Electronic poster

Prospective Motion Correction

Hall B Monday 14:00-16:00 Computer 121

14:00 5023. Combination of Prospective and Retrospective Motion Correction for Multi-Channel MRI

Chaiya Luengviriya^1,2, Jian Yun¹, Oliver Speck<sup>1

1</sup>Department of Biomedical Magnetic Resonance, Otto-von-Guericke University, Magdeburg, Germany; ²Department of Physics, Kasetsart University, Bangkok, Thailand

Real-time prospective correction is a very promising method to avoid image quality degradation caused by subject motion in MRI of human brain. The inaccuracy of measured motion data is limiting the efficiency. Simulations showed that residual motion artifacts after a prospective correction increased with the level of inaccuracy. Even for an ideal accurate case, artifacts can appear in multi-channel MRI because of coil sensitivity map errors. A retrospective correction was proposed and showed that the image quality can be further improved, especially for strong motion. In all cases, smooth motion resulted in fewer artifacts than abrupt motion.

14:30 5024. MR Compatible Sensor for Motion Artifact Corrected Reconstruction Method

Laure Rousselet^1,2, Maélène Lohezic, ²³, Damien Mandry, ^2,4, Cédric Pasquier⁵, Jacques Felblinger<sup>1,2

1</sup>IADI, Nancy-Université, Nancy, France; ²U947, Inserm, Nancy, France; ³Global Applied Science Lab., GE Healthcare, Nancy, France; ⁴CHU de Nancy, Nancy, France; ⁵CIC801, INSERM, Nancy, France

The goal of this work was to develop a MR compatible sensor which aims at measuring the acceleration of a localized region of the body. It has been integrated to a motion compensated reconstruction (GRICS) as a physiological input to reduce respiratory artifacts and improve image quality.

GRICS reconstructions, performed with respiratory belt and with accelerometer, were compared to breath-hold images and averaged free breathing acquisitions. Accelerometer gives similar results to respiratory belts and to breath-hold. Moreover, it is easier to install on patient.

15:00 5025. Self-Encoded Marker Design for Adaptive Optical Real-Time Motion Correction

Christoph Forman¹, Murat Aksoy², Matus Straka², Joachim Hornegger¹, Roland Bammer<sup>2

1</sup>Pattern Recognition Lab, Department of Computer Science, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany; ²Department of Radiology, Stanford University, Stanford, CA, United States

Patient motionis one of the unsolved problems in MRI. In this study, we propose a novel large field of view marker design to be used with a monovision based system to correct for rigid head motion. The proposed marker design has gives better accuracy and can correct for larger patient head motion.

15:30 5026. Fast Cross-Calibration Between MR Scanner and Optical System for Prospective Motion Correction

Murat Aksoy¹, Christoph Forman¹, Matus Straka¹, Samantha Jane Holdsworth¹, Stefan Tor Skare^1,2, Joachim Hornegger³, Roland Bammer<sup>1

1</sup>Department of Radiology, Stanford University, Stanford, CA, United States; ²Karolinska Institute, Stockholm, Sweden; ³Computer Science, Friedrich-Alexander-University Erlangen-Nuremberg, Erlangen, Germany

Optical systems have been proposed as a way to achieve rigid head motion correction with minimal changes to the pulse sequence. However, the usability of these systems in clinical practice have been limited mostly due to the tedious scanner-optical system cross-calibration routine that is required to link the frame of references of the optical system and the MR scanner. In this study, we propose a fast cross-calibration routine that is easy and takes only a couple of seconds. The variability of the scanner-camera cross calibration parameters with varying sequence parameters was investigated.

