Robust brain extraction across datasets and comparison with publicly available methods

Brain Structural Signature of RFC1-Related Disorder

BACKGROUND

The cerebellar ataxia, neuropathy, and vestibular areflexia syndrome was initially described in the early 1990s as a late-onset slowly progressive condition. Its underlying genetic cause was recently mapped to the RFC1 gene, and additional reports have expanded on the phenotypic manifestations related to RFC1, although little is known about the pattern and extent of structural brain abnormalities in this condition.

OBJECTIVE

The aim is to characterize the structural signature of brain damage in RFC1-related disorder, correlating the findings with clinical symptoms and normal brain RFC1 expression.

METHODS

We recruited 22 individuals with molecular confirmation of RFC1 expansions and submitted them to high-resolution 3T magnetic resonance imaging scans. We performed multimodal analyses to assess separately cerebral and cerebellar abnormalities within gray and white matter (WM). The results were compared with a group of 22 age- and sex-matched controls.

RESULTS

The mean age and disease duration of patients were 62.8 and 10.9 years, respectively. Ataxia, sensory neuronopathy, and vestibular areflexia were the most frequent manifestations, but parkinsonism and pyramidal signs were also noticed. We found that RFC1-related disorder is characterized by widespread and relatively symmetric cerebellar and basal ganglia atrophy. There is brainstem volumetric reduction along all its segments. Cerebral WM is also involved-mostly the corpus callosum and deep tracts, but cerebral cortical damage is rather restricted.

CONCLUSION

This study adds new relevant insights into the pathophysiological mechanisms of RFC1-related disorder. It should no longer be considered a purely cerebellar and sensory pathway disorder. Basal ganglia and deep cerebral WM are additional targets of damage. © 2021 International Parkinson and Movement Disorder Society.

PreQual: An automated pipeline for integrated preprocessing and quality assurance of diffusion weighted MRI images

PURPOSE

Diffusion weighted MRI imaging (DWI) is often subject to low signal-to-noise ratios (SNRs) and artifacts. Recent work has produced software tools that can correct individual problems, but these tools have not been combined with each other and with quality assurance (QA). A single integrated pipeline is proposed to perform DWI preprocessing with a spectrum of tools and produce an intuitive QA document.

METHODS

The proposed pipeline, built around the FSL, MRTrix3, and ANTs software packages, performs DWI denoising; inter-scan intensity normalization; susceptibility-, eddy current-, and motion-induced artifact correction; and slice-wise signal drop-out imputation. To perform QA on the raw and preprocessed data and each preprocessing operation, the pipeline documents qualitative visualizations, quantitative plots, gradient verifications, and tensor goodness-of-fit and fractional anisotropy analyses.

RESULTS

Raw DWI data were preprocessed and quality checked with the proposed pipeline and demonstrated improved SNRs; physiologic intensity ratios; corrected susceptibility-, eddy current-, and motion-induced artifacts; imputed signal-lost slices; and improved tensor fits. The pipeline identified incorrect gradient configurations and file-type conversion errors and was shown to be effective on externally available datasets.

CONCLUSIONS

The proposed pipeline is a single integrated pipeline that combines established diffusion preprocessing tools from major MRI-focused software packages with intuitive QA.

Generating Longitudinal Atrophy Evaluation Datasets on Brain Magnetic Resonance Images Using Convolutional Neural Networks and Segmentation Priors

Brain atrophy quantification plays a fundamental role in neuroinformatics since it permits studying brain development and neurological disorders. However, the lack of a ground truth prevents testing the accuracy of longitudinal atrophy quantification methods. We propose a deep learning framework to generate longitudinal datasets by deforming T1-w brain magnetic resonance imaging scans as requested through segmentation maps. Our proposal incorporates a cascaded multi-path U-Net optimised with a multi-objective loss which allows its paths to generate different brain regions accurately. We provided our model with baseline scans and real follow-up segmentation maps from two longitudinal datasets, ADNI and OASIS, and observed that our framework could produce synthetic follow-up scans that matched the real ones (Total scans= 584; Median absolute error: 0.03 ± 0.02; Structural similarity index: 0.98 ± 0.02; Dice similarity coefficient: 0.95 ± 0.02; Percentage of brain volume change: 0.24 ± 0.16; Jacobian integration: 1.13 ± 0.05). Compared to two relevant works generating brain lesions using U-Nets and conditional generative adversarial networks (CGAN), our proposal outperformed them significantly in most cases (p < 0.01), except in the delineation of brain edges where the CGAN took the lead (Jacobian integration: Ours - 1.13 ± 0.05 vs CGAN - 1.00 ± 0.02; p < 0.01). We examined whether changes induced with our framework were detected by FAST, SPM, SIENA, SIENAX, and the Jacobian integration method. We observed that induced and detected changes were highly correlated (Adj. R² > 0.86). Our preliminary results on harmonised datasets showed the potential of our framework to be applied to various data collections without further adjustment.

The Influence of Heart Rate Variability Biofeedback on Cardiac Regulation and Functional Brain Connectivity

Background

Heart rate variability (HRV) biofeedback has a beneficial impact on perceived stress and emotion regulation. However, its impact on brain function is still unclear. In this study, we aimed to investigate the effect of an 8-week HRV-biofeedback intervention on functional brain connectivity in healthy subjects.

Methods

HRV biofeedback was carried out in five sessions per week, including four at home and one in our lab. A control group played jump‘n’run games instead of the training. Functional magnetic resonance imaging was conducted before and after the intervention in both groups. To compute resting state functional connectivity (RSFC), we defined regions of interest in the ventral medial prefrontal cortex (VMPFC) and a total of 260 independent anatomical regions for network-based analysis. Changes of RSFC of the VMPFC to other brain regions were compared between groups. Temporal changes of HRV during the resting state recording were correlated to dynamic functional connectivity of the VMPFC.

Results

First, we corroborated the role of the VMPFC in cardiac autonomic regulation. We found that temporal changes of HRV were correlated to dynamic changes of prefrontal connectivity, especially to the middle cingulate cortex, the left insula, supplementary motor area, dorsal and ventral lateral prefrontal regions. The biofeedback group showed a drop in heart rate by 5.2 beats/min and an increased SDNN as a measure of HRV by 8.6 ms (18%) after the intervention. Functional connectivity of the VMPFC increased mainly to the insula, the amygdala, the middle cingulate cortex, and lateral prefrontal regions after biofeedback intervention when compared to changes in the control group. Network-based statistic showed that biofeedback had an influence on a broad functional network of brain regions.

Conclusion

Our results show that increased heart rate variability induced by HRV-biofeedback is accompanied by changes in functional brain connectivity during resting state.

