and that a bandwidth of 100 Hz/pixel should be chosen for the investigation of these areas to avoid signal losses while maintaining an acceptable signal-to-noise ratio. Experimental results are based on the MDEFT sequence but can be applied to other T1-weighted sequences like FLASH and MP-RAGE. Furthermore, the presented methods for improving the image quality can be combined with other artefact reduction techniques.

RINGLET motion correction for 3D MRI acquired with the elliptical centric view order

A new rigid-body motion correction algorithm is described that is compatible with 3D image sets acquired with the elliptical centric (EC) view order. With this view order, an annular ring of k-space data is acquired in the ky-kz plane during any short time interval. Images for tracking motion can be reconstructed in the yz-plane from any ring of the acquisition data. In these tracking images, a point source (such as an external marker) shows a characteristic bull's-eye pattern that permits motion monitoring and correction. The true position of the point object is located at the center of the bull's-eye pattern. Cross correlation can be performed to automatically track the positions of markers reconstructed from adjacent rings of k-space. To increase the marker signal, the markers are encased in inductively coupled RF coils. Rigid-body motion in the yz-plane is calculated directly with the Euclidean group for rotation and translation, and corrected by rotating and applying phase shifts to any corrupted rings of data. In the current work we present a theoretical analysis of this method, as well as results of volunteer and controlled phantom experiments that demonstrate its initial feasibility. Although the EC view order has mainly been used for MR angiography (MRA), it can also be used for most 3D acquisitions.

Band artifacts due to bulk motion

Band artifacts due to bulk motion were investigated in images acquired with fast gradient echo sequences. A simple analytical calculation shows that the width of the artifacts has a square-root dependence on the velocity of the imaged object, the time taken to acquire each line of k-space and the field of view in the phase-encoding direction. The theory furthermore predicts that the artifact width can be reduced using parallel imaging by a factor equal to the square root of the acceleration parameter. The analysis and results are presented for motion in the phase- and frequency-encoding directions and comparisons are made between sequential and centric ordering. The theory is validated in phantom experiments, in which bulk motion is simulated in a controlled and reproducible manner by rocking the scan table back and forth along the bore axis. Preliminary cardiac studies in healthy human volunteers show that dark bands may be observed in the endocardium in images acquired with nonsegmented fast gradient echo sequences. The fact that the position of the bands changes with the phase-encoding direction suggests that they may be artifacts due to motion of the heart walls during the image acquisition period.