A predictor-corrector phase unwrapping algorithm for temporally undersampled gradient-echo MRI.

PURPOSE: To develop a method for unwrapping temporally undersampled and nonlinear gradient recalled echo (GRE) phase. THEORY AND METHODS: Temporal unwrapping is performed as a sequential one step prediction of the echo phase, followed by a correction to the nearest integer wrap-count. A spatio-temporal extension of the 1D predictor corrector unwrapping (PCU) algorithm improves the prediction accuracy, and thereby maintains spatial continuity. The proposed method is evaluated using numerical phantom, physical phantom, and in vivo brain data at both 3 T and 9.4 T. The unwrapping performance is compared with the state-of-the-art temporal and spatial unwrapping algorithms, and the spatio-temporal iterative virtual-echo based Nyquist sampled (iVENyS) algorithm. RESULTS: Simulation results showed significant reduction in unwrapping errors at higher echoes compared with the state-of-the-art algorithms. Similar to the iVENyS algorithm, the PCU algorithm was able to generate spatially smooth phase images for in vivo data acquired at 3 T and 9.4 T, bypassing the use of additional spatial unwrapping step. A key advantage over iVENyS algorithm is the superior performance of PCU algorithm at higher echoes. CONCLUSION: PCU algorithm serves as a robust phase unwrapping method for temporally undersampled and nonlinear GRE phase, particularly in the presence of high field gradients.

Prediction of motion induced magnetic fields for human brain MRI at 3 T.

OBJECTIVE: Maps of B0 field inhomogeneities are often used to improve MRI image quality, even in a retrospective fashion. These field inhomogeneities depend on the exact head position within the static field but acquiring field maps (FM) at every position is time consuming. Here we propose a forward simulation strategy to obtain B0 predictions at different head-positions. METHODS: FM were predicted by combining (1) a multi-class tissue model for estimation of tissue-induced fields, (2) a linear k-space model for capturing gradient imperfections, (3) a dipole estimation for quantifying lower-body perturbing fields (4) and a position-dependent tissue mask to model FM alterations caused by large motion effects. The performance of the combined simulation strategy was compared with an approach based on a rigid body transformation of the FM measured in the reference position to the new position. RESULTS: The transformed FM provided inconsistent results for large head movements (> 5° rotation, approximately), while the simulation strategy had a superior prediction accuracy for all positions. The simulated FM was used to optimize B0 shims with up to 22.2% improvement with respect to the transformed FM approach. CONCLUSION: The proposed simulation strategy is able to predict movement-induced B0 field inhomogeneities yielding more precise estimates of the ground truth field homogeneity than the transformed FM.

Phase-based masking for quantitative susceptibility mapping of the human brain at 9.4T.

PURPOSE: To develop improved tissue masks for QSM. METHODS: Masks including voxels at the brain surface were automatically generated from the magnitude alone (MM) or combined with test functions from the first (PG) or second (PB) derivative of the sign of the wrapped phase. Phase images at 3T and 9.4T were simulated at different TEs and used to generate a mask, PItoh , with between-voxel phase differences less than π. MM, PG, and PB were compared with PItoh . QSM were generated from 3D multi-echo gradient-echo data acquired at 9.4T (21 subjects aged: 20-56y), and from the QSM2016 challenge 3T data using different masks, unwrapping, background removal, and dipole inversion algorithms. QSM contrast was quantified using age-based iron concentrations. RESULTS: Close to air cavities, phase wraps became denser with increasing field and echo time, yielding increased values of the test functions. Compared with PItoh , PB had the highest Dice coefficient, while PG had the lowest and MM the highest percentage of voxels outside PItoh. Artifacts observed in QSM at 9.4T with MM were mitigated by stronger background filters but yielded a reduced QSM contrast. With PB, QSM contrast was greater and artifacts diminished. Similar results were obtained with challenge data, evidencing larger effects of mask close to air cavities. CONCLUSION: Automatic, phase-based masking founded on the second derivative of the sign of the wrapped phase, including cortical voxels at the brain surface, was able to mitigate artifacts and restore QSM contrast across cortical and subcortical brain regions.

Developing formalin-based fixative agents for post mortem brain MRI at 9.4 T.