Tuesday 13:30-15:30 Computer 121

13:30 5027. Prospective Motion Correction for MRI with a Single Retro-Grate Reflector Target and a Single Camera

Maxim Zaitsev¹, Brian S. R. Armstrong², Brian Andrews-Shigaki³, Todd P. Kusik², Robert T. Barrows², Kazim Gumus³, Ilja Y. Kadashevich⁴, Thomas Prieto⁵, Oliver Speck⁴, Thomas M. Ernst<sup>3

1</sup>Dept. of Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany; ²Electrical Engineering and Computer Science, UW-Milwaukee, Milwaukee, WI, United States; ³John A. Burns School of Medicine, University of Hawaii, Honolulu, HI, United States; ⁴Biomedical Magnetic Resonance, Otto-von-Guericke University,, Magdeburg, Germany; ⁵Medical College of Wisconsin, Milwaukee, WI, United States

Even subtle motions degrade MR image quality. With optical stereoscopic motion tracking it is possible to correct for head motion in 6 degrees of freedom. However, it is extremely difficult to realise in the tight geometric constraints of the MR scanner, while keeping up with comfort and handling requirements of the clinical routine. Optical motion tracking with a single retro-grate reflector (RGR) target and a single camera has a great potential due to its versatility and accuracy. Reported here is the successful implementation of a prospective real time motion correction with RGR tracking, aiming at developing easy-to-handle motion correction strategies.

14:00 5028. Breathing Motion Artifact Reduction for MRI with Continuously Moving Table Using Motion Consistent Retrospective Data Selection

Matthias Honal¹, Jochen Leupold¹, Tobias Baumann<sup>2

1</sup>Dept. of Diagnostic Radiology, Medical Physics, University Hospital Freiburg, Freiburg, Germany; ²Dept. of Diagnostic Radiology, University Hospital Freiburg, Freiburg, Germany

A retrospective breathing motion compensation technique for axial MRI with continuously moving table is proposed. Redundant free breathing acquisitions are performed and motion consistent undersampled k-spaces are retrospectively extracted for parallel imaging reconstruction. Compared to a conventional reconstruction from free breathing data artifacts are significantly reduced. Except for increased noise, the achieved image quality is comparable to a reconstruction from a breath-hold acquisition.

14:30 5029. 3D PROMO MRI with Online Automatic Slice Positioning

Nathan Scott White¹, Josh Kuperman², Beth Ripley², Ajit Shankaranarayanan³, Eric Han³, Anders Dale<sup>2,4

1</sup>Dept. of Cognitive Science, University of California, San Diego, La Jolla, CA, United States; ²Dept. of Radiology, University of California, San Diego, La Jolla, CA, United States; ³Global Applied Science Lab, GE Healthcare, Menlo Park, CA, United States; ⁴Dept. of Neuroscience, University of California, San Diego, La Jolla, CA, United States

We present an extension to the real-time 3D "PROspective MOtion correction" (PROMO) framework for online correction of between-scan motion in 3D sequences through automatic slice plane positioning. Initial results demonstrate an intra-subject alignment precision of better than 1 mm/deg, despite initial position/landmark differences of over 13 mm. Given current scanner and computer hardware capabilities, the alignment procedure can be done in about a second or two, making it suitable to be integrated as part of a routine automatic pre-scan procedure.

15:00 5030. Motion-Induced Frequency and Shim Variations During Localized 1H MR Spectroscopy with Prospective Motion Correction

Brian Keating¹, Thomas Ernst<sup>1

1</sup>Department of Medicine, John. A. Burns School of Medicine, University of Hawaii, Honolulu, HI, United States

Motion during brain 1H MR spectroscopy may cause susceptibility-induced changes in B0. We used a PRESS sequence with prospective motion correction (PMC) to quantify the effects of motion on center frequency and shim quality. Subjects performed x- and z-head rotations while the MRS voxel tracked head motion. The center frequency displays a linear dependence on both θx (0.01ppm/deg) and θz (0.002ppm/deg). The FWHM is approximately a quadratic function of the θx, but is largely independent of θz. Our results indicate that PMC requires real-time frequency and shim updates in order to recover high-quality spectra in the presence of subject motion.

