Reproducibility of the Standard Model of diffusion in white matter on clinical MRI systems

Application of preoperative advanced diffusion magnetic resonance imaging in evaluating the postoperative recurrence of lower grade gliomas

BACKGROUND

Recurrence of lower grade glioma (LrGG) appeared to be unavoidable despite considerable research performed in last decades. Thus, we evaluated the postoperative recurrence within two years after the surgery in patients with LrGG by preoperative advanced diffusion magnetic resonance imaging (dMRI).

MATERIALS AND METHODS

48 patients with lower-grade gliomas (23 recurrence, 25 nonrecurrence) were recruited into this study. Different models of dMRI were reconstructed, including apparent fiber density (AFD), white matter tract integrity (WMTI), diffusion tensor imaging (DTI), diffusion kurtosis imaging (DKI), neurite orientation dispersion and density imaging (NODDI), Bingham NODDI and standard model imaging (SMI). Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA) was used to construct a multiparametric prediction model for the diagnosis of postoperative recurrence.

RESULTS

The parameters derived from each dMRI model, including AFD, axon water fraction (AWF), mean diffusivity (MD), mean kurtosis (MK), fractional anisotropy (FA), intracellular volume fraction (ICVF), extra-axonal perpendicular diffusivity (De⊥), extra-axonal parallel diffusivity (De∥) and free water fraction (fw), showed significant differences between nonrecurrence group and recurrence group. The extra-axonal perpendicular diffusivity (De⊥) had the highest area under curve (AUC = 0.885), which was significantly higher than others. The variable importance for the projection (VIP) value of De⊥ was also the highest. The AUC value of the multiparametric prediction model merging AFD, WMTI, DTI, DKI, NODDI, Bingham NODDI and SMI was up to 0.96.

CONCLUSION

Preoperative advanced dMRI showed great efficacy in evaluating postoperative recurrence of LrGG and De⊥ of SMI might be a valuable marker.

Improving the assessment of axonal injury in early multiple sclerosis

RATIONALE AND OBJECTIVES

Several quantitative magnetic resonance imaging (MRI) methods are available to measure tissue injury in multiple sclerosis (MS), but their pathological specificity remains limited. The multi-compartment diffusion imaging using the spherical mean technique (SMT) overcomes several technical limitations of the diffusion-weighted image signal, thus delivering metrics with increased pathological specificity. Given these premises, here we assess whether the SMT-derived apparent axonal volume (Vax) provides a better tissue classifier than the diffusion tensor imaging (DTI)-derived axial diffusivity (AD) in the white matter (WM) of MS brains.

METHODS

Forty-three treatment-naïve people with newly diagnosed MS, clinically isolated syndrome, or radiologically isolated syndrome and 18 healthy controls (HCs) underwent a 3.0 Tesla MRI inclusive of T1-weighted (T1-w) and T2-w fluid-attenuated inversion recovery (FLAIR) sequences, and multi-b shell diffusion-weighted imaging. In patients only, pre- and post-gadolinium diethylenetriamine penta-acetic acid T1-w sequences were obtained for the evaluation of contrast-active lesions (CELs). Vax and AD were calculated in T2-lesions, chronic black holes (cBHs), and normal appearing (NAWM) in patients and normal WM (NWM) in HCs. Vax and AD values were compared across all the possible combinations of these regions. CELs were excluded from the analyses.

RESULTS

Vax differed in all comparisons (p ≤ 0.047 by paired t-test); AD differed in most comparisons (p < 0.001) except between NAWM and NWM, and between cBHs and T2-lesions. Vax had higher accuracy (p ≤ 0.029 by DeLong test) and larger effect size (p ≤ 0.038 by paired t-test) than AD in differentiating areas with even minimal tissue injury.

CONCLUSIONS

Vax provides a better radiological quantitative discriminator of different degrees of axonal-mediated tissue injury even between areas with expected minimal pathology. Our data support further studies to assess the readiness of Vax as a measure of outcome for clinical trials on neuroprotection in MS.