PURPOSE: To develop fixative agents for high-field MRI with suitable dielectric properties and measure MR properties in immersion-fixed brain tissue. METHODS: Dielectric properties of formalin-based agents were assessed (100 MHz-4.5 GHz), and four candidate fixatives with/without polyvinylpyrrolidone (PVP) and different salt concentrations were formulated. B1 field and MR properties (T1 , R2∗ , R2 , R2' , and magnetic susceptibility [QSM]) were observed in white and gray matter of pig brain samples during 0.5-35 days of immersion fixation. The kinetics were fitted using exponential functions. The immersion time required to reach maximum R2∗ values at different tissue depths was used to estimate the Medawar coefficient for fixative penetration. The effect of replacing the fixatives with Fluoroinert and phosphate-buffered saline as embedding media was also evaluated. RESULTS: The dielectric properties of formalin were nonlinearly modified by increasing amounts of additives. With 5% PVP and 0.04% NaCl, the dielectric properties and B1 field reflected in vivo conditions. The highest B1 values were found in white matter with PVP and varied significantly with tissue depth and embedding media, but not with immersion time. The MR properties depended on PVP yielding lower T1 , higher R2∗ , more paramagnetic QSM values, and a lower Medawar coefficient (0.9 mm/h ; without PVP: 1.5). Regardless of fixative, switching to phosphate-buffered saline as embedder caused a paramagnetic shift in QSM and decreased R2∗ that progressed during 1 month of storage, whereas no differences were found with Fluorinert. CONCLUSION: In vivo-like B1 fields can be achieved in formalin fixatives using PVP and a low salt concentration, yielding lower T1 , higher R2∗ , and more paramagnetic QSM than without additives. The kinetics of R2∗ allowed estimation of fixative tissue penetration.

Multi-echo gradient-recalled-echo phase unwrapping using a Nyquist sampled virtual echo train in the presence of high-field gradients.

PURPOSE: To develop a spatio-temporal approach to accurately unwrap multi-echo gradient-recalled echo phase in the presence of high-field gradients. THEORY AND METHODS: Using the virtual echo-based Nyquist sampled (VENyS) algorithm, the temporal unwrapping procedure is modified by introduction of one or more virtual echoes between the first lower and the immediate higher echo, so as to reinstate the Nyquist condition at locations with high-field gradients. An iterative extension of the VENyS algorithm maintains spatial continuity by adjusting the phase rotations to make the neighborhood phase differences less than π. The algorithm is evaluated using simulated data, Gadolinium contrast-doped phantom, and in vivo brain, abdomen, and chest data sets acquired at 3 T and 9.4 T. The unwrapping performance is compared with the standard temporal unwrapping algorithm used in the morphology-enabled dipole inversion-QSM pipeline as a benchmark for validation. RESULTS: Quantitative evaluation using numerical phantom showed significant reduction in unwrapping errors in regions of large field gradients, and the unwrapped phase revealed an exact match with the linear concentration profile of vials in a gadolinium contrast-doped phantom data acquired at 9.4 T. Without the need for additional spatial unwrapping, the iterative VENyS algorithm was able to generate spatially continuous phase images. Application to in vivo data resulted in better unwrapping performance, especially in regions with large susceptibility changes such as the air/tissue interface. CONCLUSION: The iterative VENyS algorithm serves as a robust unwrapping method for multi-echo gradient-recalled echo phase in the presence of high-field gradients.

Depth relationships and measures of tissue thickness in dorsal midbrain.

Dorsal human midbrain contains two nuclei with clear laminar organization, the superior and inferior colliculi. These nuclei extend in depth between the superficial dorsal surface of midbrain and a deep midbrain nucleus, the periaqueductal gray matter (PAG). The PAG, in turn, surrounds the cerebral aqueduct (CA). This study examined the use of two depth metrics to characterize depth and thickness relationships within dorsal midbrain using the superficial surface of midbrain and CA as references. The first utilized nearest-neighbor Euclidean distance from one reference surface, while the second used a level-set approach that combines signed distances from both reference surfaces. Both depth methods provided similar functional depth profiles generated by saccadic eye movements in a functional MRI task, confirming their efficacy for delineating depth for superficial functional activity. Next, the boundaries of the PAG were estimated using Euclidean distance together with elliptical fitting, indicating that the PAG can be readily characterized by a smooth surface surrounding PAG. Finally, we used the level-set approach to measure tissue depth between the superficial surface and the PAG, thus characterizing the variable thickness of the colliculi. Overall, this study demonstrates depth-mapping schemes for human midbrain that enables accurate segmentation of the PAG and consistent depth and thickness estimates of the superior and inferior colliculi.

