eISSN 2384-1109
Open Access, Peer-reviewed
    Investig Magn Reson Imaging. 2022 Dec;26(4):208-219. English.
    Published online Dec 31, 2022.
    https://doi.org/10.13104/imri.2022.26.4.208
    Copyright © 2022 Korean Society of Magnetic Resonance in Medicine (KSMRM)
    Review

    Diffusion Encoding Methods in MRI: Perspectives and Challenges

    Alan Finkelstein,1 Xiaozhi Cao,2 Congyu Liao,2 Giovanni Schifitto,3,4,5 and Jianhui Zhong1,5,6
      • 1Department of Biomedical Engineering, University of Rochester, Rochester, NY, USA.
      • 2Department of Radiology, Stanford University, Stanford, CA, USA.
      • 3Department of Electrical Engineering, University of Rochester, Rochester, NY, USA.
      • 4Department of Neurology, University of Rochester, Rochester, NY, USA.
      • 5Department of Imaging Sciences, University of Rochester, Rochester, NY, USA.
      • 6Department of Physics and Astronomy, University of Rochester, Rochester, NY, USA.
    • Correspondence: Alan Finkelstein, MA. Department of Biomedical Engineering, University of Rochester, Rochester, NY, USA. Email: alan_finkelstein@urmc.rochester.edu
    Received September 15, 2022; Revised October 29, 2022; Accepted November 14, 2022.

    This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

    Abstract

    Diffusion MRI (dMRI) is an important imaging modality that is used extensively to diagnose and monitor diseases. dMRI measures random motion of water molecules and helps elucidate microstructural properties of tissues. Optimal diffusion encoding paradigms have been developed to reduce acquisition time, minimize artifacts, and acquire high fidelity data needed for advanced modeling of tissue properties. To further probe microstructural properties, joint diffusion-relaxometry and diffusion weighted MR fingerprinting have garnered interest. A thorough knowledge of different diffusion encoding methods is essential to accurately encode diffusion in MR experiments. Here, we review fundamental physics of diffusion encoding methods, their associated challenges, and how to address them. Advanced diffusion acquisition methods are also discussed.

    Keywords
    Diffusion MRI; Pulsed gradient spin echo; Diffusion weighted bSSFP; TRSE

    INTRODUCTION

    Diffusion MRI (dMRI) is a method to probe the molecular motion of water in biological tissues as a result of random collisions between molecules [1]. Patterns of water diffusion in the body reflect underlying tissue microstructures, such as cell membrane permeability [2], macromolecules [3], cell size and density [4, 5]. As cytoarchitecture and microenvironment change in disease, so do diffusive properties of tissues. Accordingly, dMRI has been used to study diseases in the brain [6, 7], liver [8], prostate [9], and others. Additionally, fast diffusion pulse sequences have been developed to evaluate diffusion properties in the heart [10] and high resolution dMRI has been developed to study peripheral nerve diseases [11, 12]. Therefore, development of fast, high resolution diffusion sequences without artifacts is essential for improving diagnoses and monitoring pathology.

    First described by Robert Brown in 1828 and formulated by Albert Einstein in 1905, diffusion might be interpreted as the variance of a Gaussian distribution of molecular positions over time, which is proportional to the diffusivity of species [1]. In MRI, this property is exploited to measure displacement of water over time. As spins within a voxel diffuse, there is an accumulation of phases, resulting in signal attenuation [13]. This is achieved by applying specific diffusion sensitizing gradients to probe (or “encode”) molecular motion of water. In addition, utilizing diffusion sensitizing gradients in different directions, directionality or anisotropy of water in tissues could be evaluated [14]. This provides information about microstructural properties such as cell membrane permeability [15] and axonal density [16] in the brain.

    Spin-echo (SE) single-shot echo planar imaging (SS-EPI) is the current standard for diffusion weighted (DW) image acquisition in both clinical and research settings. While EPI methods offer fast acquisition free of most motions, they suffer from several limitations such as image distortions and T2* blurring, which ultimately limit spatial resolution. In contrast, steady-state diffusion-weighted imaging (DWI) methods have been shown to overcome many limitations of SS-EPI methods. Steady-state DWI are efficient sequences that enable high resolution imaging. However, steady-state sequences suffer from complex signal dynamics because they are dependent on T1, T2, and flip angle (FA). Furthermore, shot-to-shot phase variations in steady-state sequences make them highly susceptible to motion. Recently, advanced diffusion encoding schemes such as joint diffusion-relaxometry have emerged to further probe complex microstructural properties of tissues. In this review, we will focus on basic physics and pulse sequence design of spin-echo and steady-state DWI acquisition schemes. Common challenges and solutions for each approach will be addressed. In the latter part of this review, we will focus on advanced diffusion encoding methods such as joint diffusion-relaxometry and DW MR fingerprinting (DW-MRF).

    PRINCIPLES OF DIFFUSION ENCODING

    SE Diffusion

    Current DW pulse sequences originate from the pulsed gradient spin echo (PGSE) technique (Fig. 1A) developed by Stejskal and Tanner in 1965 [17]. An SE pulse sequence is composed of a 90-degree FA followed by an inversion pulse, whereby spins in the transverse plane are rephased to generate a spin echo [18]. The PGSE method employs two strong diffusion sensitizing gradients on either side of the 180-degree inversion pulse to encode diffusion information. Phases of stationary spins are unaffected since any phase accumulation from the first gradient will be rephased by the second gradient. Spins that diffuse will accumulate a non-zero phase over time, resulting in signal attenuation. An SE is then acquired immediately after the diffusion module, usually using an EPI sequence. For PGSE, the diffusion encoding strength, b, represents the amount of diffusion weighting. It can be defined as a function of the gradient moment:

    Fig. 1
    Diffusion weighting schemes for pulsed gradient spin echo followed by an echo planar imaging (EPI) readout (A) and diffusion-weighted steady state free precession (B). TE, echo time; TR, repetition time; α, flip angle; δ, diffusion pulse length; Δ, time from the beginning of the first gradient to the beginning of the second.

    (1) b=γGδ2Δ−δ3,

    where δ is the diffusion pulse length, G is the gradient amplitude, Δ is the time from the beginning of the first gradient to the beginning of the second, and γ is the gyromagnetic ratio. A larger b-value results in a greater diffusion weighting, consequently leading to a reduced signal. In general, the diffusion-weighted signal decays exponentially as a function of the b-value and the diffusivity of the tissue:

    (2) Sb=S0e−bD,

    where D is the diffusivity of the tissue being measured. One limitation of PGSE is that the diffusion time is typically 20–40 ms, which reflects tissue boundaries on the scale of 10–15 µm. This suggests that these methods might not be effective for measuring signals from cellular or cytological components. Oscillating gradient spin echo (OGSE) is an alternative SE method with preserved diffusion sensitivity that can reduce effective diffusion time, permitting evaluation at smaller length scales [19, 20]. Multi-shell acquisitions and multi-exponential models such as intravoxel incoherent motion [21] (IVIM) and diffusion kurtosis imaging (DKI) [22] aim to expand the mono-exponential model and better resolve microstructural and perfusion properties of tissues. These are out of the scope of this review.

