Jiang Du¹, Eric Diaz¹, Won Bae¹, Sheronda Statum¹, Nikolaus Szeverenyi¹, Darryl DLima², Graeme Bydder¹, and Christine Chung^1
1Radiology, University of California, San Diego, San Diego, California, United States, ²Scripps Reseach Institution, San Diego, California, United States

The early stages of OA are associated with breakdown of collagen, decrease in proteoglycans (PG) and increase in water content (1). Recent research has focused on establishing correlations between quantitative MR parameters (T1, T2, T2*, T1ρ) and biochemical or biophysical properties of cartilage. Magic angle effects influence apparent T2 values of tissues with ordered collagen structure, such as articular cartilage, menisci, ligaments and tendons. The purpose of this study was to evaluate the effect of sample orientation on T2/T2* and T1ρ values of the deep radial layer of human patellar cartilage and ligaments.

Tao Jin¹, and Seong-Gi Kim^1
1Neuroimaging laboratory, Department of Radiology, University of Pittsburgh, Pittsburgh, PA, United States

The spin-lattice relaxation time in the rotating frame (T1rho) has been reported to be a sensitive indicator of cerebral ischemia and can provide complementary MR information of tissue status to water diffusion and perfusion. T1rho is most sensitive to molecular fluctuations with correlation time close to the inverse of the Rabi frequency of the applied spin-locking (SL) pulse. In biological tissue, previous studies have demonstrated that the chemical exchange between bulk water and labile protons of protein or metabolites is an important contributor for the T1rho relaxation in the frequency range of below several kHz. The chemical exchange contrast is related to the difference in the Larmor frequencies of the exchanging protons which increases with the magnetic field strength. Thus, to evaluate whether a large T1rho contrast can be detected at a high magnetic field of 9.4 T and to obtain some insight about the ischemia-induced changes in the tissue microenvironment, we studied the dynamic responses of T1rho during KCL-induced global ischemia for six different SL frequencies.

Anup Singh¹, Mohammad Haris¹, Kejia Cai¹, Walter RT Witschey², Hari Hariharan¹, and Ravinder Reddy^1
1CMROI, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States, ²Department of Radiology, University Hospital Freiburg, Freiburg, Germany

In the current study, MRI pulse sequence for fluid attenuated T1ρ mapping of human liver on clinical scanner was developed and implemented. Using this technique, we demonstrated the feasibility of single slice T1ρ mapping in a single breath hold on healthy volunteers. T1ρ mapping performed in single breath-hold minimized artifacts related to respiration motion as well as B₀ and B₁ in-homogeneities. In the normal liver, T1ρ relaxation times were found to be 45-60 ms and 42-48 ms on 1.5T and 3T scanners, respectively. Results from this study show the potential of quantifying biochemical changes in liver pathology using fluid suppressed T1ρ mapping.

Hye Young Heo¹, Brian J Dlouhy², Nader S Dahdaleh², Daniel R Thedens³, Bradley D Bolster⁴, John A Wemmie^2,5, and Vincent A Magnotta^1,3
1Biomedical Engineering, University of Iowa, Iowa City, IA, United States, ²Neurosurgery, University of Iowa, Iowa City, IA, United States, ³Radiology, University of Iowa, Iowa City, IA, United States, ⁴Siemens Healthcare, Rochester, MN, United States, ⁵Psychiatry, University of Iowa, Iowa City, IA, United States

The purpose of this study is to assess the ability of magnetic resonance (MR) imaging to measure relative pH changes in vivo. To evaluate pH changes using T₁ imaging, T₁ times were collected for a subject under three conditions: 1) breathing 5% CO₂, 2) breathing room air, and 3) hyperventilating. We found that widespread increases in T₁ times during the 5% CO₂ condition were consistent with acidosis, whereas reduced T₁ times during hyperventilation were consistent with alkalosis.

Catherine Anusha Mallik¹, David J Lythgoe¹, and Gareth J Barker^1
1Centre for Neuroimaging Sciences, Institute of Psychiatry, King's College London, London, United Kingdom

Increasing brain iron concentrations are part of the normal aging process, with differences in iron levels between younger and older adults in basal ganglia structures comparable to differences reported in pathological cases (compared to age-matched controls). Using the GESE (Gradient-Echo Spin-Echo) sequence R2, R2' and R2* were measured in two groups (n=6 in each group) of different ages. Significant differences (p <0.05) with age were detected in the substantia nigra for both R2' and R2*. No significant differences in any of the measures were found in white matter.

