Simulating rigid head motion artifacts on brain magnitude MRI data-Outcome on image quality and segmentation of the cerebral cortex.

Magnetic Resonance Imaging (MRI) datasets from epidemiological studies often show a lower prevalence of motion artifacts than what is encountered in clinical practice. These artifacts can be unevenly distributed between subject groups and studies which introduces a bias that needs addressing when augmenting data for machine learning purposes. Since unreconstructed multi-channel k-space data is typically not available for population-based MRI datasets, motion simulations must be performed using signal magnitude data. There is thus a need to systematically evaluate how realistic such magnitude-based simulations are. We performed magnitude-based motion simulations on a dataset (MR-ART) from 148 subjects in which real motion-corrupted reference data was also available. The similarity of real and simulated motion was assessed by using image quality metrics (IQMs) including Coefficient of Joint Variation (CJV), Signal-to-Noise-Ratio (SNR), and Contrast-to-Noise-Ratio (CNR). An additional comparison was made by investigating the decrease in the Dice-Sørensen Coefficient (DSC) of automated segmentations with increasing motion severity. Segmentation of the cerebral cortex was performed with 6 freely available tools: FreeSurfer, BrainSuite, ANTs, SAMSEG, FastSurfer, and SynthSeg+. To better mimic the real subject motion, the original motion simulation within an existing data augmentation framework (TorchIO), was modified. This allowed a non-random motion paradigm and phase encoding direction. The mean difference in CJV/SNR/CNR between the real motion-corrupted images and our modified simulations (0.004±0.054/-0.7±1.8/-0.09±0.55) was lower than that of the original simulations (0.015±0.061/0.2±2.0/-0.29±0.62). Further, the mean difference in the DSC between the real motion-corrupted images was lower for our modified simulations (0.03±0.06) compared to the original simulations (-0.15±0.09). SynthSeg+ showed the highest robustness towards all forms of motion, real and simulated. In conclusion, reasonably realistic synthetic motion artifacts can be induced on a large-scale when only magnitude MR images are available to obtain unbiased data sets for the training of machine learning based models.

Comparison of MRI methods for measuring whole-brain oxygen extraction fraction under different geometric conditions at 7T.

BACKGROUND AND PURPOSE: Cerebral tissue oxygenation is a critical brain viability parameter, and the magnetic properties of hemoglobin offer the opportunity to noninvasively quantify oxygen extraction fraction (OEF) by magnetic resonance imaging (MRI). Ultrahigh-field MRI shows advantages such as increased sensitivity to magnetic susceptibility differences and improved signal-to-noise ratio that can be translated into smaller voxel size, but also increased sensitivity to static and B1 field inhomogeneities. The aim was to produce a systematic comparison of three MRI-based methods for estimation of OEF. METHODS: OEF estimates in 16 healthy subjects were obtained at 7T utilizing susceptometry-based oximetry (SBO), quantitative susceptibility mapping (QSM), and transverse relaxation rate (R2*). Two major draining veins, that is, the superior sagittal sinus (SSS) and the straight sinus (SS), were investigated, including mutual agreement between the methods in each of the two different vessels, agreement between vessels as well as potential vessel angle and vessel size dependences. RESULTS: Very good correlation (r = .88) was found between SBO-based and QSM-based OEF estimates in SSS. Only QSM showed a moderate correlation (r = .61) between corresponding OEF estimates in SSS and SS. For SBO, a trend of increasing OEF estimates was observed as the SS vessel angle relative to the main magnetic field increased. No obvious size dependence could be established for any method. The R2*-based OEF estimates were reasonable (35%-36%), but the observed range was somewhat low. CONCLUSION: The results indicate that QSM is a promising candidate for assessment of OEF estimates, for example, providing reasonably robust estimates across a wide range of vessel orientations.

MP3RAGE: Simultaneous mapping of T1 and B 1 + in human brain at 7T.