    SS-EPI is the most common sequence for collecting DWI images. SS-EPI aims to acquire all lines of k-space in a single shot or repetition time (TR), resulting in fast acquisition, mostly free of motion artifacts. However, signal loss at boundaries of k-space can cause a low signal-to-noise ratio (SNR) and image blurring at high spatial resolution. In addition, rapid switching of diffusion gradients in blipped EPI experiments can result in eddy currents [23]. In addition, B0 inhomogeneity coupled with low bandwidth in the phase encode direction can result in geometrical distortions [23]. Furthermore, these effects are more prominent in high field (>3 T) MR systems where B0 inhomogeneity is more severe, which might become more relevant as high field MRI becomes more prevalent. Parallel imaging methods such as sensitivity encoding (SENSE) [24] and generalized autocalibrating partial parallel acquisition (GRAPPA) [25] in combination with simultaneous multislice (SMS) [26] and compressed sensing [27] can help us reduce acquisition time and improve SNR efficiency, respectively. Multishot-EPI (MS-EPI) acquisitions are alternative methods to circumvent issues associated with SS-EPI. In multishot acquisitions, data from multiple TRs are combined into one image. However, phase-accumulation due to physiological motion such as cardiac pulsation and respiration between shots can lead to image artifacts and signal dropout. Thus, shot-to-shot phase variation between each TR can cause errors in the reconstructed image if the phase is not corrected [28]. To address motion artifacts in MS-EPI, a navigator pulse is typically acquired to monitor the phase during each diffusion weighting, allowing for larger matrix sizes and higher SNR. However, navigators do not always result in artifact free images.

    Fast spin echo (FSE) methods have been introduced to mitigate some issues of EPI-based methods, such as eddy currents and T2* blurring [29]. FSE methods use a series of refocusing pulses resulting in an echo train, which can reduce artifacts by acquiring each k-space line at the center of a spin echo. Furthermore, refocusing of phase accrual at each echo can mitigate off-resonance issues. However, SS-FSE is not as sensitive to diffusion as the PGSE, due to the Carr-Purcell-Meiboom-Gill (CPMG) condition. Bulk motion and eddy currents can lead to modulation of the phase of the transverse magnetization across multiple refocusing pulses, resulting in loss of the MG condition and artifacts such as distortion and signal attenuation. In addition, FSE methods suffer from a high specific absorption rate due to multiple refocusing pulses and a low SNR as the signal degrades over several echoes. To avoid loss of CPMG condition, a non-CPMG DW-FSE sequence has been proposed [30]. A phase insensitive rapid imaging with refocused echoes (RARE) sequence has been developed by applying a 90-degree pulse at one half TE before the first refocusing pulse, thereby eliminating the Meiboom-Gill (MG) phase condition and minimizing signal attenuation. Multishot DW-FSE using PROPELLER (periodically rotated overlapping parallel lines with enhanced reconstruction) has also been proposed to minimize artifacts such as field inhomogeneity. Alternating the phase of refocusing pulses can minimize non-CPMG pulses while auto-navigation afforded by PROPELLER can minimize motion artifacts [31]. Diffusion-prepared SS-FSE with motion compensation and gating has also shown promise [32].

    Several groups have also investigated the use of a stimulated echo acquisition mode (STEAM) sequence for diffusion encoding [33, 34, 35, 36, 37]. A stimulated echo is the result of three 90-degree pulses, whereby spins stored in the z-direction are flipped back to the transverse plane where rephasing occurs, creating the stimulated echo [38]. In a STEAM sequence, the diffusion encoding occurs at the mixing time during which minimal T2 decay occurs since magnetization is along the z-axis, allowing for encoding of very long diffusion time, particularly when the diffusing substance has a long T1 [33, 39]. The capability for long diffusion encoding time coupled with minimal loss due to T2 relaxation can be significant since in most clinical scanners, the maximum gradient strength is limited and T2 decay restricts diffusion time. The long diffusion encoding time permits the use of low gradient moments to achieve a suitable b-value while minimizing eddy currents, further reducing sensitivity to bulk motion. Furthermore, analysis of the stimulated echo permits disambiguation of T2 and diffusion contributions to MR signals [39]. Stimulated echo sequences might also be advantageous for 3D diffusion encoding. A 3D diffusion-prepared turbo spin echo sequence with a stimulated echo readout has been developed, resulting in high resolution images with higher SNR and less distortion compared to DW-EPI approaches [37].

    SE sequences suffer from long acquisition time due to dead time in the diffusion preparation (DP) module. EPI sequences are subject to artifacts such as eddy currents and susceptibility artifacts. FSE sequences suffer from low SNR as the signal decays over multiple echoes. Multi-shot methods suffer from shot-to-shot phase variations, making them highly susceptible to motion. Alternatively, steady-state free precession (SSFP)-based DWI methods as described below might mitigate some related issues.

    Steady-State and Gradient Recalled Echo Diffusion

    SSFP is a specific type of gradient recalled echo (GRE) sequence that provides an alternative approach to DWI. In SSFP sequences, a steady state is achieved by applying successive radiofrequency (RF) pulses such that the TR is much less than the T2 [40]. In this way, the MR signal is a combination of a free induction decay (FID) and echo. The signal amplitude achieves a steady state. SSFP sequences have complex spin dynamics. They are the confluence of many echo pathways, including spin echoes and stimulated echoes. The measured signal for an SSFP sequence can be manipulated by placing an unbalanced gradient either before or after the readout module. Placing an unbalanced gradient after the readout results in measuring the FID (fast imaging with steady-state free precession [FISP], gradient recalled acquisition in the steady state [GRASS], and fast field echo [FFE]) while placing the unbalanced gradient before the readout results in an echo after several RF pulses (reverse FISP [PSIF], SSFP, T2-FFE). The signal magnitude decays exponentially in EPI or FSE sequences, whereas steady-state sequences provide a high-fidelity signal of constant magnitude. In 1988, Le Bihan [41] demonstrated DWI using SSFP, allowing rapid acquisition of DW images. Le Bihan’s approach exploits the spoiler gradient present in FISP sequences to encode diffusion information, which is more sensitive to diffusion than SE-DWI.

    DW-SSFP (Fig. 1B) offers several advantages over DW-EPI methods. DW-SSFP provides higher SNRs and shorter diffusion time than diffusion-prepared SE methods, albeit at the cost of a greater motion sensitivity. Various studies have since expanded Le Bihan’s original work [41]. Early on, while the FID component was recognized to have a higher signal, it was found to have a lower sensitivity to diffusion than the echo [42]. Accordingly, DW-SSFP imaging is typically acquired after the diffusion gradient using a PSIF-like sequence. Additionally, DW-PSIF sequences have been useful for imaging the peripheral nervous system due to selective suppression of vascular flow signal [43, 44, 45]. To remove the anisotropic effect of diffusion in DW-SSFP, Kim et al. [46] have developed a quasiisotropic DW-SSFP sequence by alternating the direction of the diffusion-encoding direction every other TR, yielding images comparable to trace-weighted DWI. In DW-SSFP, the echo signal is the combination of many complicated and long echo paths, over which the diffusion weighting occurs, making it difficult to describe the b-value over the sequence, confounding conventional DWI data modeling such as the apparent diffusion coefficient (ADC) and diffusion tensor imaging (DTI) [47]. Recently, Tendler et al. [48] have modeled an equivalent b-value for DW-SSFP, deriving a closed form solution dependent on the FA and the amount of phase accumulation per TR (Eq. 3), allowing for comparison with more conventional DW sequences.