Wednesday 13:30-15:30 Computer 121

13:30 5031. Head Pose Prediction for Prospectively-Corrected EPI During Rapid Subject Motion

Julian Maclaren¹, Rainer Boegle¹, Michael Herbst¹, Jürgen Hennig¹, Maxim Zaitsev<sup>1

1</sup>Medical Physics, Dept. of Diagnostic Radiology, University Hospital Freiburg, Freiburg, Germany

The final goal of this project is fMRI of moving subjects using prospective motion correction. An optical tracking system provides head pose data in six degrees of freedom, which are used to prospectively update the imaging volume. However, latency delays in the tracking system, and the effective echo time of the sequence, result in a time lag between position measurement and the acquisition of the central k-space line. This causes in errors in slice registration. This work shows that motion prediction using a Kalman filter can solve this problem.

14:00 5032. Prospective Motion Correction for Single-Voxel ¹H MR Spectroscopy

Brian Keating¹, Weiran Deng¹, J Cooper Roddey², Nathan White³, Anders Dale², V Andrew Stenger¹, Thomas Ernst<sup>1

1</sup>Department of Medicine, University of Hawaii, Honolulu, HI, United States; ²Department of Neuroscience, University of California at San Diego, La Jolla, CA, United States; ³Department of Cognitive Science, University of California at San Diego, La Jolla, CA, United States

Motion during brain ¹H MR spectroscopy acquisitions can compromise spectral quality. We adapted an image-based adaptive motion correction module for use with a PRESS sequence. Sets of three orthogonal spiral navigator images are acquired in each TR period, to estimate head motion in real-time. By applying the appropriate rotations and translations, the voxel can be made to remain stationary with respect to the brain. Adaptive motion correction recovered original metabolite values (Cho/Cr ratio) to within a few percent even for extensive head movements (20-30°), whereas non-navigated spectra showed marked changes in metabolite levels as well as increased variability.

14:30 5033. PROspective MOtion Correction (PROMO) Results in Improved Image and Segmentation Quality of High-Resolution MRI Scans of Children

Joshua M. Kuperman^1,2, Timothy T. Brown, ²³, Matthew J. Erhart, ²³, J Cooper Roddey, ²³, Nathan Cooper White, ^2,4, Ajit Shankaranarayanan⁵, Eric T. Han⁵, Daniel Rettmann⁶, Anders M. Dale, <sup>23

1</sup>Radiology, UCSD, La Jolla, CA, United States; ²Multimodal Imaging Lab, UCSD, La Jolla, CA, United States; ³Neurosciences, UCSD, La Jolla, CA, United States; ⁴Cognitive Science, UCSD, La Jolla, CA, United States; ⁵Applied Science Lab, GE Healthcare, Menlo Park, CA, United States; ⁶Applied Science Lab, GE Healthcare, Rochester, MN, United States

In order to test the utility of PROspective MOtion correction (PROMO) for pediatric MRI research, nine children, ages 9-12, were scanned four times with a high-resolution T1-weighted sequence. For each subject, PROMO on and off scans were collected in a counterbalanced alternating pattern. Results show a qualitative enhancement in image clarity and reduction of apparent motion artifacts with the use of PROMO. Furthermore, automated segmentations of PROMO-enabled images show significant improvements in quality and reliability as compared to PROMO-off images. Volumetric segmentations of structures show consistently greater percent volume overlap when PROMO is enabled.

15:00 5034. Pulsed Continuous Arterial Spin Labeling (PCASL) with Prospective Motion Correction (PROMO)

Jian Zhang^1,2, Greg Zaharchuk², Michael Moseley², Eric Han³, Nate White⁴, Cooper Roddey⁴, Daniel Rettmann⁵, Anders Dale⁴, Joshua Kuperman⁴, Ajit Shankaranarayanan<sup>3

1</sup>Department of Electrical Engineering, Stanford University, Stanford, CA, United States; ²Department of Radiology, Stanford University, Stanford, CA, United States; ³Global Applied Science Lab, GE Healthcare, Menlo Park, CA, United States; ⁴Department of Neuroscience, University of California, San Diego, La Jolla, CA, United States; ⁵Global Applied Science Lab, GE Healthcare, Rochester, MN, United States

Pulsed Continuous Arterial Spin Labeling (PCASL) is a promising whole-brain perfusion imaging technique, with good properties such as high efficiency, 3D multi-slice capability, and low hardware demands. However, this sequence is vulnerable to patient motions due to its long scan time. We demonstrate an improved perfusion imaging strategy by integrating the original PCASL sequence with a PROspective MOtion (PROMO) correction module. The new sequence is much more robust against brain motion with little interference between the imaging volume and PROMO navigators.

