Investigating the maturation of microstructure and radial orientation in the preterm human cortex with diffusion MRI.

Preterm birth disrupts and alters the complex developmental processes in the cerebral cortex. This disruption may be a contributing factor to widespread delay and cognitive difficulties in the preterm population. Diffusion-weighted magnetic resonance imaging (DW MRI) is a noninvasive imaging technique that makes inferences about cellular structures, at scales smaller than the imaging resolution. One established finding is that DW MRI shows a transient radial alignment in the preterm cortex. In this study, we quantify this maturational process with the "radiality index", a parameter that measures directional coherence, which we expect to change rapidly in the perinatal period. To measure this index, we used structural T2-weighted MRI to segment the cortex and generate cortical meshes. We obtained normal vectors for each face of the mesh and compared them to the principal diffusion direction, calculated by both the DTI and DIAMOND models, to generate the radiality index. The subjects included in this study were 89 infants born at fewer than 34 weeks completed gestation, each imaged at up to four timepoints between 27 and 42 weeks gestational age. In this manuscript, we quantify the longitudinal trajectory of radiality, fractional anisotropy and mean diffusivity from the DTI and DIAMOND models. For the radiality index and fractional anisotropy, the DIAMOND model offers improved sensitivity over the DTI model. The radiality index has a consistent progression across time, with the rate of change depending on the cortical lobe. The occipital lobe changes most rapidly, and the frontal and temporal least: this is commensurate with known developmental anatomy. Analysing the radiality index offers information complementary to other diffusion parameters.

Longitudinal development in the preterm thalamus and posterior white matter: MRI correlations between diffusion weighted imaging and T2 relaxometry.

Infants born prematurely are at increased risk of adverse neurodevelopmental outcome. The measurement of white matter tissue composition and structure can help predict functional performance. Specifically, measurements of myelination and indicators of myelination status in the preterm brain could be predictive of later neurological outcome. Quantitative imaging of myelin could thus serve to develop biomarkers for prognosis or therapeutic intervention; however, accurate estimation of myelin content is difficult. This work combines diffusion MRI and multi-component T2 relaxation measurements in a group of 37 infants born very preterm and scanned between 27 and 58 weeks equivalent gestational age. Seven infants have longitudinal data at two time points that we analyze in detail. Our aim is to show that measurement of the myelin water fraction is achievable using widely available pulse sequences and state-of-the-art algorithmic modeling of the MR imaging procedure and that a multi-component fitting routine to multi-shell diffusion weighted data can show differences in neurite density and local spatial arrangement in grey and white matter. Inference on the myelin water fraction allows us to demonstrate that the change in diffusion properties of the preterm thalamus is not solely due to myelination (that increase in myelin content accounts for about a third of the observed changes) whilst the decrease in the posterior white matter T2 has no significant component that is due to myelin water content. This work applies multi-modal advanced quantitative neuroimaging to investigate changing tissue properties in the longitudinal setting. Hum Brain Mapp 37:2479-2492, 2016. © 2016 Wiley Periodicals, Inc.

Longitudinal measurement of the developing grey matter in preterm subjects using multi-modal MRI.

Preterm birth is a major public health concern, with the severity and occurrence of adverse outcome increasing with earlier delivery. Being born preterm disrupts a time of rapid brain development: in addition to volumetric growth, the cortex folds, myelination is occurring and there are changes on the cellular level. These neurological events have been imaged non-invasively using diffusion-weighted (DW) MRI. In this population, there has been a focus on examining diffusion in the white matter, but the grey matter is also critically important for neurological health. We acquired multi-shell high-resolution diffusion data on 12 infants born at ≤ 28 weeks of gestational age at two time-points: once when stable after birth, and again at term-equivalent age. We used the Neurite Orientation Dispersion and Density Imaging model (NODDI) (Zhang et al., 2012) to analyse the changes in the cerebral cortex and the thalamus, both grey matter regions. We showed region-dependent changes in NODDI parameters over the preterm period, highlighting underlying changes specific to the microstructure. This work is the first time that NODDI parameters have been evaluated in both the cortical and the thalamic grey matter as a function of age in preterm infants, offering a unique insight into neuro-development in this at-risk population.