    (3) SSSFP=α,T1,T2,TR,q,D=−S01−E1E1E22sinα21−E1cosα·1−cosαE1A1+sin2α∑n=1∞E1cosαn−1A1n+1,

    where S0 is the equilibrium magnetization, E1 = e−TRT1,E2 = e−TRT2, α is the FA, A1 = e-q2TR*D,D is the diffusion coefficient, q = γGτ (γ is the gyromagnetic ratio, G is the diffusion gradient amplitude, and τ is the diffusion gradient duration). The first term and the second term in the bracket represent the spin-echo term and the stimulated-echo term, respectively. Thus, the SSFP signal is a weighted sum of spin and stimulated echo pathways. Challenges for DW-SSFP methods remain, such as the ability to disentangle diffusion information given the dependence of SSFP methods on T1 and T2 [49].

    Typically, diffusion information is encoded in the transverse magnetization using diffusion gradients with large gradient moments, which are responsible for eddy currents and image distortion. Recently, Tamada and Reeder [50] have proposed a 3D diffusion mapping method using RF phase modulated GRE imaging [51]. Spoiler gradients result in significant accumulation of diffusion information into the transverse magnetization. By incorporating an RF-modulated GRE, diffusion weighting is encoded in the phase of RF-phase modulated GRE signal [50]. Using a dictionary to map the signal, they were able to reconstruct ADC maps in 3D without eddy currents or image distortion artifacts commonly encountered in SE-EPI methods. However, the proposed approach still has sensitivity to motion, which could corrupt the steady-state signal.

    CHALLENGES

    SE and GRE methods both have their own unique challenges, which need to be considered when designing pulse sequences. Strong diffusion sensitizing gradients are known to result in bulk motion sensitivity and eddy currents [23]. In addition, low SNR due to long diffusion encoding time and partial volume effects can result in prolonged acquisition time and confound data analyses, respectively.

    Motion

    A major limitation of DW-MRI sequences is their sensitivity to motion [52]. The MR signal is complex and composed of magnitude and phase components. Gradients impart phase on a distribution of spins (isochromat) to encode positional information and diffusion weighting. In general, phase is accrued from susceptibility, field inhomogeneities, maxwell terms [53], eddy currents, chemical shift [54], and motion [55]. Bulk tissue motion will accrue a significant amount of phase and contribute to MR signal attenuation, which can corrupt the phase used to encode position. Patient motion and cardiac pulsation are two major sources of motion in DWI [56]. A Taylor expansion of the complex component of the MR signal is illustrated below:

    (4) ф=фoff+γ∫Gr→t*ro→+vo→t+12ao→t2+…dt,

    where Gr→(t) is the gradient waveform ro→ is ther spatial position function, vo→ is velocity, ao→ and is acceleration. Motion compensated diffusion encoding gradients can be explicitly designed with nulled zeroth (Eq. 5a) or first order moments (Eq. 5b) known as gradient moment nulling [57, 58]:

    (5a) M0=∫Gr→tdt,

    (5b) M1=∫Gr→t t dt.

    Early on, STEAM was used for cardiac diffusion imaging in vivo, where diffusion was encoded over the cardiac cycle during the mixing time, rendering the signal insensitive to cardiac motion [59]. However, STEAM methods are subject to low SNR. Assumptions regarding the periodicity of bulk motion might be invalid in patients with arrhythmias.

    Alternatively, motion compensation via gradient moment nulling can be achieved with a set of multipolar gradients. Dou et al. [60] originally showed that a STEAM sequence could be rendered insensitive to bulk motion by incorporating a pair of bipolar diffusion gradients with zero first-order moments in lieu of monopolar gradients conventionally used. Gamper et al. [61] further showed that similar results were achievable by replacing monopolar gradients in a PGSE sequence with flow compensated (M1) paired bipolar gradients. Similar work has been validated in the liver [62], which is affected by cardiac and respiratory motion [63, 64]. Additionally, rather than employing an EPI readout, SSFP can be used as a readout module coupled with a PGSE DP module, effectively minimizing the shot-to-shot signal variation [65]. In addition to M1 compensation, in human myocardium, bulk motion might be present with nonzero acceleration (M2). Nguyen et al. [66] have achieved bulk motion compensation in the heart using an M1M2 compensated diffusion-prepared balanced SSFP (bSSFP) acquisition. In their work, they modified a twice-refocused spin echo (TRSE) diffusion-prep module by replacing all monopolar gradients with two sets of paired bipolar gradients to null the first and second gradient moments. The use of bipolar gradients eliminated all phase sensitivity to the first-moment velocity motion and achieved comparable ADC maps to those achieved with cardiac triggering. Their M1M2 compensation approach yielded ADC maps with more realistic values and less motion compared to the TRSE DP without motion compensation. While effective, TRSE and the M1M2 compensated variant proposed require longer DP time, accordingly, needing longer acquisition time and TE (lower SNR). In addition, the use of spoilers after the tip-up pulse in DP approaches results in shot-to-shot magnitude inconsistencies, leading to signal drop-out and aliasing. To address magnitude variations in DP methods, Gao et al. [67] have developed a multi-shot DP magnitude stabilized bSSFP for distortion-free diffusion imaging. In their work, they employed a magnitude stabilizer prior to the tip up pulse at the end of the DP module to store the phase difference in the z-direction. This phase difference is recalled or rewound during each readout. In this way, magnitude errors due to phase are mitigated, but at a 50% reduction in SNR. In a DW-SSFP sequence, Cheung and Wu [68] have similarly included two pairs of bipolar diffusion encoding gradients to overcome motion artifacts encountered in bSSFP sequences, yielding DW images with relative motion insensitivity. A limitation of motion compensation approaches is that they can result in significantly longer DP time, leading to longer TEs and lower SNR. To address this issue, Aliotta et al. [69] have employed a convex optimized diffusion encoding scheme to minimize the echo time while compensating for bulk motion. They performed convex optimization for monopolar, M1 compensated, and M1M2 compensated sequences and achieved reductions in TE of 7, 25, and 17 ms, respectively, compared to equivalent sequences without convex optimization. A similar approach can be employed for eddy-current nulling [70].