PURPOSE: To map T1 and the local flip angle ( B 1 + ) in human brain using a single MP3RAGE sequence with 3 rapid acquisitions of gradient echoes (RAGEs). THEORY AND METHODS: A third RAGE with a relatively high flip angle was appended to an MP2RAGE sequence. Through curve fitting and a rational approximation for small flip angles and short TR, closed form solutions for T1 and B 1 + were derived. The influence of different k-space encoding schemes on precision and whether edge enhancement artifacts could be reduced with a saturation pulse applied prior to the third RAGE were explored. Validation of T1 estimates was performed using single-slice inversion recovery (IR) and a subsequent region-of-interest-based comparison, whereas validation of B 1 + was performed using a whole brain pixelwise comparison to a DREAM flip angle mapping protocol. Lastly, MP3RAGE was compared to T1 -mapping by MP2RAGE with separate B 1 + correction. RESULTS: Whole brain maps of T1 and B 1 + at 1 mm isotropic resolution were obtained with MP3RAGE in 06:37 min. A linear-reverse centric-reverse centric phase-encoding order of the 3 RAGEs improved precision, and artifacts were successfully reduced with the saturation pulse. Estimations of T1 and B 1 + deviated +2.5 ± 3.1% and -1.7 ± 8.6% from their respective references. CONCLUSION: T1 and B 1 + can be mapped simultaneously using MP3RAGE. The approach can be thought of as combining MP2RAGE with a dual flip angle T1 -mapping protocol. Both maps can be solved for analytically and will be inherently co-registered at the high resolution associated with MPRAGE.

Radiofrequency Bias Correction of Magnetization Prepared Rapid Gradient Echo MRI at 7.0 Tesla Using an External Reference in a Sequential Protocol.

At field strengths of 7 T and above, T1-weighted imaging of human brain suffers increasingly from radiofrequency (RF) B1 inhomogeneities. The well-known MP2RAGE (magnetization prepared two rapid acquisition gradient echoes) sequence provides a solution but may not be readily available for all MR systems. Here, we describe the implementation and evaluation of a sequential protocol to obtain normalized magnetization prepared rapid gradient echo (MPRAGE) images at 0.7, 0.8, or 0.9-mm isotropic spatial resolution. Optimization focused on the reference gradient-recalled echo (GRE) that was used for normalization of the MPRAGE. A good compromise between white-gray matter contrast and the signal-to-noise ratio (SNR) was reached at a flip angle of 3° and total scan time was reduced by increasing the reference voxel size by a factor of 8 relative to the MPRAGE resolution. The average intra-subject coefficient-of-variation (CV) in segmented white matter (WM) was 7.9 ± 3.3% after normalization, compared to 20 ± 8.4% before. The corresponding inter-subject average CV in WM was 7.6 ± 7.6% and 13 ± 7.8%. Maps of T1 derived from forward signal modelling showed no obvious bias after correction by a separately acquired flip angle map. To conclude, a non-interleaved acquisition for normalization of MPRAGE offers a simple alternative to MP2RAGE to obtain semi-quantitative purely T1-weighted images. These images can be converted to T1 maps, analogously to the established MP2RAGE approach. Scan time can be reduced by increasing the reference voxel size which has only a miniscule effect on image quality.

Mapping magnetization transfer saturation (MTsat ) in human brain at 7T: Protocol optimization under specific absorption rate constraints.

PURPOSE: To optimize a whole-brain magnetization transfer saturation (MTsat ) protocol at 7T, focusing on maximizing obtainable MTsat under the constraints of specific absorption rate (SAR) and transmit field inhomogeneity, while avoiding bias and keeping scan time short. THEORY AND METHODS: MTsat is a semi-quantitative metric, obtained by spoiled gradient-echo MRI in the imaging steady-state. Optimization was based on an established 7T dual flip angle protocol, and focused on MT pulse, readout flip angle, repetition time (TR), offset frequency (Δ), and correction of residual effects from transmit field inhomogeneities by separate flip angle mapping. RESULTS: A 100% SAR level was reached at a 180° MT pulse flip angle, using a compact sinc main lobe (4 ms duration) and minimum TR = 26.5 ms. The use of Δ = +2.0 kHz caused no discernible direct saturation, while Δ = -2.0 kHz resulted in 45% higher MTsat in white matter (WM) compared to Δ = +2.0 kHz. A 4° readout flip angle eliminated bias while yielding a good signal-to-noise ratio. Increased TR yielded only a little increase in MTsat , and TR = 26.5 ms (scan time 04:58 min) was thus selected. Post hoc transmit field correction clearly improved homogeneity, especially in WM. CONCLUSIONS: The range of MTsat is limited at 7T, and this can partly be overcome by the exploitation of the asymmetry of the macromolecular lineshape through the sign of Δ. To reduce scan time, a compact MT pulse with a sufficiently narrow frequency response should be used. TR and readout flip angle should be kept short/small. Transmit field correction through separate flip angle mapping is required.