Convolutional Neural Net Learning Can Achieve Production-Level Brain Segmentation in Structural Magnetic Resonance Imaging

Deep learning implementations using convolutional neural nets have recently demonstrated promise in many areas of medical imaging. In this article we lay out the methods by which we have achieved consistently high quality, high throughput computation of intra-cranial segmentation from whole head magnetic resonance images, an essential but typically time-consuming bottleneck for brain image analysis. We refer to this output as “production-level” because it is suitable for routine use in processing pipelines. Training and testing with an extremely large archive of structural images, our segmentation algorithm performs uniformly well over a wide variety of separate national imaging cohorts, giving Dice metric scores exceeding those of other recent deep learning brain extractions. We describe the components involved to achieve this performance, including size, variety and quality of ground truth, and appropriate neural net architecture. We demonstrate the crucial role of appropriately large and varied datasets, suggesting a less prominent role for algorithm development beyond a threshold of capability.

A self-supervised learning strategy for postoperative brain cavity segmentation simulating resections

PURPOSE

Accurate segmentation of brain resection cavities (RCs) aids in postoperative analysis and determining follow-up treatment. Convolutional neural networks (CNNs) are the state-of-the-art image segmentation technique, but require large annotated datasets for training. Annotation of 3D medical images is time-consuming, requires highly trained raters and may suffer from high inter-rater variability. Self-supervised learning strategies can leverage unlabeled data for training.

METHODS

We developed an algorithm to simulate resections from preoperative magnetic resonance images (MRIs). We performed self-supervised training of a 3D CNN for RC segmentation using our simulation method. We curated EPISURG, a dataset comprising 430 postoperative and 268 preoperative MRIs from 430 refractory epilepsy patients who underwent resective neurosurgery. We fine-tuned our model on three small annotated datasets from different institutions and on the annotated images in EPISURG, comprising 20, 33, 19 and 133 subjects.

RESULTS

The model trained on data with simulated resections obtained median (interquartile range) Dice score coefficients (DSCs) of 81.7 (16.4), 82.4 (36.4), 74.9 (24.2) and 80.5 (18.7) for each of the four datasets. After fine-tuning, DSCs were 89.2 (13.3), 84.1 (19.8), 80.2 (20.1) and 85.2 (10.8). For comparison, inter-rater agreement between human annotators from our previous study was 84.0 (9.9).

CONCLUSION

We present a self-supervised learning strategy for 3D CNNs using simulated RCs to accurately segment real RCs on postoperative MRI. Our method generalizes well to data from different institutions, pathologies and modalities. Source code, segmentation models and the EPISURG dataset are available at https://github.com/fepegar/resseg-ijcars .

Segmentation-Renormalized Deep Feature Modulation for Unpaired Image Harmonization

Deep networks are now ubiquitous in large-scale multi-center imaging studies. However, the direct aggregation of images across sites is contraindicated for downstream statistical and deep learning-based image analysis due to inconsistent contrast, resolution, and noise. To this end, in the absence of paired data, variations of Cycle-consistent Generative Adversarial Networks have been used to harmonize image sets between a source and target domain. Importantly, these methods are prone to instability, contrast inversion, intractable manipulation of pathology, and steganographic mappings which limit their reliable adoption in real-world medical imaging. In this work, based on an underlying assumption that morphological shape is consistent across imaging sites, we propose a segmentation-renormalized image translation framework to reduce inter-scanner heterogeneity while preserving anatomical layout. We replace the affine transformations used in the normalization layers within generative networks with trainable scale and shift parameters conditioned on jointly learned anatomical segmentation embeddings to modulate features at every level of translation. We evaluate our methodologies against recent baselines across several imaging modalities (T1w MRI, FLAIR MRI, and OCT) on datasets with and without lesions. Segmentation-renormalization for translation GANs yields superior image harmonization as quantified by Inception distances, demonstrates improved downstream utility via post-hoc segmentation accuracy, and improved robustness to translation perturbation and self-adversarial attacks.

Mapping fetal brain development based on automated segmentation and 4D brain atlasing

Fetal brain MRI has become an important tool for in utero assessment of brain development and disorders. However, quantitative analysis of fetal brain MRI remains difficult, partially due to the limited tools for automated preprocessing and the lack of normative brain templates. In this paper, we proposed an automated pipeline for fetal brain extraction, super-resolution reconstruction, and fetal brain atlasing to quantitatively map in utero fetal brain development during mid-to-late gestation in a Chinese population. First, we designed a U-net convolutional neural network for automated fetal brain extraction, which achieved an average accuracy of 97%. We then generated a developing fetal brain atlas, using an iterative linear and nonlinear registration approach. Based on the 4D spatiotemporal atlas, we quantified the morphological development of the fetal brain between 23 and 36 weeks of gestation. The proposed pipeline enabled the fully automated volumetric reconstruction for clinically available fetal brain MRI data, and the 4D fetal brain atlas provided normative templates for the quantitative characterization of fetal brain development, especially in the Chinese population.

Simultaneous imputation and classification using Multigraph Geometric Matrix Completion (MGMC): Application to neurodegenerative disease classification

Large-scale population-based studies in medicine are a key resource towards better diagnosis, monitoring, and treatment of diseases. They also serve as enablers of clinical decision support systems, in particular computer-aided diagnosis (CADx) using machine learning (ML). Numerous ML approaches for CADx have been proposed in literature. However, these approaches assume feature-complete data, which is often not the case in clinical data. To account for missing data, incomplete data samples are either removed or imputed, which could lead to data bias and may negatively affect classification performance. As a solution, we propose an end-to-end learning of imputation and disease prediction of incomplete medical datasets via Multi-graph Geometric Matrix Completion (MGMC). MGMC uses multiple recurrent graph convolutional networks, where each graph represents an independent population model based on a key clinical meta-feature like age, sex, or cognitive function. Graph signal aggregation from local patient neighborhoods, combined with multi-graph signal fusion via self-attention, has a regularizing effect on both matrix reconstruction and classification performance. Our proposed approach is able to impute class relevant features as well as perform accurate and robust classification on two publicly available medical datasets. We empirically show the superiority of our proposed approach in terms of classification and imputation performance when compared with state-of-the-art approaches. MGMC enables disease prediction in multimodal and incomplete medical datasets. These findings could serve as baseline for future CADx approaches which utilize incomplete datasets.