    Navigators represent an alternative approach for tracking motion, including acquiring an additional navigator or self-navigation methods. Navigators can be used to correct shot-to-shot phase variations at the expense of additional acquisition time. Self-navigation techniques oversample the center of k-space and use that information to estimate motion directly from the acquired image. Iterative model based reconstructions have been incorporated into spiral [71] and EPI [72] DWI acquisitions, producing images devoid of motion artifact without compromising scan time or SNR. PROPELLER is one such technique. Blades are rotated around k-space, while the center of k-space is oversampled, allowing the motion between blades to be used to correct each acquisition block during reconstruction [73]. Alternative schemes such as multiple overlapping k-space junctions for investigating translating objects (MOJITO) [74] and radial trajectories can also be used to correct for motion. In addition, self-navigated interleaved spiral (SNAILS) [75, 76] using a variable-density spiral (VDS) oversamples the center of k-space to estimate the background phase of each spiral interleaf, thereby removing the background phase by incorporating the phase term into the forward model of the parallel imaging reconstruction [77]. VDS acquisitions also benefit from having a much smaller gradient moment at the center of k-space where contrast, and hence diffusion information, is the highest [78], providing a higher SNR than EPI [79]. Spiral acquisitions can also be performed in a single shot or multiple shots by combining interleaved spiral acquisitions. This allows a fast approach to acquiring high-resolution DW images with inherent motion correction. However, inverse problems are often ill-conditioned. They may require further regularization such as a total variational regularization [80] to stabilize the problem and result in convergence. Additionally, low-rank methods are widely used to remove shot-to-shot phase variations implicitly in either k-space [81] or image space [72, 82].

    Eddy Currents, Susceptibility & SNR

    Rapid switching of large gradients in blipped EPI sequences and strong diffusion encoding gradients can result in eddy currents. The time-varying magnetic field of gradients can induce a current in conduction surfaces of the scanner, which in turn re-establishes magnetic field gradients that persist after diffusion gradients are turned off. Eddy currents ultimately result in concomitant fields which cause unwanted phase effects and image distortion. In addition, this can lead to mismatches in voxel sizes of different DW images, resulting in inaccurate ADC estimates [83]. Eddy currents might be corrected by employing compensated gradient waveforms, including pre-emphasis components to counteract concomitant fields [84]. Eddy currents may also be attenuated by adding a second refocusing pulse to the PGSE SE DW-MRI experiment. TRSE sequence employs an additional refocusing pulse, which allows splitting of field gradients into short pules of alternating polarity, thereby cancelling eddy currents caused by residual fields. Originally used in combination with an EPI acquisition, a TRSE DP module has also been used with bSSFP [67]. However, as with all DP approaches, additional refocusing pulses can result in a longer acquisition time, longer TE, and signal attenuation. In addition, various post-processing methods exist, such as the FMRIB software library (FSL) [85] eddy correction tool [86], which uses a Gaussian model to estimate displacement maps of each-diffusion direction to correct eddy current induced image distortions.

    SS-EPI methods suffer from relatively poor SNR, limiting the spatial resolution of DWI data. 3D acquisition is a promising approach to increasing SNR. However, 3D multishot methods remain challenging due to motion-induced phase errors. 3D multislab acquisition is an alternative approach associated with high SNR efficiency and short TR [87, 88]. A 2D navigator is more easily acquired than a 3D navigator. It can be used to correct for phase errors between slabs. Ultrahigh field MRI at 7 T and above is also a promising approach to increase intrinsic SNR of DWI signals [89]. However, several challenges including RF field inhomogeneity and B0 inhomogeneity remain.

    Susceptibility artifacts in EPI are caused by local field variations in the lengthy EPI readout coupled with low bandwidth in the phase encoding direction. Blipped up/down acquisitions may be acquired and reconstructed using software packages such a FSL ‘TOPUP’ to estimate and correct distortions [90]. In addition, model-based reconstruction strategies may be employed, directly reconstructing distortion free images. Liao et al. [91] have developed a blip up-down acquisition (BUDA) combined with a multislab acquisition, resulting in submillimeter-isotropic resolution distortion free dMRI images.

    FUTURE PERSPECTIVES

    Advanced Diffusion Encoding

    Various factors such as partial volume effects, crossing fibers, edema, and demyelination can affect parameters derived from conventional diffusion encoding methods such as PGSE and DW-SSFP [92]. Advanced processing techniques and models have been proposed [93, 94], yet single diffusion encoding methods ultimately fail to capture microstructural subtleties. Alternative encoding methods such as double diffusion encoding sequences [95] can provide additional information regarding cytoarchitecture. Alternatively, in contrast to pulsed gradients, time varying gradients might be used to sample a q-space trajectory, rather than a single point in q-space. This method, known as q-space trajectory encoding, can better discriminate tissue microenvironments [96]. Thus, microscopic FA, independent of orientation dispersion, can better quantify regions with crossing fibers where the conventional DTI model is inadequate [97]. Combination of linear tensor encoding (LTE) and spherical tensor encoding (STE) or b-tensor encoding further provides a measure of microscopic anisotropy [97], independent of the fiber orientation distribution function in a voxel. To address sensitivity to incoherent motion in STE, Szczepankiewicz et al. [98] have developed motion-compensated waveforms using a constrained numerical optimization framework. Moreover, motion-compensated tensor-valued gradient waveforms have been employed for cardiac imaging, yielding multidirectional diffusion data uncorrupted by motion [99]. Ultimately, advanced diffusion encoding methods such as q-space trajectory imaging and b-tensor encoding can more efficiently sample q-space than conventional LTE paradigms. This enables shorter acquisition time and higher angular resolution compared to LTE to better resolve microstructural tissue properties.

    Diffusion-Relaxometry

    Multidimensional correlation imaging is an emerging method to refine microstructural components based on joint distribution between two or more MR biophysical parameters. Joint diffusion-relaxation has provided new insights about properties of tissue microstructure. In particularly, T1-diffusion has been used to resolve crossing fibers [100] and T2-diffusion has been used to separate distinct axonal compartments [101]. Currently, diffusion-relaxation experiments are hindered by long acquisition time and complex interrelationship between diffusion and relaxation parameters. Hutter et al. [102] have developed an integrated diffusion-relaxation sequence called ZEBRA (Z-location shuffling, multiple Echos and B-interleaving for Relaxometry-diffusion Acquisitions) using slice interleaving. Fair et al. [103] have developed an alternative approach, incorporating a PGSE diffusion module with a PROPELLER echo-planar time-resolved imaging (EPTI) with dynamic encoding (PEPTIDE), allowing for joint diffusion imaging with simultaneous T2 and T2* mapping. EPTI can minimize B0-inhomogenity phase effects, while PROPELLER can minimize motion artifacts caused by shot-to-shot phase variations (Fig. 2). In addition to data sampling schemes, optimized diffusion modules could reduce acquisition time. Diffusion weighted bSSFP sequences might afford a means to encode T1, T2, and diffusion information simultaneously, drastically reducing acquisition time. As new analysis methods further refine our understanding of fractional component maps [104, 105], further reductions in acquisition time are necessary to facilitate clinical adoption of diffusion-relaxometry techniques.

    Fig. 2
    Diffusion-encoding gradient waveforms with conventional monopolar pulsed gradient spin echo (PGSE) (A), conventional bipolar PGSE (M1 nulled, velocity insensitive) (B), and M1M2 nulled PGSE (velocity and acceleration insensitive) (C). TE, echo time.