Cortical and white matter correlates of language-learning aptitudes.

People learn new languages with varying degrees of success but what are the neuroanatomical correlates of the difference in language-learning aptitude? In this study, we set out to investigate how differences in cortical morphology and white matter microstructure correlate with aptitudes for vocabulary learning, phonetic memory, and grammatical inferencing as measured by the first-language neutral LLAMA test battery. We used ultra-high field (7T) magnetic resonance imaging to estimate the cortical thickness and surface area from sub-millimeter resolved image volumes. Further, diffusion kurtosis imaging was used to map diffusion properties related to the tissue microstructure from known language-related white matter tracts. We found a correlation between cortical surface area in the left posterior-inferior precuneus and vocabulary learning aptitude, possibly indicating a greater predisposition for storing word-figure associations. Moreover, we report negative correlations between scores for phonetic memory and axial kurtosis in left arcuate fasciculus as well as mean kurtosis, axial kurtosis, and radial kurtosis of the left superior longitudinal fasciculus III, which are tracts connecting cortical areas important for phonological working memory.

Reducing bias in DREAM flip angle mapping in human brain at 7T by multiple preparation flip angles.

DREAM (Dual Refocusing Echo Acquisition Mode) is an ultra-fast multi-slice B1+-mapping technique based on the single-shot STEAM sequence. To study systematic errors at high actual flip angles (FA) and low SNR, DREAM B1+ maps at 3.75×3.75×3.50 mm3 resolution were acquired at 7T in phantoms and human brain in vivo with nominal FAs between 20° and 90° for the two STEAM preparation pulses. Predicted B1+ estimates were underestimated at actual FAs above 50° while noise was prominent below 20°. With a reliable interval of the actual FA between 20° and 50° identified, a B1+ range of 33% - 200% of nominal FA is covered by varying the nominal preparation angle through 25°, 40°, and 60°. Individual B1+ maps are thresholded according to the identified interval and combined into a single map. We demonstrate the benefit of the combined low-noise, low-bias B1+ maps for dual flip angle T1-mapping.

Reducing bias in dual flip angle T1 -mapping in human brain at 7T.

PURPOSE: To address the systematic bias in whole-brain dual flip angle (DFA) T1 -mapping at 7T by optimizing the flip angle pair and carefully selecting radiofrequency (RF) pulse shape and duration. THEORY AND METHODS: Spoiled gradient echoes can be used to estimate whole-brain maps of T1 . This can be accomplished by using only two acquisitions with different flip angles, that is, a DFA-based approach. Although DFA-based T1 -mapping is seemingly straightforward to implement, it is sensitive to bias caused by incomplete spoiling and incidental magnetization transfer effects. Further bias is introduced by the increased B0 and B1+ inhomogeneities at 7T. Experiments were performed to determine the optimal flip angle pair and appropriate RF pulse shape and duration. Obtained T1 estimates were validated using inversion recovery prepared echo planar imaging and compared to literature values. A multi-echo readout was used to increase signal-to-noise ratio, enabling quantification of R2∗ and susceptibility, χ. RESULTS: Incomplete spoiling was observed above a local flip angle of approximately 20°. An asymmetric gauss-filtered sinc pulse with a constant duration of 700 µs showed a sufficiently flat frequency response profile to avoid incomplete excitation in areas with high B0 offsets. A pulse duration of 700 µs minimized effects from incidental magnetization transfer. CONCLUSION: When performing DFA-based T1 -mapping one should (a) limit the higher flip angle to avoid incomplete spoiling, (b) use a RF pulse shape insensitive to B0 inhomogeneities and (c) apply a constant RF pulse duration, balanced to minimize incidental magnetization transfer.