Preliminary Evidence for a Relationship between Elevated Plasma TNFα and Smaller Subcortical White Matter Volume in HCV Infection Irrespective of HIV or AUD Comorbidity

Classical inflammation in response to bacterial, parasitic, or viral infections such as HIV includes local recruitment of neutrophils and macrophages and the production of proinflammatory cytokines and chemokines. Proposed biomarkers of organ integrity in Alcohol Use Disorders (AUD) include elevations in peripheral plasma levels of proinflammatory proteins. In testing this proposal, previous work included a group of human immunodeficiency virus (HIV)-infected individuals as positive controls and identified elevations in the soluble proteins TNFα and IP10; these cytokines were only elevated in AUD individuals seropositive for hepatitis C infection (HCV). The current observational, cross-sectional study evaluated whether higher levels of these proinflammatory cytokines would be associated with compromised brain integrity. Soluble protein levels were quantified in 86 healthy controls, 132 individuals with AUD, 54 individuals seropositive for HIV, and 49 individuals with AUD and HIV. Among the patient groups, HCV was present in 24 of the individuals with AUD, 13 individuals with HIV, and 20 of the individuals in the comorbid AUD and HIV group. Soluble protein levels were correlated to regional brain volumes as quantified with structural magnetic resonance imaging (MRI). In addition to higher levels of TNFα and IP10 in the 2 HIV groups and the HCV-seropositive AUD group, this study identified lower levels of IL1β in the 3 patient groups relative to the control group. Only TNFα, however, showed a relationship with brain integrity: in HCV or HIV infection, higher peripheral levels of TNFα correlated with smaller subcortical white matter volume. These preliminary results highlight the privileged status of TNFα on brain integrity in the context of infection.

White matter hyperintensities segmentation using the ensemble U-Net with multi-scale highlighting foregrounds

White matter hyperintensities (WMHs) are abnormal signals within the white matter region on the human brain MRI and have been associated with aging processes, cognitive decline, and dementia. In the current study, we proposed a U-Net with multi-scale highlighting foregrounds (HF) for WMHs segmentation. Our method, U-Net with HF, is designed to improve the detection of the WMH voxels with partial volume effects. We evaluated the segmentation performance of the proposed approach using the Challenge training dataset. Then we assessed the clinical utility of the WMH volumes that were automatically computed using our method and the Alzheimer’s Disease Neuroimaging Initiative database. We demonstrated that the U-Net with HF significantly improved the detection of the WMH voxels at the boundary of the WMHs or in small WMH clusters quantitatively and qualitatively. Up to date, the proposed method has achieved the best overall evaluation scores, the highest dice similarity index, and the best F1-score among 39 methods submitted on the WMH Segmentation Challenge that was initially hosted by MICCAI 2017 and is continuously accepting new challengers. The evaluation of the clinical utility showed that the WMH volume that was automatically computed using U-Net with HF was significantly associated with cognitive performance and improves the classification between cognitive normal and Alzheimer’s disease subjects and between patients with mild cognitive impairment and those with Alzheimer’s disease. The implementation of our proposed method is publicly available using Dockerhub (https://hub.docker.com/r/wmhchallenge/pgs).

Modeling Healthy Anatomy with Artificial Intelligence for Unsupervised Anomaly Detection in Brain MRI

PURPOSE

To develop an unsupervised deep learning model on MR images of normal brain anatomy to automatically detect deviations indicative of pathologic states on abnormal MR images.

MATERIALS AND METHODS

In this retrospective study, spatial autoencoders with skip-connections (which can learn to compress and reconstruct data) were leveraged to learn the normal variability of the brain from MR scans of healthy individuals. A total of 100 normal, in-house MR scans were used for training. Subsequently, as the model was unable to reconstruct anomalies well, this characteristic was exploited for detecting and delineating various diseases by computing the difference between the input data and their reconstruction. The unsupervised model was compared with a supervised U-Net- and threshold-based classifier trained on data from 50 patients with multiple sclerosis (in-house dataset) and 50 patients from The Cancer Imaging Archive. Both the unsupervised and supervised U-Net models were tested on five different datasets containing MR images of microangiopathy, glioblastoma, and multiple sclerosis. Precision-recall statistics and derivations thereof (mean area under the precision-recall curve, Dice score) were used to quantify lesion detection and segmentation performance.

RESULTS

The unsupervised approach outperformed the naive thresholding approach in lesion detection (mean F1 scores ranging from 17% to 62% vs 6.4% to 15% across the five different datasets) and performed similarly to the supervised U-Net (20%-64%) across a variety of pathologic conditions. This outperformance was mostly driven by improved precision compared with the thresholding approach (mean precisions, 15%-59% vs 3.4%-10%). The model was also developed to create an anomaly heatmap display.

CONCLUSION

The unsupervised deep learning model was able to automatically detect anomalies on brain MR images with high performance. Supplemental material is available for this article. Keywords: Brain/Brain Stem Computer Aided Diagnosis (CAD), Convolutional Neural Network (CNN), Experimental Investigations, Head/Neck, MR-Imaging, Quantification, Segmentation, Stacked Auto-Encoders, Technology Assessment, Tissue Characterization © RSNA, 2021.

Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation

Segmentation of brain images from Magnetic Resonance Images (MRI) is an indispensable step in clinical practice. Morphological changes of sub-cortical brain structures and quantification of brain lesions are considered biomarkers of neurological and neurodegenerative disorders and used for diagnosis, treatment planning, and monitoring disease progression. In recent years, deep learning methods showed an outstanding performance in medical image segmentation. However, these methods suffer from generalisability problem due to inter-centre and inter-scanner variabilities of the MRI images. The main objective of the study is to develop an automated deep learning segmentation approach that is accurate and robust to the variabilities in scanner and acquisition protocols. In this paper, we propose a transductive transfer learning approach for domain adaptation to reduce the domain-shift effect in brain MRI segmentation. The transductive scenario assumes that there are sets of images from two different domains: (1) source-images with manually annotated labels; and (2) target-images without expert annotations. Then, the network is jointly optimised integrating both source and target images into the transductive training process to segment the regions of interest and to minimise the domain-shift effect. We proposed to use a histogram loss in the feature level to carry out the latter optimisation problem. In order to demonstrate the benefit of the proposed approach, the method has been tested in two different brain MRI image segmentation problems using multi-centre and multi-scanner databases for: (1) sub-cortical brain structure segmentation; and (2) white matter hyperintensities segmentation. The experiments showed that the segmentation performance of a pre-trained model could be significantly improved by up to 10%. For the first segmentation problem it was possible to achieve a maximum improvement from 0.680 to 0.799 in average Dice Similarity Coefficient (DSC) metric and for the second problem the average DSC improved from 0.504 to 0.602. Moreover, the improvements after domain adaptation were on par or showed better performance compared to the commonly used traditional unsupervised segmentation methods (FIRST and LST), also achieving faster execution time. Taking this into account, this work presents one more step toward the practical implementation of deep learning algorithms into the clinical routine.