In-vivo quantitative structural imaging of the human midbrain and the superior colliculus at 9.4T.

We explored anatomical details of the superior colliculus (SC) by in vivo magnetic resonance imaging (MRI) at 9.4T. The high signal-to-noise ratio allowed the acquisition of high resolution, multi-modal images with voxel sizes ranging between 176 × 132 × 600 µm and (800)3µm. Quantitative mapping of the longitudinal relaxation rate R1, the effective transverse relaxation rate R2*, and the magnetic susceptibility QSM was performed in 14 healthy volunteers. The images were analyzed in native space as well as after normalization to a common brain space (MNI). The coefficient-of-variation (CoV) across subjects was evaluated in prominent regions of the midbrain, reaching the best reproducibility (CoV of 5%) in the R2* maps of the SC in MNI space, while the CoV in the QSM maps remained high regardless of brain-space. To investigate whether more complex neurobiological architectural features could be detected, depth profiles through the SC layers towards the red nucleus (RN) were evaluated at different levels of the SC along the rostro-caudal axis. This analysis revealed alterations of the quantitative MRI parameters concordant with previous post mortem histology studies of the cyto- and myeloarchitecture of the SC. In general, the R1 maps were hyperintense in areas characterized by the presence of abundant myelinated fibers, and likely enabled detection of the deep white layer VII of the SC adjacent to the periaqueductal gray. While R1 maps failed to reveal finer details, possibly due to the relatively coarse spatial sampling used for this modality, these could be recovered in R2* maps and in QSM. In the central part of the SC along its rostro-caudal axis, increased R2* values and decreased susceptibility values were observed 2 mm below the SC surface, likely reflecting the myelinated fibers in the superficial optic layer (layer III). Towards the deeper layers, a second increase in R2* was paralleled by a paramagnetic shift in QSM suggesting the presence of an iron-rich layer about 3 mm below the surface of the SC, attributed to the intermediate gray layer (IV) composed of multipolar neurons. These results dovetail observations in histological specimens and animal studies and demonstrate that high-resolution multi-modal MRI at 9.4T can reveal several microstructural features of the SC in vivo.

"Wrong Way Up": Temporal and Spatial Dynamics of the Networks for Body Motion Processing at 9.4 T.

Body motion delivers a wealth of socially relevant information. Yet display inversion severely impedes biological motion (BM) processing. It is largely unknown how the brain circuits for BM are affected by display inversion. As upright and upside-down point-light BM displays are similar, we addressed this issue by using ultrahigh field functional MRI at 9.4 T providing for high sensitivity and spatial resolution. Whole-brain analysis along with exploration of the temporal dynamics of the blood-oxygen-level-dependent response reveals that in the left hemisphere, inverted BM activates anterior networks likely engaged in decision making and cognitive control, whereas readily recognizable upright BM activates posterior areas solely. In the right hemisphere, multiple networks are activated in response to upright BM as compared with scarce activation to inversion. With identical visual input with display inversion, a large-scale network in the right hemisphere is detected in perceivers who do not constantly interpret displays as shown the "wrong way up." For the first time, we uncover (1) (multi)functional involvement of each region in the networks underpinning BM processing and (2) large-scale ensembles of regions playing in unison with distinct temporal dynamics. The outcome sheds light on the neural circuits underlying BM processing as an essential part of the social brain.

Depth-dependence of visual signals in the human superior colliculus at 9.4 T.

The superior colliculus (SC) is a layered structure located in the midbrain. We exploited the improved spatial resolution and BOLD signal strength available at 9.4 T to investigate the depth profile of visual BOLD responses in the human SC based on distortion-corrected EPI data with a 1 mm isotropic resolution. We used high resolution (350 µm in-plane) anatomical images to determine regions-of-interest of the SC and applied a semi-automated method to segment it into superficial, intermediate, and deep zones. A greater than linear increase in sensitivity of the functional signal at 9.4 T allowed us to detect a statistically significant depth pattern in a group analysis with a 20 min stimulation paradigm. Descriptive data showed consistent depth profiles also in single individuals. The highest signals were localized to the superficial layers of the right and left SC during contralateral stimulation, which was in good agreement with its functional architecture known from non-human primates. This study thus demonstrates the potential of 9.4 T MRI for functional neuroimaging even in deeply located, particularly challenging brain structures such as the SC. Hum Brain Mapp 38:574-587, 2017. © 2016 Wiley Periodicals, Inc.