Thursday 13:30-15:30 Computer 121

13:30 5035. Catadioptric RGR Motion Tracking for Prospective Motion Compensation in MR Acquisitions

Brian S. R. Armstrong¹, Todd P. Kusik¹, Robert T. Barrows¹, Brian Andrews-Shigaki², Julian Maclaren³, Maxim Zaitsev³, Oliver Speck⁴, Thomas Prieto⁵, Thomas Ernst<sup>2

1</sup>Electrical Engineering, Univ. Wisc.-Milwaukee, Milwaukee, WI, United States; ²Medicine, University of Hawaii, Honolulu, HI, United States; ³Dept. of Diagnostic Radiology, University Hospital Freiburg, Freiburg, Germany; ⁴Biomedical Magnetic Resonance, Otto-von-Guericke University; ⁵Neurology, Medical College of Wisconsin, Wauwatosa, WI, United States

A retro-grate reflector (RGR) optical system for tracking motion in an MR bore is presented, including an RGR motion tracking camera comprising a camera, lighting system and custom drive electronics in an RF enclosure, and a rib that has been engineered to grip the inside surface of the MR bore and support a mirror, which permits viewing through a head coil opening with the RGR camera positioned outside the head end of the MR bore. Evaluations of RF interference, mirror stability and tracking system noise are presented.

14:00 5036. Real-Time Motion Correction for High-Resolution Imaging of the Larynx: Implementation and Initial Results

Joëlle Karine Barral¹, Juan M. Santos¹, Dwight G. Nishimura<sup>1

1</sup>Electrical Engineering, Stanford University, Stanford, CA, United States

Respiratory motion is currently the main limitation to high-resolution MR imaging of the larynx. A novel algorithm integrating Compressed Sensing and the Diminishing Variance Algorithm is proposed and implemented within the framework of the real-time system RTHawk. The effectiveness of the approach is demonstrated on phantoms and in vivo.

14:30 5037. Real-Time Motion Detection for Structural Brain Imaging Using Multi-Coil FID Navigators

Tobias Kober^1,2, José P. Marques^1,3, Rolf Gruetter^1,4, Gunnar Krueger<sup>2

1</sup>Laboratory for functional and metabolic imaging, Ecole Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; ²Advanced Clinical Imaging Technology, Siemens Suisse SA - CIBM, Lausanne, Switzerland; ³Department of Radiology, University of Lausanne, Switzerland; ⁴Departments of Radiology, Universities of Lausanne and Geneva, Switzerland

Subject motion is often affecting the quality of MRI data. In this work, we investigate the feasibility to detect motion by monitoring repetitive FID navigator signals from arrays of receive coils. Object motion is shown to induce changes in the FID signal intensity. For proof of concept, the technique was applied to structural MP-RAGE scans of the brain. Subject motion was reliably detected. The technique has the potential to provide a scan quality measure and motion parameters for real-time or retrospective correction. It could be used in other MR sequences without time or signal penalty.

15:00 5038. Real-Time Intra-Volume Motion Correction in EPI Using Active Markers

Melvyn Boon King Ooi^1,2, Sascha Krueger³, William J. Thomas², Truman R. Brown<sup>1,2

1</sup>Biomedical Engineering, Columbia University, New York, NY, United States; ²Radiology, Columbia University, New York, NY, United States; ³Philips Research Europe, Hamburg, Germany

Head motion is a fundamental problem in fMRI. A prospective, slice-by-slice compensation strategy for rigid-body motion is presented for EPI sequences. Before the acquisition of each EPI-slice, a short tracking pulse-sequence is used to measure the positions of three micro RF-coil "active markers" integrated into a headband worn by the subject. During head motion, the rigid-body transformation that realigns these markers to their initial positions is fed back to update the image-plane – maintaining it at a fixed orientation relative to the head – before the next EPI-slice is acquired. EPI time-series are obtained that demonstrate real-time image-plane realignment during volunteer motion.