Modified Diffusion Tensor Image Processing Pipeline for Archived Studies of Patients With Leukoencephalopathy

BACKGROUND

In archived diffusion tensor imaging (DTI) studies, a reversed-phase encoding (PE) scan required to correct the distortion in single-shot echo-planar imaging (EPI) may not have been acquired. Furthermore, DTI tractography is adversely affected by incorrect white matter segmentation due to leukoencephalopathy (LE). All these issues need to be addressed.

PURPOSE

To propose and evaluate a modified DTI processing pipeline with DIstortion COrrection using pseudo T₂ -weighted images (DICOT) to overcome limitations in existing acquisition protocols.

STUDY TYPE

Retrospective feasibility.

SUBJECTS

DICOT was assessed in simulated data and 84 acute lymphoblastic leukemia (ALL) patients with reversed PE acquired. The pipeline was then tested in 522 scans from 261 ALL patients without a reversed PE acquired.

FIELD STRENGTH/SEQUENCE

A 3 T; diffusion-weighted EPI; 3D magnetization prepared rapid acquisition gradient echo (MPRAGE).

STATISTICAL TESTS

Repeated measures analysis of variance and Tukey post hoc tests were performed to compare fractional anisotropy (FA) values obtained by different methods.

ASSESSMENT

FA and corresponding absolute error maps were obtained using TOPUP, DICOT, INVERSION (Inverse contrast Normalization for VERy Simple registratION) and NO CORR (no correction). Each method was assessed by comparing to TOPUP. The pipeline in the ALL patients was evaluated based on the failure rate of the distortion correction using the global correlation values.

RESULTS

Using DICOT reduced the mean absolute errors by an average of 32% in FA in simulation datasets. In 84 patients, the error reductions were approximately 15% in FA with DICOT, while it was 5% with INVERSION. No significant differences between the TOPUP and DICOT were observed in FA with P = 0.090/0.894(AP/PA). Only 15 of 516 examinations requiring any additional manual intervention.

CONCLUSION

This modified pipeline produced better results than the INVERSION. Furthermore, robust performance was demonstrated in archived patient scans acquired without an inverse PE necessary for TOPUP correction.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

Processing pipeline for image reconstructed fNIRS analysis using both MRI templates and individual anatomy

Significance: Image reconstruction of fNIRS data is a useful technique for transforming channel-based fNIRS into a volumetric representation and managing spatial variance based on optode location. We present an innovative integrated pipeline for image reconstruction of fNIRS data using either MRI templates or individual anatomy. Aim: We demonstrate a pipeline with accompanying code to allow users to clean and prepare optode location information, prepare and standardize individual anatomical images, create the light model, run the 3D image reconstruction, and analyze data in group space. Approach: We synthesize a combination of new and existing software packages to create a complete pipeline, from raw data to analysis. Results: This pipeline has been tested using both templates and individual anatomy, and on data from different fNIRS data collection systems. We show high temporal correlations between channel-based and image-based fNIRS data. In addition, we demonstrate the reliability of this pipeline with a sample dataset that included 74 children as part of a longitudinal study taking place in Scotland. We demonstrate good correspondence between data in channel space and image reconstructed data. Conclusions: The pipeline presented here makes a unique contribution by integrating multiple tools to assemble a complete pipeline for image reconstruction in fNIRS. We highlight further issues that may be of interest to future software developers in the field.

Patient-specific prediction of SEEG electrode bending for stereotactic neurosurgical planning

Purpose 

Electrode bending observed after stereotactic interventions is typically not accounted for in either computer-assisted planning algorithms, where straight trajectories are assumed, or in quality assessment, where only metrics related to entry and target points are reported. Our aim is to provide a fully automated and validated pipeline for the prediction of stereo-electroencephalography (SEEG) electrode bending.

Methods 

We transform electrodes of 86 cases into a common space and compare features-based and image-based neural networks on their ability to regress local displacement (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathbf{lu} $$\end{document}lu) or electrode bending (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hat{\mathbf{eb }}$$\end{document}eb^). Electrodes were stratified into six groups based on brain structures at the entry and target point. Models, both with and without Monte Carlo (MC) dropout, were trained and validated using tenfold cross-validation.

Results 

mage-based models outperformed features-based models for all groups, and models that predicted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathbf{lu} $$\end{document}lu performed better than for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hat{\mathbf{eb }}$$\end{document}eb^. Image-based model prediction with MC dropout resulted in lower mean squared error (MSE) with improvements up to 12.9% (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathbf{lu} $$\end{document}lu) and 39.9% (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\hat{\mathbf{eb }}$$\end{document}eb^), compared to no dropout. Using an image of brain tissue types (cortex, white and deep grey matter) resulted in similar, and sometimes better performance, compared to using a T1-weighted MRI when predicting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathbf{lu} $$\end{document}lu. When inferring trajectories of image-based models (brain tissue types), 86.9% of trajectories had an MSE\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\le 1$$\end{document}≤1 mm.

Conclusion 

An image-based approach regressing local displacement with an image of brain tissue types resulted in more accurate electrode bending predictions compared to other approaches, inputs, and outputs. Future work will investigate the integration of electrode bending into planning and quality assessment algorithms.

Supplementary Information

The online version supplementary material available at 10.1007/s11548-021-02347-8.

Automated workflow for volumetric assessment of signal intensity ratio on T1-weighted MR images after multiple gadolinium administrations

Purpose: Repeated injections of linear gadolinium-based contrast agent (GBCA) have shown correlations with increased signal intensities (SI) on unenhanced T1-weighted (T1w) images. Assessment is usually performed manually on a single slice and the SI as an average of a freehand region-of-interest is reported. We aim to develop a fully automated software that segments and computes SI ratio of dentate nucleus (DN) to pons (DN/P) and globus pallidus (GP) to thalamus (GP/T) for the assessment of gadolinium presence in the brain after a serial GBCA administrations. Approach: All patients ( N = 113 ) underwent at least eight GBCA enhanced scans. The modal SI in the DN, GP, pons, and thalamus were measured volumetrically on unenhanced T1w images and corrected based on the reference protocol (measurement 1) and compared to the SI-uncorrected-modal-volume (measurement 2), SI-corrected-mean-volume (measurement 3), as well as SI-corrected-modal-single slice (measurement 4) approaches. Results: Automatic processing worked on all 2119 studies (1150 at 1.5 T and 969 at 3 T). DN/P were 1.085 ± 0.048 (1.5 T) and 0.979 ± 0.061 (3 T). GP/T were 1.084 ± 0.039 (1.5 T) and 1.069 ± 0.042 (3 T). Modal DN/P ratios from volumetric assessment at 1.5 T failed to show a statistical difference with or without SI corrections ( p = 0.71 ). All other t -tests demonstrated significant differences (measurement 2, 3, 4 compared to 1, p < 0.001 ). Conclusion: The fully automatic method is an effective powerful tool to streamline the analysis of SI ratios in the deep brain tissues. Divergent SI ratios using different approaches reinforces the need to standardize the measurement for the research in this field.