MR spectroscopy for in vivo assessment of the oncometabolite 2-hydroxyglutarate and its effects on cellular metabolism in human brain gliomas at 9.4T.

PURPOSE: To examine in vivo metabolic alterations in the isocitrate dehydrogenase (IDH) mutated gliomas using magnetic resonance spectroscopy (MRS) at magnetic field 9.4T. MATERIALS AND METHODS: Spectra were acquired with a 9.4T whole-body scanner with the use of a custom-built head coil (16 channel transmit and 31 channel receive). A modified stimulated echo acquisition mode (STEAM) sequence was used for localization. Eighteen patients with brain tumors of probable glial origin participated in this study. The study was performed in accordance with the guidelines of the local Ethics Committee. RESULTS: The increased spectral resolution allowed us to directly address metabolic alterations caused by the specific pathophysiology of IDH mutations including the presence of the oncometabolite 2-hydroxglutarate (2HG) and a significant decrease of the pooled glutamate and glutamine (20%, P = 0.024), which probably reflects an attempt to replenish α-ketoglutarate lost by conversion to 2HG. We also observed significantly reduced glutathione (GSH) levels (39%, P = 0.019), which could be similarly caused by depletion of dihydronicotinamide-adenine dinucleotide phosphate (NADPH) during this conversion in IDH mutant gliomas. CONCLUSION: We demonstrate that MRS at 9.4T provides a noninvasive measure of 2HG in vivo, which may be used for therapy planning and prognostication, and may provide insights into related pathophysiologic metabolic alterations associated with IDH mutations. J. MAGN. RESON. IMAGING 2016;44:823-833.

Functional quantitative susceptibility mapping (fQSM).

Blood oxygenation level dependent (BOLD) functional magnetic resonance imaging (fMRI) is a powerful technique, typically based on the statistical analysis of the magnitude component of the complex time-series. Here, we additionally interrogated the phase data of the fMRI time-series and used quantitative susceptibility mapping (QSM) in order to investigate the potential of functional QSM (fQSM) relative to standard magnitude BOLD fMRI. High spatial resolution data (1mm isotropic) were acquired every 3 seconds using zoomed multi-slice gradient-echo EPI collected at 7 T in single orientation (SO) and multiple orientation (MO) experiments, the latter involving 4 repetitions with the subject's head rotated relative to B0. Statistical parametric maps (SPM) were reconstructed for magnitude, phase and QSM time-series and each was subjected to detailed analysis. Several fQSM pipelines were evaluated and compared based on the relative number of voxels that were coincidentally found to be significant in QSM and magnitude SPMs (common voxels). We found that sensitivity and spatial reliability of fQSM relative to the magnitude data depended strongly on the arbitrary significance threshold defining "activated" voxels in SPMs, and on the efficiency of spatio-temporal filtering of the phase time-series. Sensitivity and spatial reliability depended slightly on whether MO or SO fQSM was performed and on the QSM calculation approach used for SO data. Our results present the potential of fQSM as a quantitative method of mapping BOLD changes. We also critically discuss the technical challenges and issues linked to this intriguing new technique.

Dual-frequency calcium-responsive MRI agents.

Responsive or smart magnetic resonance imaging (MRI) contrast agents are molecular sensors that alter the MRI signal upon changes in a particular parameter in their microenvironment. Consequently, they could be exploited for visualization of various biochemical events that take place at molecular and cellular levels. In this study, a set of dual-frequency calcium-responsive MRI agents are reported. These are paramagnetic, fluorine-containing complexes that produce remarkably high MRI signal changes at the (1)H and (19)F frequencies at varying Ca(2+) concentrations. The nature of the processes triggered by Ca(2+) was revealed, allowing a better understanding of these complex systems and their further improvement. The findings indicate that these double-frequency tracers hold great promise for development of novel functional MRI methods.

Assessment of Calcaneal Spongy Bone Magnetic Resonance Characteristics in Women: A Comparison between Measures Obtained at 0.3 T, 1.5 T, and 3.0 T.