Antonella Meloni¹, Pasquale Pepe¹, Maria Chiara Dell'Amico¹, Gennaro Restaino², Gianluca Valeri³, Massimo Midiri⁴, Vincenzo Positano¹, Petra Keilberg¹, Antonio Cardinale⁵, Massimo Lombardi¹, and Alessia Pepe^1
1Fondazione G.Monasterio CNR-Regione Toscana and Institute of Clinical Physiology, Pisa, Italy, ²Università Cattolica del Sacro Cuore, Campobasso, Italy, ³Azienda Ospedaliero-Universitaria Ospedali Riuniti "Umberto I-Lancisi-Salesi", Ancona, Italy, ⁴Policlinico “Paolo Giaccone”, Palermo, Italy, ⁵Ospedale S Maria alla Gruccia, Montevarchi, Italy

The validated multislice multiecho T2* CMR technique has permitted to quantify segmental and global myocardial iron overload (MIO) detecting different patterns. Biventricular dysfunction is correlated with MIO distribution decreasing from the patients with homogeneous MIO (all segments with T2* values <20 ms) to the patients with no MIO (all segments with T2* values ≥20 ms). Homogeneous MIO and heterogeneous MIO (some segments with T2* values≥20 ms and other segments with T2* values <20 ms) with a global heart T2*<20 predicts a significantly higher risk to develop cardiac dysfunction evaluated by using CMR.

Antonella Meloni¹, John C Wood², Alessia Pepe¹, Leila J Noetzli², Maria Chiara Dell'Amico¹, Gianluca Valeri³, Claudio Ascioti⁴, Petra Keilberg¹, Massimo Lombardi¹, and Vincenzo Positano^1
1Fondazione G.Monasterio CNR-Regione Toscana and Institute of Clinical Physiology, Pisa, Italy, ²Children’s Hospital, Los Angeles, United States, ³Azienda Ospedaliero-Universitaria Ospedali Riuniti "Umberto I-Lancisi-Salesi", Ancona, Italy, ⁴P.O. “Giovanni Paolo II”, Lamezia Terme, Italy

The gradient echo multiecho T2* MRI technique is the most robust method for the sensitive, fast, and reproducible quantification of cardiac and hepatic iron overload in thalassemia patients. Longitudinal analysis of cardiac and hepatic iron overload by regular MRI screening allows the characterization of chelation therapies efficacy and mechanism in thalassemia patients.

Anna-Katinka Bracher¹, Axel Bornstedt¹, Erich Hell², Johannes Ulrici², and Volker Rasche^1
1Department of Internal Medicine II, University Hospital of Ulm, Ulm, Germany, ²Sirona Dental Systems, Bensheim, Germany

The T2* value of dentin and caries lesions was assessed by ultra-short echo time (UTE) MRI. A significant difference between the mean T2* values in caries lesions and healthy dentin was observed. Significant changes in T2* occur even before the lesions become visible in conventional X-ray imaging. This indicates the potential of MRI for very early identification of dental demineralizations and initial caries lesions.

Yongxian Qian¹, and Fernando E. Boada^1,2
1Radiology, University of Pittsburgh, Pittsburgh, PA, United States, ²Bioengineering, University of Pittsburgh, Pittsburgh, PA, United States

This work investigates potential sources responsible for MR signal delay we observed and reported before. Computer simulations and phantom studies showed that mobile ions (electrically-charged particles) may be the source for MR signal delay, rather than imperfect performance of the MRI system such as gradient dephasing, B0-field inhomogeneity or chemical shift.

Kejia Cai¹, Adam Shore¹, Anup Singh¹, Mohammad Haris¹, Damodar Reddy¹, Hari Hariharan¹, Mark Elliott¹, and Ravinder Reddy^1
1CMROI, Department of Radiology, University of Pennsylvania, Philadelphia, PA, United States

Tumor angiogenesis is the primary facilitator of rapid tumor growth and metastasis. Current MRI techniques using endogenous contrast cannot provide sufficient sensitivity for high-resolution imaging of tumor vasculature. In this study, we use hypoxic challenge to generate sufficient BOLD contrast (~40%), with which high-resolution tumor vasculature maps were able to be obtained. 3D high-resolution tumor vasculature maps can be acquired in minutes. This technique allows us to quantify tumor size, measure tumor vasculature density, provide early detection of tumor metastasis, monitor the effectiveness of cancer treatment drugs, and could potentially characterize tumor grade and aggressiveness.