White Matter Microstructural Changes Using Ultra-Strong Diffusion Gradient MRI in Adult-Onset Idiopathic Focal Cervical Dystonia

BACKGROUND AND OBJECTIVES

Adult-onset idiopathic focal cervical dystonia (AOIFCD) involves abnormal posturing of the cervical musculature and, in some individuals, an associated head tremor. Existing neuroimaging studies have implicated key motor networks. However, measures used to date lack specificity toward underlying pathophysiologic differences. We aim to assess white matter motor pathways for localized, microstructural differences, which may aid in understanding underlying mechanisms.

METHODS

Individuals diagnosed with AOIFCD and an age- and sex-matched control group were prospectively recruited through the Welsh Movement Disorders Research Network. All participants underwent in-depth clinical phenotyping and MRI (structural and diffusion sequences) using ultra-strong diffusion gradients. Tractography (whole-tract median values) and tractometry (along tract profiling) were performed for key white matter motor pathways assessing diffusion kurtosis imaging (DKI), neurite orientation dispersion and density imaging (NODDI), and standard model parameters. Groups were compared using linear model analysis with Bonferroni multiple comparison correction.

RESULTS

Fifty participants with AOIFCD and 30 healthy control participants were recruited, with 46 with AOIFCD and 30 healthy controls included for analysis (33 without head tremor, 13 with head tremor). Significant differences were observed in the anterior thalamic radiations (lower mid-tract fractional anisotropy [estimate = -0.046, p = 3.07 × 10-3], radial kurtosis [estimate = -0.165, p = 1.42 × 10-4], f-intra-axonal signal fraction [estimate = -0.044, p = 2.78 × 10-3], p2 orientation coherence [estimate = -0.043, p = 1.64 × 10-3], higher Orientation Dispersion Index [ODI, estimate = 0.023, p = 2.22 × 10-3]) and thalamopremotor tracts (higher mid-tract mean kurtosis [estimate = 0.064, p = 7.56 × 10-4], lower Neurite Density Index [estimate = 0.062, p = 2.1 × 10-3], higher distal tract ODI [estimate = 0.062, p = 3.1 × 10-3], lower f [estimate = -0.1, p = 2.3 × 10-3], and striatopremotor tracts [proximal lower f: estimate = -0.075, p = 1.06 × 10-3]). These measures correlated with clinical measures: dystonia duration (right thalamopremotor distal ODI: r = -0.9, p = 1.29 × 10-14), psychiatric symptoms (obsessive compulsive symptoms: left anterior thalamic radiation p2 r = 0.92, p = 2.797 × 10-11), sleep quality (Sleep Disorders Questionnaire Score: left anterior thalamic radiation ODI: r = -0.84, p = 4.84 × 10-11), pain (left anterior thalamic radiation ODI: r = -0.89, p = 1.4 × 10-13), and cognitive functioning (paired associated learning task p2, r = 0.94, p = 6.68 × 10-20).

DISCUSSION

Overall, localized microstructural differences were identified within tracts linking the prefrontal and premotor cortices with thalamic and basal ganglia regions, suggesting pathophysiologic processes involve microstructural aberrances of motor system modulatory pathways, particularly involving intra-axonal and fiber orientation dispersion measures.

Callosal Interhemispheric Communication in Mild Traumatic Brain Injury: A Mediation Analysis on WM Microstructure Effects

BACKGROUND AND PURPOSE

Because the corpus callosum connects the left and right hemispheres and a variety of WM bundles across the brain in complex ways, damage to the neighboring WM microstructure may specifically disrupt interhemispheric communication through the corpus callosum following mild traumatic brain injury. Here we use a mediation framework to investigate how callosal interhemispheric communication is affected by WM microstructure in mild traumatic brain injury.