Learning skillful medium-range global weather forecasting.

Global medium-range weather forecasting is critical to decision-making across many social and economic domains. Traditional numerical weather prediction uses increased compute resources to improve forecast accuracy but does not directly use historical weather data to improve the underlying model. Here, we introduce GraphCast, a machine learning-based method trained directly from reanalysis data. It predicts hundreds of weather variables for the next 10 days at 0.25° resolution globally in under 1 minute. GraphCast significantly outperforms the most accurate operational deterministic systems on 90% of 1380 verification targets, and its forecasts support better severe event prediction, including tropical cyclone tracking, atmospheric rivers, and extreme temperatures. GraphCast is a key advance in accurate and efficient weather forecasting and helps realize the promise of machine learning for modeling complex dynamical systems.

As Easy as 1, 2 4? Uncertainty in Counting Tasks for Medical Imaging.

Counting is a fundamental task in biomedical imaging and count is an important biomarker in a number of conditions. Estimating the uncertainty in the measurement is thus vital to making definite, informed conclusions. In this paper, we first compare a range of existing methods to perform counting in medical imaging and suggest ways of deriving predictive intervals from these. We then propose and test a method for calculating intervals as an output of a multi-task network. These predictive intervals are optimised to be as narrow as possible, while also enclosing a desired percentage of the data. We demonstrate the effectiveness of this technique on histopathological cell counting and white matter hyperintensity counting. Finally, we offer insight into other areas where this technique may apply.

Multi-domain Adaptation in Brain MRI Through Paired Consistency and Adversarial Learning.

Supervised learning algorithms trained on medical images will often fail to generalize across changes in acquisition parameters. Recent work in domain adaptation addresses this challenge and successfully leverages labeled data in a source domain to perform well on an unlabeled target domain. Inspired by recent work in semi-supervised learning we introduce a novel method to adapt from one source domain to n target domains (as long as there is paired data covering all domains). Our multi-domain adaptation method utilises a consistency loss combined with adversarial learning. We provide results on white matter lesion hyperintensity segmentation from brain MRIs using the MICCAI 2017 challenge data as the source domain and two target domains. The proposed method significantly outperforms other domain adaptation baselines.

Differential Rates of Perinatal Maturation of Human Primary and Nonprimary Auditory Cortex.

Primary and nonprimary cerebral cortex mature along different timescales; however, the differences between the rates of maturation of primary and nonprimary cortex are unclear. Cortical maturation can be measured through changes in tissue microstructure detectable by diffusion magnetic resonance imaging (MRI). In this study, diffusion tensor imaging (DTI) was used to characterize the maturation of Heschl's gyrus (HG), which contains both primary auditory cortex (pAC) and nonprimary auditory cortex (nAC), in 90 preterm infants between 26 and 42 weeks postmenstrual age (PMA). The preterm infants were in different acoustical environments during their hospitalization: 46 in open ward beds and 44 in single rooms. A control group consisted of 15 term-born infants. Diffusion parameters revealed that (1) changes in cortical microstructure that accompany cortical maturation had largely already occurred in pAC by 28 weeks PMA, and (2) rapid changes were taking place in nAC between 26 and 42 weeks PMA. At term equivalent PMA, diffusion parameters for auditory cortex were different between preterm infants and term control infants, reflecting either delayed maturation or injury. No effect of room type was observed. For the preterm group, disturbed maturation of nonprimary (but not primary) auditory cortex was associated with poorer language performance at age two years.

NiftyNet: a deep-learning platform for medical imaging.