Daniel Lee Shefchik¹, Andrew Scott Nencka¹, Andrzej Jesmanowicz¹, and James S Hyde^1
1Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States

The gradient-echo asymmetric spin-echo pulse sequence (GREASE) originally allowed for the production of T2 and/or T2* maps. We have modified GREASE to allow simultaneous mapping of T2, T2*, and T1. This was accomplished by accelerating image acquisition by adding partial k-space and generalized autocalibrating partially parallel acquisition (GRAPPA) methods, and utilizing additional excitation and refocusing pulses. These modifications to the original GREASE pulse sequence allow the acquisition of six images following an excitation pulse, enabling the generation of multi-parametric relaxivity maps.

Anna Izabella Blazejewska¹, Samuel Wharton¹, Penny A. Gowland¹, and Richard Bowtell^1
1Sir Peter Mansfield Magnetic Resonance Centre, University of Nottingham, Nottingham, United Kingdom

Magnetic susceptibility differences generate inhomogeneities in the local magnetic field which modulate the magnitude and phase of the signals acquired in MRI. With the advent of susceptibility mapping it has become particularly important to understand the relationship between microscopic structure in the distribution of susceptibility and the induced variation in the NMR frequency. We simulated the field perturbations generated in a uniform matrix containing oriented ellipsoidal inclusions of different susceptibility and used this information to explore the effect of the shape of the inclusions on the phase and magnitude of the MR signal in the "static dephasing" regime.

Christian Langkammer^1,2, Nikolaus Krebs², Walter Goessler³, Eva Scheurer², Michaela Soellinger¹, Kathrin Yen², Franz Fazekas¹, and Stefan Ropele^1
1Department of Neurology, Medical University of Graz, Graz, Austria, ²Ludwig Boltzmann Institute for Clinical-Forensic Imaging, Graz, Austria, ³Institute of Chemistry - Analytical Chemistry, University of Graz, Graz, Austria

This study investigated possible contributions of iron deposition, myelin density and fiber orientation on R2* relaxation rates in white matter by using quantitative MRI of postmortem brains in situ and by assessing regional iron concentrations with inductively coupled plasma mass spectrometry. The results suggest that the density of myelin is the dominant determinant of R2* in white matter while the presence of iron additionally has a strong effect. Our results do not support the hypothesis that fiber orientation contributes to the phenomenon of T2* relaxation.

Fumiyuki Mitsumori¹, Hidehiro Watanabe¹, and Nobuhiro Takaya^1
1Natl. Inst. Environmental Studies, Tsukuba, Ibaraki, Japan

Iron is an essential element in human body, but its overabundance is toxic through the production of reactive oxygen species. We previously reported that the apparent transverse relaxation rate (R₂^�õ = 1/T₂^�õ) in human brain obtained using a MASE sequence is well explained by a linear combination of the regional non-hemin iron concentration [Fe] and the macromolecular mass fraction f_M defined as 1 - water fraction. In the present study we attempted the quantitative iron mapping based on the above relationship between R₂^�õ, [Fe], and f_M. Iron distribution in normal brain and that with cavernous hemangioma was successfully visualized.

Catherine D. G. Hines¹, Calista Roen¹, Diego Hernando¹, and Scott B Reeder^1,2
1Radiology, University of Wisconsin-Madison, Madison, WI, United States, ²Biomedical Engineering, University of Wisconsin-Madison, Madison, WI, United States

As iron differentially affects the T2* of water and fat, the purpose of this work is to investigate the impact of fat particle size on the behavior of R2* of water and fat for T2* correction for fat quantification. A homogeneous fat-water-SPIO phantom of constant fat-fraction was constructed with varying particle sizes of fat. Phantom results show that in the presence of iron, fat and water have different R2* values, which are highly dependent on fat particle size. Thus, fat particle size impacts the ability of a quantitative MRI method that uses T2* correction to quantify fat, depending on how T2* correction is performed.