MATERIALS AND METHODS

Multishell diffusion MR imaging was performed on 23 patients with mild traumatic brain injury within 1 month of injury and 17 healthy controls, deriving 11 diffusion metrics, including DTI, diffusional kurtosis imaging, and compartment-specific standard model parameters. Interhemispheric processing speed was assessed using the interhemispheric speed of processing task (IHSPT) by measuring the latency between word presentation to the 2 hemivisual fields and oral word articulation. Mediation analysis was performed to assess the indirect effect of neighboring WM microstructures on the relationship between the corpus callosum and IHSPT performance. In addition, we conducted a univariate correlation analysis to investigate the direct association between callosal microstructures and IHSPT performance as well as a multivariate regression analysis to jointly evaluate both callosal and neighboring WM microstructures in association with IHSPT scores for each group.

RESULTS

Several significant mediators in the relationships between callosal microstructure and IHSPT performance were found in healthy controls. However, patients with mild traumatic brain injury appeared to lose such normal associations when microstructural changes occurred compared with healthy controls.

CONCLUSIONS

This study investigates the effects of neighboring WM microstructure on callosal interhemispheric communication in healthy controls and patients with mild traumatic brain injury, highlighting that neighboring noncallosal WM microstructures are involved in callosal interhemispheric communication and information transfer. Further longitudinal studies may provide insight into the temporal dynamics of interhemispheric recovery following mild traumatic brain injury.

Cerebral microstructural alterations in Post-COVID-condition are related to cognitive impairment, olfactory dysfunction and fatigue

After contracting COVID-19, a substantial number of individuals develop a Post-COVID-Condition, marked by neurologic symptoms such as cognitive deficits, olfactory dysfunction, and fatigue. Despite this, biomarkers and pathophysiological understandings of this condition remain limited. Employing magnetic resonance imaging, we conduct a comparative analysis of cerebral microstructure among patients with Post-COVID-Condition, healthy controls, and individuals that contracted COVID-19 without long-term symptoms. We reveal widespread alterations in cerebral microstructure, attributed to a shift in volume from neuronal compartments to free fluid, associated with the severity of the initial infection. Correlating these alterations with cognition, olfaction, and fatigue unveils distinct affected networks, which are in close anatomical-functional relationship with the respective symptoms.

Optimisation of quantitative brain diffusion-relaxation MRI acquisition protocols with physics-informed machine learning

Diffusion-relaxation MRI aims to extract quantitative measures that characterise microstructural tissue properties such as orientation, size, and shape, but long acquisition times are typically required. This work proposes a physics-informed learning framework to extract an optimal subset of diffusion-relaxation MRI measurements for enabling shorter acquisition times, predict non-measured signals, and estimate quantitative parameters. In vivo and synthetic brain 5D-Diffusion-T1-T2∗-weighted MRI data obtained from five healthy subjects were used for training and validation, and from a sixth participant for testing. One fully data-driven and two physics-informed machine learning methods were implemented and compared to two manual selection procedures and Cramér-Rao lower bound optimisation. The physics-informed approaches could identify measurement-subsets that yielded more consistently accurate parameter estimates in simulations than other approaches, with similar signal prediction error. Five-fold shorter protocols yielded error distributions of estimated quantitative parameters with very small effect sizes compared to estimates from the full protocol. Selected subsets commonly included a denser sampling of the shortest and longest inversion time, lowest echo time, and high b-value. The proposed framework combining machine learning and MRI physics offers a promising approach to develop shorter imaging protocols without compromising the quality of parameter estimates and signal predictions.