    MRF is a framework for encoding multiple biophysical tissue properties simultaneously [106]. MRF utilizes an SSFP type sequence with varying acquisition parameters such as FA and TR, resulting in a unique signal or fingerprint for each combination of biophysical parameters. MRF was originally performed using a bSSFP sequence. It was quickly adapted to an FISP sequence to mitigate off-resonance effects encountered by bSSFP [107]. However, Kobayashi and Terada [108] have noted that failure to account for diffusion encoding caused by the additional spoiler gradients can bias T2 measurements. Accordingly, MRF has been extended beyond simply T1, T2, and proton density to encode other tissue parameters such as diffusivity, velocity [109], and T2* [110]. Recent works have sought to explicitly encode diffusion information in MRF experiments. However, several limitations remain. Diffusion encoding results in shot-to-shot phase variations during diffusion encoding, results in undesirable magnitude variations in the data. Cao et al. [111] have addressed this by employing an M1 motion compensated DP pulse in combination with cardiac gating. The M1-compensated DP can mitigate motion from constant-velocity motion, albeit with an increased preparation time. Navigator echoes could also be used to evaluate excessive motion. M1-compensated with cardiac trigger and navigator can result in less phase variation in diffusion maps than without either M1-compensation or cardiac trigger. This resulted in whole brain 3D MRF in approximately 6 minutes. However, this work was limited to diffusion in one direction only. Pirk et al. [112] have used deep learning for joint relaxation and diffusion tensor MR fingerprinting. In their work, they exploited the spoiler gradient in MRF-FISP experiments to encode diffusion information. They randomly varied the magnitude of gradients for Gx, Gy, and Gz to sample all q-space. Images were then reconstructed using a neural network. The network was trained using ground truth data obtained using multi-inversion and multi-echo spin-echo relaxometry for T1 and T2 mapping, respectively. An SS-EPI DTI sequence was used to obtain diffusion tensor data. Using this approach, they were able to inherently model shot-to-shot phase variations in the data. However, these approaches require significant amounts of data. Thus, they might not be applicable in pathology. Deep learning may circumvent this problem using simulated data to estimate directionless parameters such as ADC [113]. Afzali et al. [114] have recently shown that mapping of T1, T2, and ADC was feasible using MRF from a single scan. They employed linear and spherical b-tensor encoding (LTE and STE) to encode T1, T2, and diffusion. The pulse sequence was composed of five different preparatory modules. They also compared three different readout paradigms to mitigate phase errors. They compared FISP-based MRF, fast low-angle shot MRF with RF and gradient spoiling, and phase stabilizers used in diffusion prep sequences. However, their current method was limited to a single slice, taking approximately 24 s/slice to acquire. Despite current limitations, DW-MRF can encode complex signal dynamics by simultaneously encoding multiple biophysical parameters. Conversely, conventional diffusion-relaxation methods can sample a multiparametric space acquiring diffusion-weighted and relaxometry data independent of each other. Accordingly, DW-MRF is a promising approach to reduce acquisition time while concomitantly capturing complex signal dynamics more representative of the underlying relationships between biophysical parameters.

    CONCLUSION

    Diffusion MRI has dramatically impacted research and clinical diagnosis. Currently, there are many different methods of encoding diffusion, each with their own advantages and unique challenges. Spin echo techniques such as SS-EPI remain the clinical standard albeit they have relatively poor SNR, spatial resolution, and image artifacts. Variations such as multishot-EPI and FSE methods in combination with parallel imaging provide an alternative to circumvent issues encountered by SS-EPI. GRE based methods such as DW-SSFP offer great potential for fast imaging with high resolution and SNR, although they suffer from shot-to-shot phase variations. Advances in acquisition and reconstruction techniques aim to resolve these problems and yield fast imaging methods devoid of artifacts. Further development of advanced diffusion encoding schemes has the potential to offer nuanced insights into the microstructure of various tissue pathologies. In addition, higher fidelity data acquisition will improve diagnostic accuracy in a clinical setting. Ultimately, the development of optimized diffusion encoding schemes will facilitate diffusion-relaxometry methods clinically to better resolve microstructural properties in tissues.

    Notes

    Conflicts of Interest:The authors have no potential conflicts of interest to disclose.

    Author Contributions:

    • Conceptualization: Alan Finkelstein, Jianhui Zhong.

    • Funding Acquisition: Giovanni Schifitto.

    • Supervision: Giovanni Schifitto, Jianhui Zhong.

    • Writing—original draft: Alan Finkelstein.

    • Writing—review & editing: all authors.

    • Approval of final manuscript: all authors.

    Funding Statement:This work was supported by NIH grants R01MH118020 and R01AG054328.

    Availability of Data and Material

    Data sharing does not apply to this article as no datasets were generated or analyzed during the current study.