Sam Wharton¹, and Richard Bowtell^1
1Sir Peter Mansfield Magnetic Resonance Centre, University of Nottingham, Nottingham, United Kingdom

Recently, it was shown that the magnetic susceptibility of white matter has measurable anisotropy, but calculation of the full anisotropic susceptibility tensor requires a 6-parameter fit, using phase maps acquired at multiple orientations to B0. Here, we present a method for reducing the parameters in the fitting process to two, based on the assumption of cylindrical symmetry and a priori knowledge of the orientation of the principal axis of the susceptibility tensor. The 2-parameter method is demonstrated on simulated data and shown to require less sampling orientations and a smaller range of rotation angles compared to the 6-parameter method.

Elizabeth Mary Tunnicliffe^1,2, Martin John Graves^1,3, and Matthew D Robson^4
1Department of Medical Physics & Clinical Engineering, Addenbrooke's Hospital, Cambridge, United Kingdom, ²AVIC, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, United Kingdom, ³Department of Radiology, University of Cambridge, Cambridge, United Kingdom, ⁴OCMR, Department of Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom

The noise covariance matrix of a phased array coil provides information about both the noise level in individual channels and the correlations between channels. It can therefore be useful in quality assurance procedures as these characteristics are liable to change in the case of coil failure. This work presents examples of noise covariance matrices of functioning and failing arrays and suggests quantitative limits, beyond which remedial action may be required.

Ivan I. Maximov¹, Ezequiel A. Farrher¹, Farida Grinberg¹, and Nadim Jon Shah^1,2
1Institute of Neuroscience and Medicine 4, Forschungszentrum Juelich, Juelich, Germany, ²Department of Neurology, Faculty of Medicine, JARA, RWTH Aachen University, Aachen, Germany

We propose a new algorithm for noise correction in DTI experiments based on the hypothesis of spatially-variable noise fields. Application of the robust estimator followed by Rician correction of the initially assumed Gaussian standard deviation allows us to produce a more stable and precise scheme for the noise evaluation at arbitrary signal-to-noise ratio levels.

Santiago Aja-Fernandez¹, and Antonio Tristan-Vega^2
1Universidad de Valladolid, Valladolid, VA, Spain, ²Harvard Medical School, Boston, MA, United States

Noise in the composite magnitude signal (CMS) from multiple-coil systems is usually assumed to follow a noncentral chi distribution. However, this is true only if the variance of noise is the same for all coils, and no correlation exists between them. If the covariance matrix is not diagonal, the distribution of the CMS is not strictly a noncentral chi. It could be modeled as such only if the coefficient of correlation is small enough and effective values are considered. This will imply a reduced effective number of coils and an increased effective variance of noise.

Elizabeth Mary Tunnicliffe^1,2, Martin John Graves^1,3, and Matthew D Robson^4
1Department of Medical Physics & Clinical Engineering, Addenbrooke's Hospital, Cambridge, United Kingdom, ²AVIC, Nuffield Department of Clinical Medicine, University of Oxford, Oxford, United Kingdom, ³Department of Radiology, University of Cambridge, Cambridge, United Kingdom, ⁴OCMR, Department of Cardiovascular Medicine, University of Oxford, Oxford, United Kingdom

The simulated-multi-image SNR method allows SNR measurement when the usual multi-image method is impractical, for example in vivo. However, it is limited to linear reconstruction methods, which excludes the usual root-sum-of-squares reconstruction mechanism. Here, we show that by using simulated-multi-power images, this method can be extended to be used with root-sum-of-squares reconstructed images. The multi-power image method also has the advantage that it accounts for noise bias in the signal measurement at lower SNR, unlike the usual multi-image method.

Anna Andreychenko¹, Sjoerd Crijns¹, Ingmar Voogt¹, Wouter Koning¹, Peter Luijten¹, Jan J.W. Lagendijk¹, and Cornelis A.T. van den Berg^1
1University Medical Center Utrecht, Utrecht, Utrecht, Netherlands

The SNR differences between optimal SNR, magnitude SNR and Sum-of-Squares (SoS) signal combination methods for an array of multiple receiver coil are generally thought to be modest. However, at high fields wave propagation effects arise which can significantly degrade SNR in the images reconstructed with the SoS. Here we have demonstrated both, theoretically and experimentally that optimal signal combination at high field (7T) is beneficial as it can considerably increase image SNR whereas at low field (1.5 T) there is no relevant SNR gain.