Performance ramifications of abnormal functional connectivity of ventral posterior lateral thalamus with cerebellum in abstinent individuals with Alcohol Use Disorder

The extant literature supports the involvement of the thalamus in the cognitive and motor impairment associated with chronic alcohol consumption, but clear structure/function relationships remain elusive. Alcohol effects on specific nuclei rather than the entire thalamus may provide the basis for differential cognitive and motor decline in Alcohol Use Disorder (AUD). This functional MRI (fMRI) study was conducted in 23 abstinent individuals with AUD and 27 healthy controls to test the hypothesis that functional connectivity between anterior thalamus and hippocampus would be compromised in those with an AUD diagnosis and related to mnemonic deficits. Functional connectivity between 7 thalamic structures [5 thalamic nuclei: anterior ventral (AV), mediodorsal (MD), pulvinar (Pul), ventral lateral posterior (VLP), and ventral posterior lateral (VPL); ventral thalamus; the entire thalamus] and 14 "functional regions" was evaluated. Relative to controls, the AUD group exhibited different VPL-based functional connectivity: an anticorrelation between VPL and a bilateral middle temporal lobe region observed in controls became a positive correlation in the AUD group; an anticorrelation between the VPL and the cerebellum was stronger in the AUD than control group. AUD-associated altered connectivity between anterior thalamus and hippocampus as a substrate of memory compromise was not supported; instead, connectivity differences from controls selective to VPL and cerebellum demonstrated a relationship with impaired balance. These preliminary findings support substructure-level evaluation in future studies focused on discerning the role of the thalamus in AUD-associated cognitive and motor deficits.

Virtual brain grafting: Enabling whole brain parcellation in the presence of large lesions

Brain atlases and templates are at the heart of neuroimaging analyses, for which they facilitate multimodal registration, enable group comparisons and provide anatomical reference. However, as atlas-based approaches rely on correspondence mapping between images they perform poorly in the presence of structural pathology. Whilst several strategies exist to overcome this problem, their performance is often dependent on the type, size and homogeneity of any lesions present. We therefore propose a new solution, referred to as Virtual Brain Grafting (VBG), which is a fully-automated, open-source workflow to reliably parcellate magnetic resonance imaging (MRI) datasets in the presence of a broad spectrum of focal brain pathologies, including large, bilateral, intra- and extra-axial, heterogeneous lesions with and without mass effect. The core of the VBG approach is the generation of a lesion-free T1-weighted image, which enables further image processing operations that would otherwise fail. Here we validated our solution based on Freesurfer recon-all parcellation in a group of 10 patients with heterogeneous gliomatous lesions, and a realistic synthetic cohort of glioma patients (n = 100) derived from healthy control data and patient data. We demonstrate that VBG outperforms a non-VBG approach assessed qualitatively by expert neuroradiologists and Mann-Whitney U tests to compare corresponding parcellations (real patients U(6,6) = 33, z = 2.738, P < .010, synthetic-patients U(48,48) = 2076, z = 7.336, P < .001). Results were also quantitatively evaluated by comparing mean dice scores from the synthetic-patients using one-way ANOVA (unilateral VBG = 0.894, bilateral VBG = 0.903, and non-VBG = 0.617, P < .001). Additionally, we used linear regression to show the influence of lesion volume, lesion overlap with, and distance from the Freesurfer volumes of interest, on labeling accuracy. VBG may benefit the neuroimaging community by enabling automated state-of-the-art MRI analyses in clinical populations using methods such as FreeSurfer, CAT12, SPM, Connectome Workbench, as well as structural and functional connectomics. To fully maximize its availability, VBG is provided as open software under a Mozilla 2.0 license (https://github.com/KUL-Radneuron/KUL_VBG).

Replicability, Repeatability, and Long-term Reproducibility of Cerebellar Morphometry

To identify robust and reproducible methods of cerebellar morphometry that can be used in future large-scale structural MRI studies, we investigated the replicability, repeatability, and long-term reproducibility of three fully automated software tools: FreeSurfer, CEREbellum Segmentation (CERES), and automatic cerebellum anatomical parcellation using U-Net with locally constrained optimization (ACAPULCO). Replicability was defined as computational replicability, determined by comparing two analyses of the same high-resolution MRI data set performed with identical analysis software and computer hardware. Repeatability was determined by comparing the analyses of two MRI scans of the same participant taken during two independent MRI sessions on the same day for the Kirby-21 study. Long-term reproducibility was assessed by analyzing two MRI scans of the same participant in the longitudinal OASIS-2 study. We determined percent difference, the image intraclass correlation coefficient, the coefficient of variation, and the intraclass correlation coefficient between two analyses. Our results show that CERES and ACAPULCO use stochastic algorithms that result in surprisingly high differences between identical analyses for ACAPULCO and small differences for CERES. Changes between two consecutive scans from the Kirby-21 study were less than ± 5% in most cases for FreeSurfer and CERES (i.e., demonstrating high repeatability). As expected, long-term reproducibility was lower than repeatability for all software tools. In summary, CERES is an accurate, as demonstrated before, and reproducible tool for fully automated segmentation and parcellation of the cerebellum. We conclude with recommendations for the assessment of replicability, repeatability, and long-term reproducibility in future studies on cerebellar structure.

Autoencoders for unsupervised anomaly segmentation in brain MR images: A comparative study

Deep unsupervised representation learning has recently led to new approaches in the field of Unsupervised Anomaly Detection (UAD) in brain MRI. The main principle behind these works is to learn a model of normal anatomy by learning to compress and recover healthy data. This allows to spot abnormal structures from erroneous recoveries of compressed, potentially anomalous samples. The concept is of great interest to the medical image analysis community as it i) relieves from the need of vast amounts of manually segmented training data-a necessity for and pitfall of current supervised Deep Learning-and ii) theoretically allows to detect arbitrary, even rare pathologies which supervised approaches might fail to find. To date, the experimental design of most works hinders a valid comparison, because i) they are evaluated against different datasets and different pathologies, ii) use different image resolutions and iii) different model architectures with varying complexity. The intent of this work is to establish comparability among recent methods by utilizing a single architecture, a single resolution and the same dataset(s). Besides providing a ranking of the methods, we also try to answer questions like i) how many healthy training subjects are needed to model normality and ii) if the reviewed approaches are also sensitive to domain shift. Further, we identify open challenges and provide suggestions for future community efforts and research directions.