BACKGROUND AND OBJECTIVES: Medical image analysis and computer-assisted intervention problems are increasingly being addressed with deep-learning-based solutions. Established deep-learning platforms are flexible but do not provide specific functionality for medical image analysis and adapting them for this domain of application requires substantial implementation effort. Consequently, there has been substantial duplication of effort and incompatible infrastructure developed across many research groups. This work presents the open-source NiftyNet platform for deep learning in medical imaging. The ambition of NiftyNet is to accelerate and simplify the development of these solutions, and to provide a common mechanism for disseminating research outputs for the community to use, adapt and build upon. METHODS: The NiftyNet infrastructure provides a modular deep-learning pipeline for a range of medical imaging applications including segmentation, regression, image generation and representation learning applications. Components of the NiftyNet pipeline including data loading, data augmentation, network architectures, loss functions and evaluation metrics are tailored to, and take advantage of, the idiosyncracies of medical image analysis and computer-assisted intervention. NiftyNet is built on the TensorFlow framework and supports features such as TensorBoard visualization of 2D and 3D images and computational graphs by default. RESULTS: We present three illustrative medical image analysis applications built using NiftyNet infrastructure: (1) segmentation of multiple abdominal organs from computed tomography; (2) image regression to predict computed tomography attenuation maps from brain magnetic resonance images; and (3) generation of simulated ultrasound images for specified anatomical poses. CONCLUSIONS: The NiftyNet infrastructure enables researchers to rapidly develop and distribute deep learning solutions for segmentation, regression, image generation and representation learning applications, or extend the platform to new applications.

Longitudinal measurement of the developing thalamus in the preterm brain using multi-modal MRI.

Preterm birth is a significant public health concern. For infants born very preterm (≤ 32 weeks completed gestation), there is a high instance of developmental disability. Due to the heterogeneity of patient outcomes, it is important to investigate early markers of future ability to provide effective and targeted intervention. As a neuronal relay centre, the thalamus is critical for effective cognitive function and, thus, development of white matter connections between the thalamus and cortex is vital. By non-invasively examining the state of the thalamus we can monitor development in the preterm period. To track the development we develop a novel registration technique to combine data from multiple modalities, in order to derive the transformation from a preterm scan, to a scan of the same infant at term-equivalent age. By measuring the changes in diffusion parameters over this period on a per-voxel basis, we hope to provide unique insight into neurodevelopment.

Multi-modal measurement of the myelin-to-axon diameter g-ratio in preterm-born neonates and adult controls.

Infants born prematurely are at increased risk of adverse functional outcome. The measurement of white matter tissue composition and structure can help predict functional performance and this motivates the search for new multi-modal imaging biomarkers. In this work we develop a novel combined biomarker from diffusion MRI and multi-component T2 relaxation measurements in a group of infants born very preterm and scanned between 30 and 40 weeks equivalent gestational age. We also investigate this biomarker on a group of seven adult controls, using a multi-modal joint model-fitting strategy. The proposed emergent biomarker is tentatively related to axonal energetic efficiency (in terms of axonal membrane charge storage) and conduction velocity and is thus linked to the tissue electrical properties, giving it a good theoretical justification as a predictive measurement of functional outcome.

Measurement of myelin in the preterm brain: multi-compartment diffusion imaging and multi-component T2 relaxometry.

Measurements of myelination and indicators of myelination status in the preterm brain could be predictive of later neurological outcome. Quantitative imaging of myelin could thus serve to develop predictive biomarkers; however, accurate estimation of myelin content is difficult. In this work we show that measurement of the myelin water fraction (MWF) is achievable using widely available pulse sequences and state-of-the-art algorithmic modelling of the MR imaging. We show results of myelin water fraction measurement at both 30 (4 infants) and 40 (2 infants) weeks equivalent gestational age (EGA) and show that the spatial pattern of myelin is different between these ages. Furthermore we apply a multi-component fitting routine to multi-shell diffusion weighted data to show differences in neurite density and local spatial arrangement in grey and white matter. Finally we combine these results to investigate the relationships between the diffusion and myelin measurements to show that MWF in the preterm brain may be measured alongside multi-component diffusion characteristics using clinically feasible MR sequences.