Nicolas Robitaille¹, Abderazzak Mouiha¹, and Simon Duchesne^1,2
1Laboratoire MEDICS, Centre de Recherche Université Laval Robert-Giffard, Québec, Québec, Canada, ²Radiology Department, Université Laval, Québec, Québec, Canada

Intensity standardization aims at correcting scanner-dependent intensity variations. Most existing techniques aim at matching the input image histogram onto a standard, while we believe the goal should be to match spatially corresponding tissue intensities. In this study, we present STI, a novel technique implementing such approach. We compared STI to an existing histogram-matching method on the ADNI multi-centric dataset by evaluating the mean absolute intensity error (MAE) of the standardized images with respect to the standard, after non-linear registration onto the standard. STI gave better results by showing significantly lower mean MAE for white matter.

Muqing Lin¹, Siwa Chan², Jeon-Hor Chen^1,3, Daniel H-E. Chang¹, Ke Nie¹, Shih-Ting Chen⁴, Cheng-Ju Lin⁴, Tzu-Ching Shih⁴, Orhan Nalcioglu¹, and Min-Ying Lydia Su^1
1Tu & Yuen Center for Functional Onco-Imaging and Department of Radiological Sciences, University of California, Irvine, CA, United States, ²Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan, ³Department of Radiology, China Medical University, Taichung, Taiwan, ⁴Department of Biomedical Imaging and Radiological Science, China Medical University, Taichung, Taiwan

The purpose is to test a new bias-field correction method by combining N3 (nonparametric non-uniformity normalization) and Fuzzy-C-Means clustering based methods on breast MR images. The new algorithm is based on N3 for an initial correction; then FCM-based correction with smoothing using B-spline surface fitting was repeated iteratively for gradually refined improvements. The results indicated that the N3+FCM correction method performs significantly better than N3 or FCM, and comparable to CLIC. This new bias-field correction method can be implemented to improve the segmentation quality of breast density on inhomogeneous breast MRI, or other images acquired using a surface coil.

Stathis Hadjidemetriou¹, Michael Weiner², and Juergen Hennig^1
1Department of Radiology, Medical Physics, University Medical Center Freiburg, Freiburg, Germany, ²Department of Radiology, VA UCSF, San Francisco, CA 94121, United States

MRI assumes a uniform static radio-frequency field. However, this is not physically possible and results in an intensity non-uniformity artifact across an image that complicates its further quantitative data analysis. Typically, an acquisition protocol provides images of different contrasts that may suffer from different non-uniformities. This work presents an effective joint intensity uniformity restoration method for two such images. It is performed with Wiener restoration of the auto-co-occurrence and cross-co-occurrence statistics of the images together with several regularity constrains. The effectiveness of the method has been demonstrated with phantom and real brain data.

Torben Schneider¹, David L Thomas², Carolina Kachramanoglou², Olga Ciccarelli², Daniel C Alexander³, and Claudia AM Wheeler-Kingshott^1
1Department of Neuroinflammation, UCL Institute of Neurology, London, United Kingdom, ²Department of Brain Repair & Rehabilitation, UCL Institute of Neurology, London, United Kingdom, ³Centre for Medical Image Computing, Department of Computer Science, UCL, London, United Kingdom

We present a novel correction method for partial volume effects on the estimation of average parameters derived from diffusion tensor imaging in the cervical spinal cord. We demonstrate improved accuracy of our approach in healthy volunteers and demonstrate that our method significantly reduces bias from partial volume effects.

Stephan William Anderson¹, Jorge A Soto¹, Osamu Sakai¹, and Hernan Jara^1
1Radiology, Boston University Medical Center, Boston, MA, United States

Purpose: To develop a T2 qMRI algorithm such that the number of echoes used for semi-logarithmic regression is adaptively and iteratively determined on a pixel-by-pixel basis for maximum pixelwise Pearson correlation. Methods: The adaptive iterative T2 algorithm was programmed and used to process CPMG images of mouse liver specimen at 11.7T. Results: T2 maps were generated that provide high anatomical detail as well as high Pearson correlation coefficients across the field of view. Processing time for 16 slices was 90s. Conclusion: Adaptive iterative T2 mapping with short processing times is feasible. This algorithm could be useful for monitoring subtle T2 changes caused by disease in animal models and also for processing in vivo CPMG data.