Learning joint segmentation of tissues and brain lesions from task-specific hetero-modal domain-shifted datasets

Brain tissue segmentation from multimodal MRI is a key building block of many neuroimaging analysis pipelines. Established tissue segmentation approaches have, however, not been developed to cope with large anatomical changes resulting from pathology, such as white matter lesions or tumours, and often fail in these cases. In the meantime, with the advent of deep neural networks (DNNs), segmentation of brain lesions has matured significantly. However, few existing approaches allow for the joint segmentation of normal tissue and brain lesions. Developing a DNN for such a joint task is currently hampered by the fact that annotated datasets typically address only one specific task and rely on task-specific imaging protocols including a task-specific set of imaging modalities. In this work, we propose a novel approach to build a joint tissue and lesion segmentation model from aggregated task-specific hetero-modal domain-shifted and partially-annotated datasets. Starting from a variational formulation of the joint problem, we show how the expected risk can be decomposed and optimised empirically. We exploit an upper bound of the risk to deal with heterogeneous imaging modalities across datasets. To deal with potential domain shift, we integrated and tested three conventional techniques based on data augmentation, adversarial learning and pseudo-healthy generation. For each individual task, our joint approach reaches comparable performance to task-specific and fully-supervised models. The proposed framework is assessed on two different types of brain lesions: White matter lesions and gliomas. In the latter case, lacking a joint ground-truth for quantitative assessment purposes, we propose and use a novel clinically-relevant qualitative assessment methodology.

Multisite Comparison of MRI Defacing Software Across Multiple Cohorts

With improvements to both scan quality and facial recognition software, there is an increased risk of participants being identified by a 3D render of their structural neuroimaging scans, even when all other personal information has been removed. To prevent this, facial features should be removed before data are shared or openly released, but while there are several publicly available software algorithms to do this, there has been no comprehensive review of their accuracy within the general population. To address this, we tested multiple algorithms on 300 scans from three neuroscience research projects, funded in part by the Ontario Brain Institute, to cover a wide range of ages (3-85 years) and multiple patient cohorts. While skull stripping is more thorough at removing identifiable features, we focused mainly on defacing software, as skull stripping also removes potentially useful information, which may be required for future analyses. We tested six publicly available algorithms (afni_refacer, deepdefacer, mri_deface, mridefacer, pydeface, quickshear), with one skull stripper (FreeSurfer) included for comparison. Accuracy was measured through a pass/fail system with two criteria; one, that all facial features had been removed and two, that no brain tissue was removed in the process. A subset of defaced scans were also run through several preprocessing pipelines to ensure that none of the algorithms would alter the resulting outputs. We found that the success rates varied strongly between defacers, with afni_refacer (89%) and pydeface (83%) having the highest rates, overall. In both cases, the primary source of failure came from a single dataset that the defacer appeared to struggle with - the youngest cohort (3-20 years) for afni_refacer and the oldest (44-85 years) for pydeface, demonstrating that defacer performance not only depends on the data provided, but that this effect varies between algorithms. While there were some very minor differences between the preprocessing results for defaced and original scans, none of these were significant and were within the range of variation between using different NIfTI converters, or using raw DICOM files.

Multimodal structural neuroimaging markers of risk and recovery from posttrauma anhedonia: A prospective investigation

BACKGROUND

Anhedonic symptoms of posttraumatic stress disorder (PTSD) reflect deficits in reward processing that have significant functional consequences. Although recent evidence suggests that disrupted integrity of fronto-limbic circuitry is related to PTSD development, including anhedonic PTSD symptoms (posttrauma anhedonia [PTA]), little is known about potential structural biomarkers of long-term PTA as well as structural changes in fronto-limbic pathways associated with recovery from PTA over time.

METHODS

We investigated associations between white matter microstructure, gray matter volume, and PTA in 75 recently traumatized individuals, with a subset of participants (n = 35) completing follow-up assessment 12 months after trauma exposure. Deterministic tractography and voxel-based morphometry were used to assess changes in white and gray matter structure associated with changes in PTA.

RESULTS

Reduced fractional anisotropy (FA) of the uncinate fasciculus at around the time of trauma predicted greater PTA at 12-months posttrauma. Further, increased FA of the fornix over time was associated with lower PTA between 1 and 12-months posttrauma. Increased gray matter volume of the ventromedial prefrontal cortex and precuneus over time was also associated with reduced PTA.

CONCLUSIONS

The microstructure of the uncinate fasciculus, an amygdala-prefrontal white matter connection, may represent a biomarker of vulnerability for later PTA. Conversely, development and recovery from PTA appear to be facilitated by white and gray matter structural changes in a major hippocampal pathway, the fornix. The present findings shed new light on neuroanatomical substrates of recovery from PTA and characterize white matter biomarkers of risk for posttraumatic dysfunction.

Radiomics Repeatability Pitfalls in a Scan-Rescan MRI Study of Glioblastoma

PURPOSE

To determine the influence of preprocessing on the repeatability and redundancy of radiomics features extracted using a popular open-source radiomics software package in a scan-rescan glioblastoma MRI study.

MATERIALS AND METHODS

In this study, a secondary analysis of T2-weighted fluid-attenuated inversion recovery (FLAIR) and T1-weighted postcontrast images from 48 patients (mean age, 56 years [range, 22-77 years]) diagnosed with glioblastoma were included from two prospective studies (ClinicalTrials.gov NCT00662506 [2009-2011] and NCT00756106 [2008-2011]). All patients underwent two baseline scans 2-6 days apart using identical imaging protocols on 3-T MRI systems. No treatment occurred between scan and rescan, and tumors were essentially unchanged visually. Radiomic features were extracted by using PyRadiomics (https://pyradiomics.readthedocs.io/) under varying conditions, including normalization strategies and intensity quantization. Subsequently, intraclass correlation coefficients were determined between feature values of the scan and rescan.

RESULTS

Shape features showed a higher repeatability than intensity (adjusted P < .001) and texture features (adjusted P < .001) for both T2-weighted FLAIR and T1-weighted postcontrast images. Normalization improved the overlap between the region of interest intensity histograms of scan and rescan (adjusted P < .001 for both T2-weighted FLAIR and T1-weighted postcontrast images), except in scans where brain extraction fails. As such, normalization significantly improves the repeatability of intensity features from T2-weighted FLAIR scans (adjusted P = .003 [z score normalization] and adjusted P = .002 [histogram matching]). The use of a relative intensity binning strategy as opposed to default absolute intensity binning reduces correlation between gray-level co-occurrence matrix features after normalization.

CONCLUSION

Both normalization and intensity quantization have an effect on the level of repeatability and redundancy of features, emphasizing the importance of both accurate reporting of methodology in radiomics articles and understanding the limitations of choices made in pipeline design. Supplemental material is available for this article. © RSNA, 2020See also the commentary by Tiwari and Verma in this issue.