BACKGROUND: There is a growing interest in bone tissue MRI and an even greater interest in using low-cost MR scanners. However, the characteristics of bone MRI remain to be fully defined, especially at low field strength. This study aimed to characterize the signal-to-noise ratio (SNR), T2, and T2* in spongy bone at 0.3 T, 1.5 T, and 3.0 T. Furthermore, relaxation times were characterized as a function of bone-marrow lipid/water ratio content and trabecular bone density. METHODS: Thirty-two women in total underwent an MR-imaging investigation of the calcaneus at 0.3 T, 1.5 T, and 3.0 T. MR-spectroscopy was performed at 3.0 T to assess the fat/water ratio. SNR, T2, and T2* were quantified in distinct calcaneal regions (ST, TC, and CC). ANOVA and Pearson correlation statistics were used. RESULTS: SNR increase depends on the magnetic field strength, acquisition sequence, and calcaneal location. T2* was different at 3.0 T and 1.5 T in ST, TC, and CC. Relaxation times decrease as much as the magnetic field strength increases. The significant linear correlation between relaxation times and fat/water found in healthy young is lost in osteoporotic subjects. CONCLUSION: The results have implications for the possible use of relaxation vs. lipid/water marrow content for bone quality assessment and the development of quantitative MRI diagnostics at low field strength.

MP2RAGE vs. MPRAGE surface-based morphometry in focal epilepsy.

In drug-resistant focal epilepsy, detecting epileptogenic lesions using MRI poses a critical diagnostic challenge. Here, we assessed the utility of MP2RAGE-a T1-weighted sequence with self-bias correcting properties commonly utilized in ultra-high field MRI-for the detection of epileptogenic lesions using a surface-based morphometry pipeline based on FreeSurfer, and compared it to the common approach using T1w MPRAGE, both at 3T. We included data from 32 patients with focal epilepsy (5 MRI-positive, 27 MRI-negative with lobar seizure onset hypotheses) and 94 healthy controls from two epilepsy centres. Surface-based morphological measures and intensities were extracted and evaluated in univariate GLM analyses as well as multivariate unsupervised 'novelty detection' machine learning procedures. The resulting prediction maps were analyzed over a range of possible thresholds using alternative free-response receiver operating characteristic (AFROC) methodology with respect to the concordance with predefined lesion labels or hypotheses on epileptogenic zone location. We found that MP2RAGE performs at least comparable to MPRAGE and that especially analysis of MP2RAGE image intensities may provide additional diagnostic information. Secondly, we demonstrate that unsupervised novelty-detection machine learning approaches may be useful for the detection of epileptogenic lesions (maximum AFROC AUC 0.58) when there is only a limited lesional training set available. Third, we propose a statistical method of assessing lesion localization performance in MRI-negative patients with lobar hypotheses of the epileptogenic zone based on simulation of a random guessing process as null hypothesis. Based on our findings, it appears worthwhile to study similar surface-based morphometry approaches in ultra-high field MRI (≥ 7 T).

Aryl-phosphonate lanthanide complexes and their fluorinated derivatives: investigation of their unusual relaxometric behavior and potential application as dual frequency 1H/19F†MRI probes.

A series of low molecular weight lanthanide complexes were developed that have high (1)H longitudinal relaxivities (r1) and the potential to be used as dual frequency (1)H and (19)F MR probes. Their behavior was investigated in more detail through relaxometry, pH-potentiometry, luminescence, and multinuclear NMR spectroscopy. Fitting of the (1)H NMRD and (17)O NMR profiles demonstrated a very short water residence lifetime (<10 ns) and an appreciable second sphere effect. At lower field strengths (20 MHz), two of the complexes displayed a peak in r1 (21.7 and 16.3 mM(-1) s(-1)) caused by an agglomeration, that can be disrupted through the addition of phosphate anions. NMR spectroscopy revealed that at least two species are present in solution interconverting through an intramolecular binding process. Two complexes provided a suitable signal in (19)F NMR spectroscopy and through the selection of optimized imaging parameters, phantom images were obtained in a MRI scanner at concentrations as low as 1 mM. The developed probes could be visualized through both (1)H and (19)F MRI, showing their capability to function as dual frequency MRI contrast agents.

Distribution of corpora amylacea in the human midbrain: using synchrotron radiation phase-contrast microtomography, high-field magnetic resonance imaging, and histology.