Image Correction - Non-Motion

Hall B Monday 14:00-16:00 Computer 122

14:00 5039. Application of K-Space Energy Spectrum Analysis for Inherent and Dynamic B₀ Mapping and Distortion Correction in DTI

Trong-Kha Truong¹, Nan-kuei Chen¹, Allen W. Song<sup>1

1</sup>Brain Imaging and Analysis Center, Duke University, Durham, NC, United States

Diffusion tensor imaging (DTI) is vulnerable to spatial and temporal B₀ variations due to susceptibility effects, eddy currents, subject motion, physiological noise, and system instabilities, resulting in geometric distortions and subsequent errors in the derivation of the diffusion tensor. Here, we propose a novel method based on k-space energy spectrum analysis, which can inherently and dynamically generate a B₀ map from the k-space data for each baseline (b = 0) and diffusion-weighted image, without requiring any additional data acquisition, to effectively and efficiently correct for such artifacts and achieve a high spatial fidelity and accuracy.

14:30 5040. A Blood Flow Navigator for Assessing Physiologic Noise in EPI Data

Andre van der Kouwe¹, Matthew Dylan Tisdall¹, Oliver Hinds¹, Aaron Hess², David Salat¹, Douglas Greve<sup>1

1</sup>Radiology, Massachusetts General Hospital, Charlestown, MA, United States; ²Biomedical Engineering, University of Cape Town, Cape Town, South Africa

A rapid single slice EPI acquisition (neck blood flow navigator) is interleaved between slices of a conventional multislice 2D BOLD EPI acquisition in a single sequence that provides information with high temporal resolution describing blood flow in the major arteries of the neck. This signal is tightly coupled to blood flow in the brain and may be used to assess and correct for physiologic noise in the BOLD signal. Navigator images are reconstructed in real-time during acquisition and it is shown that the timing of the cardiac signal derived from the navigators closely matches the timing of the photoplethysmograph.

15:00 5041. CINE Images of a Beating Rodent Cardiac Phantom

Steven Fortune¹, Ian Marshall¹, Maurits A. Jansen¹, Peter R. Hoskins¹, Tom Anderson<sup>1

1</sup>Medical Physics, University of Edinburgh, Edinburgh, United Kingdom

Small animal cardiac MRI is challenging due to small dimensions and fast heart rates. In order to assist in the development and testing of MRI sequences a rodent cardiac phantom has been designed and tested. It consists of a single chamber of PVAC housed in a water bath, expanded by an external pump. Initial CINE images of the phantom show this phantom functions with similar parameters to a rat left ventricle. Strong flow artefacts are present in the image. Thus this phantom could be used for the development of flow compensation methods as well as fast imaging methods.

15:30 5042. ECG-Triggered FASTERMAP Shimming Allows for Reproducible Shim Convergence in Spinal Cord Spectroscopy

Andreas Hock¹, Anke Henning², Michael Schär^3,4, Alexander Fuchs¹, Spyros Kollias⁵, Peter Boesiger<sup>1

1</sup>University and ETH Zurich, Institute for Biomedical Engineering, Zurich, Switzerland; ²University Hospital of Zurich, Institute for Biomedical Engineering, Zurich, Switzerland; ³The Johns Hopkins University School of Medicine, 1Russel H. Morgan Department of Radiology and Radiological Science, Baltimore, MD, United States; ⁴Philips Healthcare, Cleveland, OH, United States; ⁵University Hospital of Zurich, Institute of Neuroradiology, Zurich, Switzerland

ECG-triggered FASTERMAP shimming is introduced and compared to conventional FASTERMAP shimming and shimming based on ECG-triggered B0 field mapping for single voxel proton spectroscopy at 3T. This investigation shows that the use of ECG-triggered FASTERMAP seems to be a robust and applicable method for clinical spinal cord spectroscopy. It is significantly faster than ECG-triggered B0 mapping and in contrast to the conventional FASTERMAP shimming artifacts due to pulsatile CSF flow introduced by cardiac motion may have less influence.