    References

      1. Le Bihan D, Iima M. Diffusion magnetic resonance imaging: what water tells us about biological tissues. PLoS Biol 2015;13:e1002203
      1. Li H, Jiang X, Xie J, McIntyre JO, Gore JC, Xu J. Time-dependent influence of cell membrane permeability on MR diffusion measurements. Magn Reson Med 2016;75:1927–1934.
      1. Goebell E, Fiehler J, Siemonsen S, et al. Macromolecule content influences proton diffusibility in gliomas. Eur Radiol 2011;21:2626–2632.
      1. Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002;224:177–183.
      1. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999;9:53–60.
      1. Tournier JD. Diffusion MRI in the brain - Theory and concepts. Prog Nucl Magn Reson Spectrosc 2019;112-113:1–16.
      1. Alexander DC, Dyrby TB, Nilsson M, Zhang H. Imaging brain microstructure with diffusion MRI: practicality and applications. NMR Biomed 2019;32:e3841
      1. Kele PG, van der Jagt EJ. Diffusion weighted imaging in the liver. World J Gastroenterol 2010;16:1567–1576.
      1. Haider MA, van der Kwast TH, Tanguay J, et al. Combined T2-weighted and diffusion-weighted MRI for localization of prostate cancer. AJR Am J Roentgenol 2007;189:323–328.
      1. Mekkaoui C, Reese TG, Jackowski MP, Bhat H, Sosnovik DE. Diffusion MRI in the heart. NMR Biomed 2017;30:e3426
      1. Takahara T, Hendrikse J, Kwee TC, et al. Diffusion-weighted MR neurography of the sacral plexus with unidirectional motion probing gradients. Eur Radiol 2010;20:1221–1226.
      1. Zhou Y, Narayana PA, Kumaravel M, Athar P, Patel VS, Sheikh KA. High resolution diffusion tensor imaging of human nerves in forearm. J Magn Reson Imaging 2014;39:1374–1383.
      1. Kiselev VG. Fundamentals of diffusion MRI physics. NMR Biomed 2017;30:e3602
      1. Alexander AL, Lee JE, Lazar M, Field AS. Diffusion tensor imaging of the brain. Neurotherapeutics 2007;4:316–329.
      1. Sen PN, Basser PJ. A model for diffusion in white matter in the brain. Biophys J 2005;89:2927–2938.
      1. Beaulieu C. The basis of anisotropic water diffusion in the nervous system - a technical review. NMR Biomed 2002;15:435–455.
      1. Stejskal EO, Tanner JE. Spin diffusion measurements: spin echoes in the presence of a time-dependent field gradient. J Chem Phys 1965;42:288–292.
      1. Jung BA, Weigel M. Spin echo magnetic resonance imaging. J Magn Reson Imaging 2013;37:805–817.
      1. Drobnjak I, Zhang H, Ianuş A, Kaden E, Alexander DC. PGSE, OGSE, and sensitivity to axon diameter in diffusion MRI: insight from a simulation study. Magn Reson Med 2016;75:688–700.
      1. Baron CA, Beaulieu C. Oscillating gradient spin-echo (OGSE) diffusion tensor imaging of the human brain. Magn Reson Med 2014;72:726–736.
      1. Le Bihan D. What can we see with IVIM MRI? Neuroimage 2019;187:56–67.
      1. Steven AJ, Zhuo J, Melhem ER. Diffusion kurtosis imaging: an emerging technique for evaluating the microstructural environment of the brain. AJR Am J Roentgenol 2014;202:W26–W33.
      1. Le Bihan D, Poupon C, Amadon A, Lethimonnier F. Artifacts and pitfalls in diffusion MRI. J Magn Reson Imaging 2006;24:478–488.
      1. Hu J, Li M, Dai Y, et al. Combining SENSE and reduced field-of-view for high-resolution diffusion weighted magnetic resonance imaging. Biomed Eng Online 2018;17:77
      1. Nana R, Zhao T, Hu X. Single-shot multiecho parallel echo-planar imaging (EPI) for diffusion tensor imaging (DTI) with improved signal-to-noise ratio (SNR) and reduced distortion. Magn Reson Med 2008;60:1512–1517.
      1. Ho M, Becker A, Ulbrich E, et al. Comparison of simultaneous multi-slice readout-segmented EPI and conventional single-shot EPI for diffusion tensor imaging of the ulnar nerve. Heliyon 2018;4:e00853
      1. Tamada T, Ueda Y, Kido A, et al. Clinical application of single-shot echo-planar diffusion-weighted imaging with compressed SENSE in prostate MRI at 3T: preliminary experience. Magn Reson Mater Phy 2022;35:549–556.
      1. Feinberg DA, Oshio K. Phase errors in multi-shot echo planar imaging. Magn Reson Med 1994;32:535–539.
      1. Sarlls JE, Pierpaoli C. Diffusion-weighted radial fast spin-echo for high-resolution diffusion tensor imaging at 3T. Magn Reson Med 2008;60:270–276.
      1. Gibbons EK, Le Roux P, Vasanawala SS, Pauly JM, Kerr AB. Body diffusion weighted imaging using non-CPMG fast spin echo. IEEE Trans Med Imaging 2017;36:549–559.
      1. Pipe JG, Farthing VG, Forbes KP. Multishot diffusion-weighted FSE using PROPELLER MRI. Magn Reson Med 2002;47:42–52.
      1. Lee SY, Meyer BP, Kurpad SN, Budde MD. Diffusion-prepared fast spin echo for artifact-free spinal cord imaging. Magn Reson Med 2021;86:984–994.
      1. Tanner JE. Use of the stimulated echo in NMR diffusion studies. J Chem Phys 1970;52:2523–2526.
      1. Noehren B, Andersen A, Feiweier T, Damon B, Hardy P. Comparison of twice refocused spin echo versus stimulated echo diffusion tensor imaging for tracking muscle fibers. J Magn Reson Imaging 2015;41:624–632.
      1. Zhang Y, Wells SA, Triche BL, Kelcz F, Hernando D. Stimulated-echo diffusion-weighted imaging with moderate b values for the detection of prostate cancer. Eur Radiol 2020;30:3236–3244.
      1. Zhang H, Sun A, Li H, Saiviroonporn P, Wu EX, Guo H. Stimulated echo diffusion weighted imaging of the liver at 3 tesla. Magn Reson Med 2017;77:300–309.
      1. Zhang Q, Coolen BF, Versluis MJ, Strijkers GJ, Nederveen AJ. Diffusion-prepared stimulated-echo turbo spin echo (DPsti-TSE): an eddy current-insensitive sequence for three-dimensional high-resolution and undistorted diffusion-weighted imaging. NMR Biomed 2017;30:e3719
      1. Burstein D. Stimulated echoes: description, applications, practical hints. Concepts Magn Reson 1996;8:269–278.
      1. Zhang Y, Wells SA, Hernando D. Stimulated echo based mapping (STEM) of T1, T2, and apparent diffusion coefficient: validation and protocol optimization. Magn Reson Med 2019;81:167–181.
      1. Bieri O, Scheffler K. Fundamentals of balanced steady state free precession MRI. J Magn Reson Imaging 2013;38:2–11.
      1. Le Bihan D. Intravoxel incoherent motion imaging using steady-state free precession. Magn Reson Med 1988;7:346–351.
      1. Buxton RB. The diffusion sensitivity of fast steady-state free precession imaging. Magn Reson Med 1993;29:235–243.
      1. Chhabra A, Soldatos T, Subhawong TK, et al. The application of three-dimensional diffusion-weighted PSIF technique in peripheral nerve imaging of the distal extremities. J Magn Reson Imaging 2011;34:962–967.
      1. Chhabra A, Subhawong TK, Bizzell C, Flammang A, Soldatos T. 3T MR neurography using three-dimensional diffusion-weighted PSIF: technical issues and advantages. Skeletal Radiol 2011;40:1355–1360.