Insights and improvements in correspondence between axonal volume fraction measured with diffusion-weighted MRI and electron microscopy

Biophysical diffusion-weighted imaging (DWI) models are increasingly used in neuroscience to estimate the axonal water fraction (f AW ), which in turn is key for noninvasive estimation of the axonal volume fraction (f A ). These models require thorough validation by comparison with a reference method, for example, electron microscopy (EM). While EM studies often neglect the unmyelinated axons and solely report the fraction of myelinated axons, in DWI both myelinated and unmyelinated axons contribute to the DWI signal. However, DWI models often include simplifications, for example, the neglect of differences in the compartmental relaxation times or fixed diffusivities, which in turn might affect the estimation off AW . We investigate whether linear calibration parameters (scaling and offset) can improve the comparability between EM- and DWI-based metrics off A . To this end, we (a) used six DWI models based on the so-called standard model of white matter (WM), including two models with fixed compartmental diffusivities (e.g., neurite orientation dispersion and density imaging, NODDI) and four models that fitted the compartmental diffusivities (e.g., white matter tract integrity, WMTI), and (b) used a multimodal data set including ex vivo diffusion DWI and EM data in mice with a broad dynamic range of fibre volume metrics. We demonstrated that the offset is associated with the volume fraction of unmyelinated axons and the scaling factor is associated with different compartmentalT 2 and can substantially enhance the comparability between EM- and DWI-based metrics off A . We found that DWI models that fitted compartmental diffusivities provided the most accurate estimates of the EM-basedf A . Finally, we introduced a more efficient hybrid calibration approach, where only the offset is estimated but the scaling is fixed to a theoretically predicted value. Using this approach, a similar one-to-one correspondence to EM was achieved for WMTI. The method presented can pave the way for use of validated DWI-based models in clinical research and neuroscience.

Assessment of Precision and Accuracy of Brain White Matter Microstructure using Combined Diffusion MRI and Relaxometry

Joint modeling of diffusion and relaxation has seen growing interest due to its potential to provide complementary information about tissue microstructure. For brain white matter, we designed an optimal diffusion-relaxometry MRI protocol that samples multiple b-values, B-tensor shapes, and echo times (TE). This variable-TE protocol (27 min) has as subsets a fixed-TE protocol (15 min) and a 2-shell dMRI protocol (7 min), both characterizing diffusion only. We assessed the sensitivity, specificity and reproducibility of these protocols with synthetic experiments and in six healthy volunteers. Compared with the fixed-TE protocol, the variable-TE protocol enables estimation of free water fractions while also capturing compartmental T 2 relaxation times. Jointly measuring diffusion and relaxation offers increased sensitivity and specificity to microstructure parameters in brain white matter with voxelwise coefficients of variation below 10%.

Changes in Cortical Microstructure of the Human Brain Resulting from Long-Term Motor Learning

The mechanisms subserving motor skill acquisition and learning in the intact human brain are not fully understood. Previous studies in animals have demonstrated a causal relationship between motor learning and structural rearrangements of synaptic connections, raising the question of whether neurite-specific changes are also observable in humans. Here, we use advanced diffusion magnetic resonance imaging (MRI), sensitive to dendritic and axonal processes, to investigate neuroplasticity in response to long-term motor learning. We recruited healthy male and female human participants (age range 19-29) who learned a challenging dynamic balancing task (DBT) over four consecutive weeks. Diffusion MRI signals were fitted using Neurite Orientation Dispersion and Density Imaging (NODDI), a theory-driven biophysical model of diffusion, yielding measures of tissue volume, neurite density and the organizational complexity of neurites. While NODDI indices were unchanged and reliable during the control period, neurite orientation dispersion increased significantly during the learning period mainly in primary sensorimotor, prefrontal, premotor, supplementary, and cingulate motor areas. Importantly, reorganization of cortical microstructure during the learning phase predicted concurrent behavioral changes, whereas there was no relationship between microstructural changes during the control phase and learning. Changes in neurite complexity were independent of alterations in tissue density, cortical thickness, and intracortical myelin. Our results are in line with the notion that structural modulation of neurites is a key mechanism supporting complex motor learning in humans.SIGNIFICANCE STATEMENT The structural correlates of motor learning in the human brain are not fully understood. Results from animal studies suggest that synaptic remodeling (e.g., reorganization of dendritic spines) in sensorimotor-related brain areas is a crucial mechanism for the formation of motor memory. Using state-of-the-art diffusion magnetic resonance imaging (MRI), we found a behaviorally relevant increase in the organizational complexity of neocortical microstructure, mainly in primary sensorimotor, prefrontal, premotor, supplementary, and cingulate motor regions, following training of a challenging dynamic balancing task (DBT). Follow-up analyses suggested structural modulation of synapses as a plausible mechanism driving this increase, while colocalized changes in cortical thickness, tissue density, and intracortical myelin could not be detected. These results advance our knowledge about the neurobiological basis of motor learning in humans.