Stephan William Anderson¹, Osamu Sakai¹, Jorge A Soto¹, and Hernan Jara^1
1Radiology, Boston University Medical Center, Boston, MA, United States

Purpose: To develop a method for correcting phase encoding profile order effects that can increase the accuracy of T2 qMRI with DE-FSE sequences. Methods: Algorithm incorporating such effects is developed and tested with data of the ACR MRI accreditation phantom and of the brain of a research subject. Results: The correction method leads to T2 values in excellent agreement to those reported with CPMG-32 echoes. Conclusion: Accuracy of T2 qMRI with the DE-FSE pulse sequence can be improved by correcting for profile order effects. This work could have implications for large scale studies using DE-FSE pulse sequences.

Pei-Hsin Wu¹, Nan-Kuei Chen², and Hsiao-Wen Chung^1
1Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, Taiwan, ²Brain Imaging and Analysis Center, Duke University Medical Center, Durham, NC, United States

The accuracy of T2* quantification plays an important role in many researches and applications. However, when there exists background susceptibility field gradients, T2* values may be distorted due to TE-shifting effect and echo-shifting effect. In this study, we proposed an improved fitting procedure to quantify the £GTE and £Gky for correcting TE-shifting effect and avoiding signal-loss artifact from echo-shifting effect. Experimental results in the phantom and human brain suggest that the correction scheme proposed in this study has potential to help achieving accurate T2* mapping, particularly at regions with prominent background field gradient.

Yo Taniguchi¹, Suguru Yokosawa¹, and Yoshitaka Bito^1
1Central Research Laboratory, Hitachi, Ltd., Kokubunji, Tokyo, Japan

In MR parameter mapping, parameters are estimated from images obtained with various acquisition parameters. For the estimation, the intensity function, which defines the relationship of image intensity to acquisition and MR parameters, needs to be formulated analytically in a simple form. A method to formulate the intensity function numerically by computer simulation based on Bloch equations is proposed. Intensity functions of arbitrary pulse sequences are formulated using this method so that rapid imaging is applied for the mapping. The intensity function for partially RF-spoiled gradient echo was formulated numerically, and we confirmed that T1, T2, and B1 maps were well estimated simultaneously from images obtained in a phantom experiment.

Joshua Trzasko¹, Petrice M. Mostardi¹, Stephen J Riederer², and Armando Manduca^1
1Mayo Clinic, Rochester, MN, United States, ²Mayo Clinic

Variable flip angle SPGR imaging, despite challenges due to transmit field inhomogeneities, remains a widely used strategy for T1 mapping. Nonetheless, there has been little work on numerical strategies for computing T1 values from such image series, and most estimations are performed using either a two-parameter nonlinear least squares fitting, which is accurate but computationally intensive, or a linear regression approximation, which is significantly faster but prone to error. In this work, we describe a simple numerical optimization strategy which can accelerate NLLS computation by over an order of magnitude without compromising its accuracy.

Qing Ji¹, Junyu Guo¹, John O. Glass¹, and Wilburn E. Reddick^1
1Radiological Science, St.Jude Children's Research Hospital, Memphis, TN, United States

Quantification of myelin water fraction (MWF) involves the calculation of the T2 distributions of relaxation spectra using a regularized non-negative least square (NNLS) algorithm. Simulated and experimental T2 decay data were used to investigate the influence of the NNLS regularization on the MWF calculation. The relationship between the MWF and regularization is a “V’ shape curve. The regularization which gave the most accurate and stable agreement with the simulated data was on the strong side of the curve in contrast to previous studies which suggested a weak regularization. Use of the stronger regularization was demonstrated for in vivo data.

Xiaolu Zhu¹, Boguslaw Tomanek¹, and Jonathan Sharp^1
1Institute for Biodiagnostics (West), National Research Council of Canada, Calgary, AB, Canada

Pixels visible as small square patches of uniform intensity do not represent object structures – and are therefore artifacts. To quantitatively evaluate this we designed a Fourier-based method to decompose a pixelated visual scene into two parts: one representing the information in the original MRI k-space data, and the other the visible structure related to pixelized patchwork display. Results show that pixelized patchwork displays: A) attenuate actual high spatial frequency content (because patchwork display enforces image uniformity within each patch); and B) introduce artifactual high spatial frequencies (the patch edges). The solution is to zero-fill until individual pixels become invisible.