      1. Zare M, Faeghi F, Hosseini A, Ardekani MS, Heidari MH, Zarei E. Comparison between three-dimensional diffusion-weighted PSIF technique and routine imaging sequences in evaluation of peripheral nerves in healthy people. Basic Clin Neurosci 2018;9:65–71.
      1. Kim TS, Gold GE, Pauly JM. Isotropic steady-state diffusion-weighted imaging; Proceedings of the ISMRM 14th Scientific Meeting & Exhibition; 2006 May 6-12; Seattle, WA, USA. ISMRM; 2006. pp. 1051.
      1. O’Donnell LJ, Westin CF. An introduction to diffusion tensor image analysis. Neurosurg Clin N Am 2011;22:185–196. viii.
      1. Tendler BC, Foxley S, Cottaar M, Jbabdi S, Miller KL. Modeling an equivalent b-value in diffusion-weighted steady-state free precession. Magn Reson Med 2020;84:873–884.
      1. Biffar A, Baur-Melnyk A, Schmidt GP, Reiser MF, Dietrich O. Quantitative analysis of the diffusion-weighted steady-state free precession signal in vertebral bone marrow lesions. Invest Radiol 2011;46:601–609.
      1. Tamada D, Reeder SB. Phase-based 3D diffusion mapping using RF phase-modulated gradient echo imaging; Proceedings of the Joint Annual Meeting ISMRM-ESMRMB & ISMRT 31st Annual Meeting; 2022 May 7-12; London, United Kingdom. ISMRM; 2022.
      1. Wang X, Hernando D, Reeder SB. Phase-based T2 mapping with gradient echo imaging. Magn Reson Med 2020;84:609–619.
      1. Norris DG. Implications of bulk motion for diffusion-weighted imaging experiments: effects, mechanisms, and solutions. J Magn Reson Imaging 2001;13:486–495.
      1. Bernstein MA, Zhou XJ, Polzin JA, et al. Concomitant gradient terms in phase contrast MR: analysis and correction. Magn Reson Med 1998;39:300–308.
      1. Hood MN, Ho VB, Smirniotopoulos JG, Szumowski J. Chemical shift: the artifact and clinical tool revisited. Radiographics 1999;19:357–371.
      1. Zaitsev M, Maclaren J, Herbst M. Motion artifacts in MRI: a complex problem with many partial solutions. J Magn Reson Imaging 2015;42:887–901.
      1. Bopp MHA, Yang J, Nimsky C, Carl B. The effect of pulsatile motion and cardiac-gating on reconstruction and diffusion tensor properties of the corticospinal tract. Sci Rep 2018;8:11204
      1. Hinks RS, Constable RT. Gradient moment nulling in fast spin echo. Magn Reson Med 1994;32:698–706.
      1. Ehman RL, Felmlee JP. Flow artifact reduction in MRI: a review of the roles of gradient moment nulling and spatial presaturation. Magn Reson Med 1990;14:293–307.
      1. Frahm J, Hänicke W, Bruhn H, Gyngell ML, Merboldt KD. High-speed STEAM MRI of the human heart. Magn Reson Med 1991;22:133–142.
      1. Dou J, Reese TG, Tseng WY, Wedeen VJ. Cardiac diffusion MRI without motion effects. Magn Reson Med 2002;48:105–114.
      1. Gamper U, Boesiger P, Kozerke S. Diffusion imaging of the in vivo heart using spin echoes--considerations on bulk motion sensitivity. Magn Reson Med 2007;57:331–337.
      1. Ozaki M, Inoue Y, Miyati T, et al. Motion artifact reduction of diffusion-weighted MRI of the liver: use of velocity-compensated diffusion gradients combined with tetrahedral gradients. J Magn Reson Imaging 2013;37:172–178.
      1. Kwee TC, Takahara T, Niwa T, et al. Influence of cardiac motion on diffusion-weighted magnetic resonance imaging of the liver. MAGMA 2009;22:319–325.
      1. Nasu K, Kuroki Y, Sekiguchi R, Nawano S. The effect of simultaneous use of respiratory triggering in diffusion-weighted imaging of the liver. Magn Reson Med Sci 2006;5:129–136.
      1. Lu L, Erokwu B, Lee G, et al. Diffusion-prepared fast imaging with steady-state free precession (DP-FISP): a rapid diffusion MRI technique at 7 T. Magn Reson Med 2012;68:868–873.
      1. Nguyen C, Fan Z, Sharif B, et al. In vivo three-dimensional high resolution cardiac diffusion-weighted MRI: a motion compensated diffusion-prepared balanced steady-state free precession approach. Magn Reson Med 2014;72:1257–1267.
      1. Gao Y, Han F, Zhou Z, et al. Multishot diffusion-prepared magnitude-stabilized balanced steady-state free precession sequence for distortion-free diffusion imaging. Magn Reson Med 2019;81:2374–2384.
      1. Cheung MM, Wu EX. Diffusion imaging with balanced steady state free precession; Proceedings of the 34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC); 2012 Aug 28-Sep 1; San Diego, CA, USA. IEEE; 2012. pp. 90-93.
      1. Aliotta E, Wu HH, Ennis DB. Convex optimized diffusion encoding (CODE) gradient waveforms for minimum echo time and bulk motion-compensated diffusion-weighted MRI. Magn Reson Med 2017;77:717–729.
      1. Aliotta E, Moulin K, Ennis DB. Eddy current-nulled convex optimized diffusion encoding (EN-CODE) for distortion-free diffusion tensor imaging with short echo times. Magn Reson Med 2018;79:663–672.
      1. Truong TK, Chen NK, Song AW. Inherent correction of motion-induced phase errors in multishot spiral diffusion-weighted imaging. Magn Reson Med 2012;68:1255–1261.
      1. Chen NK, Guidon A, Chang HC, Song AW. A robust multi-shot scan strategy for high-resolution diffusion weighted MRI enabled by multiplexed sensitivity-encoding (MUSE). Neuroimage 2013;72:41–47.
      1. Wang FN, Huang TY, Lin FH, et al. PROPELLER EPI: an MRI technique suitable for diffusion tensor imaging at high field strength with reduced geometric distortions. Magn Reson Med 2005;54:1232–1240.
      1. Bookwalter CA, Griswold MA, Duerk JL. Multiple overlapping k-space junctions for investigating translating objects (MOJITO). IEEE Trans Med Imaging 2010;29:339–349.
      1. Glover GH, Lai S. Self-navigated spiral fMRI: interleaved versus single-shot. Magn Reson Med 1998;39:361–368.
      1. Liu C, Bammer R, Kim DH, Moseley ME. Self-navigated interleaved spiral (SNAILS): application to high-resolution diffusion tensor imaging. Magn Reson Med 2004;52:1388–1396.
      1. Li TQ, Kim DH, Moseley ME. High-resolution diffusion-weighted imaging with interleaved variable-density spiral acquisitions. J Magn Reson Imaging 2005;21:468–475.
      1. Chang C, Glover GH. Variable-density spiral-in/out functional magnetic resonance imaging. Magn Reson Med 2011;65:1287–1296.
      1. Lee Y, Wilm BJ, Brunner DO, et al. On the signal-to-noise ratio benefit of spiral acquisition in diffusion MRI. Magn Reson Med 2021;85:1924–1937.
      1. Hu Z, Ma X, Truong TK, Song AW, Guo H. Phase-updated regularized SENSE for navigator-free multishot diffusion imaging. Magn Reson Med 2017;78:172–181.
      1. Haldar JP. Low-rank modeling of local k-space neighborhoods (LORAKS) for constrained MRI. IEEE Trans Med Imaging 2014;33:668–681.
      1. Hu Y, Levine EG, Tian Q, et al. Motion-robust reconstruction of multishot diffusion-weighted images without phase estimation through locally low-rank regularization. Magn Reson Med 2019;81:1181–1190.