A simulation-driven supervised learning framework to estimate brain microstructure using diffusion MRI

We propose a framework to train supervised learning models on synthetic data to estimate brain microstructure parameters using diffusion magnetic resonance imaging (dMRI). Although further validation is necessary, the proposed framework aims to seamlessly incorporate realistic simulations into dMRI microstructure estimation. Synthetic data were generated from over 1,000 neuron meshes converted from digital neuronal reconstructions and linked to their neuroanatomical parameters (such as soma volume and neurite length) using an optimized diffusion MRI simulator that produces intracellular dMRI signals from the solution of the Bloch-Torrey partial differential equation. By combining random subsets of simulated neuron signals with a free diffusion compartment signal, we constructed a synthetic dataset containing dMRI signals and 40 tissue microstructure parameters of 1.45 million artificial brain voxels. To implement supervised learning models we chose multilayer perceptrons (MLPs) and trained them on a subset of the synthetic dataset to estimate some microstructure parameters, namely, the volume fractions of soma, neurites, and the free diffusion compartment, as well as the area fractions of soma and neurites. The trained MLPs perform satisfactorily on the synthetic test sets and give promising in-vivo parameter maps on the MGH Connectome Diffusion Microstructure Dataset (CDMD). Most importantly, the estimated volume fractions showed low dependence on the diffusion time, the diffusion time independence of the estimated parameters being a desired property of quantitative microstructure imaging. The synthetic dataset we generated will be valuable for the validation of models that map between the dMRI signals and microstructure parameters. The surface meshes and microstructures parameters of the aforementioned neurons have been made publicly available.

Mapping tissue microstructure across the human brain on a clinical scanner with soma and neurite density image metrics

Soma and neurite density image (SANDI) is an advanced diffusion magnetic resonance imaging biophysical signal model devised to probe in vivo microstructural information in the gray matter (GM). This model requires acquisitions that include b values that are at least six times higher than those used in clinical practice. Such high b values are required to disentangle the signal contribution of water diffusing in soma from that diffusing in neurites and extracellular space, while keeping the diffusion time as short as possible to minimize potential bias due to water exchange. These requirements have limited the use of SANDI only to preclinical or cutting-edge human scanners. Here, we investigate the potential impact of neglecting water exchange in the SANDI model and present a 10-min acquisition protocol that enables to characterize both GM and white matter (WM) on 3 T scanners. We implemented analytical simulations to (i) evaluate the stability of the fitting of SANDI parameters when diminishing the number of shells; (ii) estimate the bias due to potential exchange between neurites and extracellular space in such reduced acquisition scheme, comparing it with the bias due to experimental noise. Then, we demonstrated the feasibility and assessed the repeatability and reproducibility of our approach by computing microstructural metrics of SANDI with AMICO toolbox and other state-of-the-art models on five healthy subjects. Finally, we applied our protocol to five multiple sclerosis patients. Results suggest that SANDI is a practical method to characterize WM and GM tissues in vivo on performant clinical scanners.

Neurite orientation dispersion and density imaging of white matter microstructure in sensory processing dysfunction with versus without comorbid ADHD

Introduction

Sensory Processing Dysfunction (SPD) is common yet understudied, affecting up to one in six children with 40% experiencing co-occurring challenges with attention. The neural architecture of SPD with Attention Deficit and Hyperactivity Disorder (ADHD) (SPD+ADHD) versus SPD without ADHD (SPD-ADHD) has yet to be explored in diffusion tensor imaging (DTI) and Neurite Orientation Dispersion and Density Imaging (NODDI) has yet to be examined.