PREEMACS: Pipeline for preprocessing and extraction of the macaque brain surface

Accurate extraction of the cortical brain surface is critical for cortical thickness estimation and a key element to perform multimodal imaging analysis, where different metrics are integrated and compared in a common space. While brain surface extraction has become widespread practice in human studies, several challenges unique to neuroimaging of non-human primates (NHP) have hindered its adoption for the study of macaques. Although, some of these difficulties can be addressed at the acquisition stage, several common artifacts can be minimized through image preprocessing. Likewise, there are several image analysis pipelines for human MRIs, but very few automated methods for extraction of cortical surfaces have been reported for NHPs and none have been tested on data from diverse sources. We present PREEMACS, a pipeline that standardizes the preprocessing of structural MRI images (T1- and T2-weighted) and carries out an automatic surface extraction of the macaque brain. Building upon and extending pre-existing tools, the first module performs volume orientation, image cropping, intensity non-uniformity correction, and volume averaging, before skull-stripping through a convolutional neural network. The second module performs quality control using an adaptation of MRIqc method to extract objective quality metrics that are then used to determine the likelihood of accurate brain surface estimation. The third and final module estimates the white matter (wm) and pial surfaces from the T1-weighted volume (T1w) using an NHP customized version of FreeSurfer aided by the T2-weighted volumes (T2w). To evaluate the generalizability of PREEMACS, we tested the pipeline using 57 T1w/T2w NHP volumes acquired at 11 different sites from the PRIME-DE public dataset. Results showed an accurate and robust automatic brain surface extraction from images that passed the quality control segment of our pipeline. This work offers a robust, efficient and generalizable pipeline for the automatic standardization of MRI surface analysis on NHP.

DIKA-Nets: Domain-invariant knowledge-guided attention networks for brain skull stripping of early developing macaques

As non-human primates, macaques have a close phylogenetic relationship to human beings and have been proven to be a valuable and widely used animal model in human neuroscience research. Accurate skull stripping (aka. brain extraction) of brain magnetic resonance imaging (MRI) is a crucial prerequisite in neuroimaging analysis of macaques. Most of the current skull stripping methods can achieve satisfactory results for human brains, but when applied to macaque brains, especially during early brain development, the results are often unsatisfactory. In fact, the early dynamic, regionally-heterogeneous development of macaque brains, accompanied by poor and age-related contrast between different anatomical structures, poses significant challenges for accurate skull stripping. To overcome these challenges, we propose a fully-automated framework to effectively fuse the age-specific intensity information and domain-invariant prior knowledge as important guiding information for robust skull stripping of developing macaques from 0 to 36 months of age. Specifically, we generate Signed Distance Map (SDM) and Center of Gravity Distance Map (CGDM) based on the intermediate segmentation results as guidance. Instead of using local convolution, we fuse all information using the Dual Self-Attention Module (DSAM), which can capture global spatial and channel-dependent information of feature maps. To extensively evaluate the performance, we adopt two relatively-large challenging MRI datasets from rhesus macaques and cynomolgus macaques, respectively, with a total of 361 scans from two different scanners with different imaging protocols. We perform cross-validation by using one dataset for training and the other one for testing. Our method outperforms five popular brain extraction tools and three deep-learning-based methods on cross-source MRI datasets without any transfer learning.

Automatic Brain Extraction for Rodent MRI Images

Rodent models are increasingly important in translational neuroimaging research. In rodent neuroimaging, particularly magnetic resonance imaging (MRI) studies, brain extraction is a critical data preprocessing component. Current brain extraction methods for rodent MRI usually require manual adjustment of input parameters due to widely different image qualities and/or contrasts. Here we propose a novel method, termed SHape descriptor selected Extremal Regions after Morphologically filtering (SHERM), which only requires a brain template mask as the input and is capable of automatically and reliably extracting the brain tissue in both rat and mouse MRI images. The method identifies a set of brain mask candidates, extracted from MRI images morphologically opened and closed sequentially with multiple kernel sizes, that match the shape of the brain template. These brain mask candidates are then merged to generate the brain mask. This method, along with four other state-of-the-art rodent brain extraction methods, were benchmarked on four separate datasets including both rat and mouse MRI images. Without involving any parameter tuning, our method performed comparably to the other four methods on all datasets, and its performance was robust with stably high true positive rates and low false positive rates. Taken together, this study provides a reliable automatic brain extraction method that can contribute to the establishment of automatic pipelines for rodent neuroimaging data analysis.

Regional growth trajectories of cortical myelination in adolescents and young adults: longitudinal validation and functional correlates

Adolescence is a time of continued cognitive and emotional evolution occurring with continuing brain development involving synaptic pruning and cortical myelination. The hypothesis of this study is that heavy myelination occurs in cortical regions with relatively direct, predetermined circuitry supporting unimodal sensory or motor functions and shows a steep developmental slope during adolescence (12-21 years) until young adulthood (22-35 years) when further myelination decelerates. By contrast, light myelination occurs in regions with highly plastic circuitry supporting complex functions and follows a delayed developmental trajectory. In support of this hypothesis, cortical myelin content was estimated and harmonized across publicly available datasets provided by the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA) and the Human Connectome Project (HCP). The cross-sectional analysis of 226 no-to-low alcohol drinking NCANDA adolescents revealed relatively steeper age-dependent trajectories of myelin growth in unimodal primary motor cortex and flatter age-dependent trajectories in multimodal mid/posterior cingulate cortices. This pattern of continued myelination showed smaller gains when the same analyses were performed on 686 young adults of the HCP cohort free of neuropsychiatric diagnoses. Critically, a predicted correlation between a motor task and myelin content in motor or cingulate cortices was found in the NCANDA adolescents, supporting the functional relevance of this imaging neurometric. Furthermore, the regional trajectory slopes were confirmed by performing longitudinally consistent analysis of cortical myelin. In conclusion, coordination of myelin content and circuit complexity continues to develop throughout adolescence, contributes to performance maturation, and may represent active cortical development climaxing in young adulthood.

Differential reinforcement encoding along the hippocampal long axis helps resolve the explore-exploit dilemma

When making decisions, should one exploit known good options or explore potentially better alternatives? Exploration of spatially unstructured options depends on the neocortex, striatum, and amygdala. In natural environments, however, better options often cluster together, forming structured value distributions. The hippocampus binds reward information into allocentric cognitive maps to support navigation and foraging in such spaces. Here we report that human posterior hippocampus (PH) invigorates exploration while anterior hippocampus (AH) supports the transition to exploitation on a reinforcement learning task with a spatially structured reward function. These dynamics depend on differential reinforcement representations in the PH and AH. Whereas local reward prediction error signals are early and phasic in the PH tail, global value maximum signals are delayed and sustained in the AH body. AH compresses reinforcement information across episodes, updating the location and prominence of the value maximum and displaying goal cell-like ramping activity when navigating toward it.