      1. Bodammer N, Kaufmann J, Kanowski M, Tempelmann C. Eddy current correction in diffusion-weighted imaging using pairs of images acquired with opposite diffusion gradient polarity. Magn Reson Med 2004;51:188–193.
      1. Spees WM, Buhl N, Sun P, Ackerman JJ, Neil JJ, Garbow JR. Quantification and compensation of eddy-current-induced magnetic-field gradients. J Magn Reson 2011;212:116–123.
      1. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004;23 Suppl 1:S208–S219.
      1. Andersson JLR, Sotiropoulos SN. An integrated approach to correction for off-resonance effects and subject movement in diffusion MR imaging. Neuroimage 2016;125:1063–1078.
      1. Engström M, Skare S. Diffusion-weighted 3D multislab echo planar imaging for high signal-to-noise ratio efficiency and isotropic image resolution. Magn Reson Med 2013;70:1507–1514.
      1. Frost R, Miller KL, Tijssen RH, Porter DA, Jezzard P. 3D multi-slab diffusion-weighted readout-segmented EPI with real-time cardiac-reordered K-space acquisition. Magn Reson Med 2014;72:1565–1579.
      1. Eichner C, Setsompop K, Koopmans PJ, et al. Slice accelerated diffusion-weighted imaging at ultra-high field strength. Magn Reson Med 2014;71:1518–1525.
      1. Andersson JL, Skare S, Ashburner J. How to correct susceptibility distortions in spin-echo echo-planar images: application to diffusion tensor imaging. Neuroimage 2003;20:870–888.
      1. Liao C, Bilgic B, Tian Q, et al. Distortion-free, high-isotropic-resolution diffusion MRI with gSlider BUDA-EPI and multicoil dynamic B0 shimming. Magn Reson Med 2021;86:791–803.
      1. Jeurissen B, Leemans A, Tournier JD, Jones DK, Sijbers J. Investigating the prevalence of complex fiber configurations in white matter tissue with diffusion magnetic resonance imaging. Hum Brain Mapp 2013;34:2747–2766.
      1. Zhang H, Schneider T, Wheeler-Kingshott CA, Alexander DC. NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. Neuroimage 2012;61:1000–1016.
      1. Assaf Y, Basser PJ. Composite hindered and restricted model of diffusion (CHARMED) MR imaging of the human brain. Neuroimage 2005;27:48–58.
      1. Henriques RN, Palombo M, Jespersen SN, Shemesh N, Lundell H, Ianuş A. Double diffusion encoding and applications for biomedical imaging. J Neurosci Methods 2021;348:108989
      1. Westin CF, Knutsson H, Pasternak O, et al. Q-space trajectory imaging for multidimensional diffusion MRI of the human brain. Neuroimage 2016;135:345–362.
      1. Lasič S, Szczepankiewicz F, Eriksson S, Nilsson M, Topgaard D. Microanisotropy imaging: quantification of microscopic diffusion anisotropy and orientational order parameter by diffusion MRI with magic-angle spinning of the q-vector. Front Phys 2014;2:11
      1. Szczepankiewicz F, Sjölund J, Dall’Armellina E, et al. Motion-compensated gradient waveforms for tensor-valued diffusion encoding by constrained numerical optimization. Magn Reson Med 2021;85:2117–2126.
      1. Lasič S, Szczepankiewicz F, Dall’Armellina E, et al. Motion-compensated b-tensor encoding for in vivo cardiac diffusion-weighted imaging. NMR Biomed 2020;33:e4213
      1. De Santis S, Barazany D, Jones DK, Assaf Y. Resolving relaxometry and diffusion properties within the same voxel in the presence of crossing fibres by combining inversion recovery and diffusion-weighted acquisitions. Magn Reson Med 2016;75:372–380.
      1. Veraart J, Novikov DS, Fieremans E. TE dependent diffusion imaging (TEdDI) distinguishes between compartmental T2 relaxation times. Neuroimage 2018;182:360–369.
      1. Hutter J, Slator PJ, Christiaens D, et al. Integrated and efficient diffusion-relaxometry using ZEBRA. Sci Rep 2018;8:15138
      1. Fair MJ, Liao C, Manhard MK, Setsompop K. Diffusion-PEPTIDE: distortion- and blurring-free diffusion imaging with self-navigated motion-correction and relaxometry capabilities. Magn Reson Med 2021;85:2417–2433.
      1. Pas K, Komlosh ME, Perl DP, Basser PJ, Benjamini D. Retaining information from multidimensional correlation MRI using a spectral regions of interest generator. Sci Rep 2020;10:3246
      1. Slator PJ, Hutter J, Marinescu RV, et al. Data-driven multi-contrast spectral microstructure imaging with InSpect: INtegrated SPECTral component estimation and mapping. Med Image Anal 2021;71:102045
      1. Ma D, Gulani V, Seiberlich N, et al. Magnetic resonance fingerprinting. Nature 2013;495:187–192.
      1. Jiang Y, Ma D, Seiberlich N, Gulani V, Griswold MA. MR fingerprinting using fast imaging with steady state precession (FISP) with spiral readout. Magn Reson Med 2015;74:1621–1631.
      1. Kobayashi Y, Terada Y. Diffusion-weighting caused by spoiler gradients in the fast imaging with steady-state precession sequence may lead to inaccurate T2 measurements in MR fingerprinting. Magn Reson Med Sci 2019;18:96–104.
      1. Su P, Fan H, Liu P, et al. MR fingerprinting ASL: sequence characterization and comparison with dynamic susceptibility contrast (DSC) MRI. NMR Biomed 2020;33:e4202
      1. Lima da Cruz GJ, Velasco C, Lavin B, Jaubert O, Botnar RM, Prieto C. Myocardial T1, T2, T2*, and fat fraction quantification via low-rank motion-corrected cardiac MR fingerprinting. Magn Reson Med 2022;87:2757–2774.
      1. Cao X, Congyu L, Zhong Z, et al. 3D Diffusion-prepared MRF (3DM) with cardiac gating for rapid high resolution whole-brain T1, T2, proton density and diffusivity mapping; Proceedings of the Joint Annual Meeting ISMRM-ESMRMB & ISMRT 31st Annual Meeting; 2022 May 7-12; London, United Kingdom. ISMRM; 2022.
      1. Pirk CM, Gómez PA, Lipp I, et al. Deep learning-based parameter mapping for joint relaxation and diffusion tensor MR fingerprinting; Proceedings of the Third Conference on Medical Imaging with Deep Learning; 2020 Jul 6-8; Montreal, Canada. MIDL; 2020. pp. 638-654.
      1. Finkelstein A, Liao C, Cao X, Zhong J. Transfer learning promotes robust parametric mapping of diffusion encoded MR fingerprinting; Proceedings of the Medical Imaging with Deep Learning; 2022 Jul 6-8; Zürich, Switzerland. MIDL; 2022.
      1. Afzali M, Mueller L, Sakaie K, et al. MR fingerprinting with b-tensor encoding for simultaneous quantification of relaxation and diffusion in a single scan. Magn Reson Med 2022;88:2043–2057.
    Cited by
    Crossref 4
    Google Scholar 
    MeSH Terms
    Artifacts
    Dermatoglyphics
    Diagnosis
    Diffusion Magnetic Resonance Imaging
    Diffusion*
    Joints
    Magnetic Resonance Imaging*
    Methods*
    Water
    Metrics
    Share
    Figures
    slide
    slide

    1 / 2

    ORCID IDs
    Funding Information
    PERMALINK