Methods

The present study computed DTI and NODDI biophysical model parameter maps of one hundred children with SPD. Global, regional and voxel-level white matter tract measures were analyzed and compared between SPD+ADHD and SPD-ADHD groups.

Results

SPD+ADHD children had global WM Fractional Anisotropy (FA) and Neurite Density Index (NDI) that trended lower than SPD-ADHD children, primarily in boys only. Data-driven voxelwise and WM tract-based analysis revealed statistically significant decreases of NDI in boys with SPD+ADHD compared to those with SPD-ADHD, primarily in projection tracts of the internal capsule and commissural fibers of the splenium of the corpus callosum.

Conclusion

We conclude that WM microstructure is more delayed/disrupted in boys with SPD+ADHD compared to SPD-ADHD, with NODDI showing a larger effect than DTI. This may represent the combined WM pathology of SPD and ADHD, or it may result from a greater degree of SPD WM pathology causing the development of ADHD.

Differentiation of white matter histopathology using b-tensor encoding and machine learning

Diffusion-weighted magnetic resonance imaging (DW-MRI) is a non-invasive technique that is sensitive to microstructural geometry in neural tissue and is useful for the detection of neuropathology in research and clinical settings. Tensor-valued diffusion encoding schemes (b-tensor) have been developed to enrich the microstructural data that can be obtained through DW-MRI. These advanced methods have proven to be more specific to microstructural properties than conventional DW-MRI acquisitions. Additionally, machine learning methods are particularly useful for the study of multidimensional data sets. In this work, we have tested the reach of b-tensor encoding data analyses with machine learning in different histopathological scenarios. We achieved this in three steps: 1) We induced different levels of white matter damage in rodent optic nerves. 2) We obtained ex vivo DW-MRI data with b-tensor encoding schemes and calculated quantitative metrics using Q-space trajectory imaging. 3) We used a machine learning model to identify the main contributing features and built a voxel-wise probabilistic classification map of histological damage. Our results show that this model is sensitive to characteristics of microstructural damage. In conclusion, b-tensor encoded DW-MRI data analyzed with machine learning methods, have the potential to be further developed for the detection of histopathology and neurodegeneration.

Parameter estimation for WMTI-Watson model of white matter using encoder-decoder recurrent neural network

PURPOSE

Biophysical modeling of the diffusion MRI (dMRI) signal provides estimates of specific microstructural tissue properties. Although non-linear least squares (NLLS) is the most widespread fitting method, it suffers from local minima and high computational cost. Deep learning approaches are steadily replacing NLLS, but come with the limitation that the model needs to be retrained for each acquisition protocol and noise level. In this study, a novel fitting approach was proposed based on the encoder-decoder recurrent neural network (RNN) to accelerate model estimation with good generalization to various datasets.

METHODS

The white matter tract integrity (WMTI)-Watson model as an implementation of the Standard Model of diffusion in white matter derives its parameters indirectly from the diffusion and kurtosis tensors (DKI). The RNN-based solver, which estimates the WMTI-Watson model from DKI, is therefore more readily translatable to various data, irrespective of acquisition protocols as long as the DKI was pre-computed from the signal. An embedding approach was also used to render the model insensitive to potential differences in distributions between training data and experimental data. The analytical solution, NLLS, RNN-, and a multilayer perceptron (MLP)-based methods were evaluated on synthetic and in vivo datasets of rat and human brain.

RESULTS

The proposed RNN solver showed highly reduced computation time over the analytical solution and NLLS, with similar accuracy but improved robustness, and superior generalizability over MLP.

CONCLUSION

The RNN estimator can be easily applied to various datasets without retraining, which shows great potential for a widespread use.