Corpora amylacea (CA) are polyglucosan aggregated granules that accumulate in the human body throughout aging. In the cerebrum, CA have been found in proximity to ventricular walls, pial surfaces, and blood vessels. However, studies showing their three-dimensional spatial distribution are sparse. In this study, volumetric images of four human brain stems were obtained with MRI and phase-contrast X-ray microtomography, followed up by Periodic acid Schiff stain for validation. CA appeared as hyperintense spheroid structures with diameters up to 30 µm. An automatic pipeline was developed to segment the CA, and the spatial distribution of over 200,000 individual corpora amylacea could be investigated. A threefold-or higher-density of CA was detected in the dorsomedial column of the periaqueductal gray (860-4,200 CA count/mm3) than in the superior colliculus (150-340 CA count/mm3). We estimated that about 2% of the CA were located in the immediate vicinity of the vessels or in the peri-vascular space. While CA in the ependymal lining of the cerebral aqueduct was rare, the sub-pial tissue of the anterior and posterior midbrain contained several CA. In the sample with the highest CA density, quantitative maps obtained with MRI revealed high R2* values and a diamagnetic shift in a region which spatially coincided with the CA dense region.

Simulated and experimental phantom data for multi-center quality assurance of quantitative susceptibility maps at 3 T, 7 T and 9.4 T.

PURPOSE: To develop methods for quality assurance of quantitative susceptibility mapping (QSM) using MRI at different magnetic field strengths, and scanners, using different MR-sequence protocols, and post-processing pipelines. METHODS: We built a custom phantom based on iron in two forms: homogeneous susceptibility ('free iron') and with fine-scaled variations in susceptibility ('clustered iron') at different iron concentrations. The phantom was measured at 3.0 T (two scanners), 7.0 T and 9.4 T using multi-echo, gradient echo acquisition sequences. A digital phantom analogue to the iron-phantom, tailored to obtain similar results as in experimentation was developed, with similar geometry and susceptibility values. Morphology enabled dipole inversion was applied to the phase images to obtain QSM for experimental and simulated data using the MEDI + 0 approach for background regularization. RESULTS: Across all scanners, QSM-values showed a linear increase with iron concentrations. The QSM-relaxivity was 0.231 ± 0.047 ppm/mM for free and 0.054 ± 0.013 ppm/mM for clustered iron, with adjusted determination coefficients (DoC) ≥ 0.87. Similarly, the simulations yielded linear increases (DoC ≥ 0.99). In both the experimental and digital phantoms, the estimated molar susceptibility was lower with clustered iron, because clustering led to highly localized field effects. CONCLUSION: Our iron phantom can be used to evaluate the capability of QSM to detect local variations in susceptibility across different field strengths, when using different MR-sequence protocols. The devised simulation method captures the effect of iron clustering in QSM as seen experimentally and could be used in the future to optimize QSM processing pipelines and achieve higher accuracy for local field effects, as also seen in Alzheimer's beta-amyloid plaques.

Phase stability in fMRI time series: effect of noise regression, off-resonance correction and spatial filtering techniques.

Although the majority of fMRI studies exploit magnitude changes only, there is an increasing interest regarding the potential additive information conveyed by the phase signal. This integrated part of the complex number furnished by the MR scanners can also be used for exploring direct detection of neuronal activity and for thermography. Few studies have explicitly addressed the issue of the available signal stability in the context of phase time-series, and therefore we explored the spatial pattern of frequency specific phase fluctuations, and evaluated the effect of physiological noise components (heart beat and respiration) on the phase signal. Three categories of retrospective noise reduction techniques were explored and the temporal signal stability was evaluated in terms of a physiologic noise model, for seven fMRI measurement protocols in eight healthy subjects at 3T, for segmented CSF, gray and white matter voxels. We confirmed that for most processing methods, an efficient use of the phase information is hampered by the fact that noise from physiological and instrumental sources contributes significantly more to the phase than to the magnitude instability. Noise regression based on the phase evolution of the central k-space point, RETROICOR, or an orthonormalized combination of these were able to reduce their impact, but without bringing phase stability down to levels expected from the magnitude signal. Similar results were obtained after targeted removal of scan-to-scan variations in the bulk magnetic field by the dynamic off-resonance in k-space (DORK) method and by the temporal off-resonance alignment of single-echo time series technique (TOAST). We found that spatial high-pass filtering was necessary, and in vivo a Gaussian filter width of 20mm was sufficient to suppress physiological noise and bring the phase fluctuations to magnitude levels. Stronger filters brought the fluctuations down to levels dictated by thermal noise contributions, and for 62.5mm(3) voxels the phase stability was as low as 5 mrad (0.27°). In conditions of low SNR(o) and high temporal sampling rate (short TR); we achieved an upper bound for the phase instabilities at 0.0017 ppm, which is close to the dHb contribution to the GM/WM phase contrast.