Brain extraction on MRI scans in presence of diffuse glioma: Multi-institutional performance evaluation of deep learning methods and robust modality-agnostic training

Brain extraction, or skull-stripping, is an essential pre-processing step in neuro-imaging that has a direct impact on the quality of all subsequent processing and analyses steps. It is also a key requirement in multi-institutional collaborations to comply with privacy-preserving regulations. Existing automated methods, including Deep Learning (DL) based methods that have obtained state-of-the-art results in recent years, have primarily targeted brain extraction without considering pathologically-affected brains. Accordingly, they perform sub-optimally when applied on magnetic resonance imaging (MRI) brain scans with apparent pathologies such as brain tumors. Furthermore, existing methods focus on using only T1-weighted MRI scans, even though multi-parametric MRI (mpMRI) scans are routinely acquired for patients with suspected brain tumors. In this study, we present a comprehensive performance evaluation of recent deep learning architectures for brain extraction, training models on mpMRI scans of pathologically-affected brains, with a particular focus on seeking a practically-applicable, low computational footprint approach, generalizable across multiple institutions, further facilitating collaborations. We identified a large retrospective multi-institutional dataset of n=3340 mpMRI brain tumor scans, with manually-inspected and approved gold-standard segmentations, acquired during standard clinical practice under varying acquisition protocols, both from private institutional data and public (TCIA) collections. To facilitate optimal utilization of rich mpMRI data, we further introduce and evaluate a novel ''modality-agnostic training'' technique that can be applied using any available modality, without need for model retraining. Our results indicate that the modality-agnostic approach1 obtains accurate results, providing a generic and practical tool for brain extraction on scans with brain tumors.

A naturalistic neuroimaging database for understanding the brain using ecological stimuli

Neuroimaging has advanced our understanding of human psychology using reductionist stimuli that often do not resemble information the brain naturally encounters. It has improved our understanding of the network organization of the brain mostly through analyses of 'resting-state' data for which the functions of networks cannot be verifiably labelled. We make a 'Naturalistic Neuroimaging Database' (NNDb v1.0) publically available to allow for a more complete understanding of the brain under more ecological conditions during which networks can be labelled. Eighty-six participants underwent behavioural testing and watched one of 10 full-length movies while functional magnetic resonance imaging was acquired. Resulting timeseries data are shown to be of high quality, with good signal-to-noise ratio, few outliers and low movement. Data-driven functional analyses provide further evidence of data quality. They also demonstrate accurate timeseries/movie alignment and how movie annotations might be used to label networks. The NNDb can be used to answer questions previously unaddressed with standard neuroimaging approaches, progressing our knowledge of how the brain works in the real world.

Deep learning of brain magnetic resonance images: A brief review

Magnetic resonance imaging (MRI) is one of the most popular techniques in brain science and is important for understanding brain function and neuropsychiatric disorders. However, the processing and analysis of MRI is not a trivial task with lots of challenges. Recently, deep learning has shown superior performance over traditional machine learning approaches in image analysis. In this survey, we give a brief review of the recent popular deep learning approaches and their applications in brain MRI analysis. Furthermore, popular brain MRI databases and deep learning tools are also introduced. The strength and weaknesses of different approaches are addressed, and challenges as well as future directions are also discussed.

Infant FreeSurfer: An automated segmentation and surface extraction pipeline for T1-weighted neuroimaging data of infants 0-2 years

The development of automated tools for brain morphometric analysis in infants has lagged significantly behind analogous tools for adults. This gap reflects the greater challenges in this domain due to: 1) a smaller-scaled region of interest, 2) increased motion corruption, 3) regional changes in geometry due to heterochronous growth, and 4) regional variations in contrast properties corresponding to ongoing myelination and other maturation processes. Nevertheless, there is a great need for automated image-processing tools to quantify differences between infant groups and other individuals, because aberrant cortical morphologic measurements (including volume, thickness, surface area, and curvature) have been associated with neuropsychiatric, neurologic, and developmental disorders in children. In this paper we present an automated segmentation and surface extraction pipeline designed to accommodate clinical MRI studies of infant brains in a population 0-2 year-olds. The algorithm relies on a single channel of T1-weighted MR images to achieve automated segmentation of cortical and subcortical brain areas, producing volumes of subcortical structures and surface models of the cerebral cortex. We evaluated the algorithm both qualitatively and quantitatively using manually labeled datasets, relevant comparator software solutions cited in the literature, and expert evaluations. The computational tools and atlases described in this paper will be distributed to the research community as part of the FreeSurfer image analysis package.

Automatic cerebellum anatomical parcellation using U-Net with locally constrained optimization

The cerebellum plays a central role in sensory input, voluntary motor action, and many neuropsychological functions and is involved in many brain diseases and neurological disorders. Cerebellar parcellation from magnetic resonance images provides a way to study regional cerebellar atrophy and also provides an anatomical map for functional imaging. In a recent comparison, a multi-atlas approach proved to be superior to other parcellation methods including some based on convolutional neural networks (CNNs) which have a considerable speed advantage. In this work, we developed an alternative CNN design for cerebellar parcellation, yielding a method that achieves the leading performance to date. The proposed method was evaluated on multiple data sets to show its broad applicability, and a Singularity container has been made publicly available.

Deep learning identifies morphological determinants of sex differences in the pre-adolescent brain

The application of data-driven deep learning to identify sex differences in developing brain structures of pre-adolescents has heretofore not been accomplished. Here, the approach identifies sex differences by analyzing the minimally processed MRIs of the first 8144 participants (age 9 and 10 years) recruited by the Adolescent Brain Cognitive Development (ABCD) study. The identified pattern accounted for confounding factors (i.e., head size, age, puberty development, socioeconomic status) and comprised cerebellar (corpus medullare, lobules III, IV/V, and VI) and subcortical (pallidum, amygdala, hippocampus, parahippocampus, insula, putamen) structures. While these have been individually linked to expressing sex differences, a novel discovery was that their grouping accurately predicted the sex in individual pre-adolescents. Another novelty was relating differences specific to the cerebellum to pubertal development. Finally, we found that reducing the pattern to a single score not only accurately predicted sex but also correlated with cognitive behavior linked to working memory. The predictive power of this score and the constellation of identified brain structures provide evidence for sex differences in pre-adolescent neurodevelopment and may augment understanding of sex-specific vulnerability or resilience to psychiatric disorders and presage sex-linked learning disabilities.

Extracting diffusion tensor fractional anisotropy and mean diffusivity from 3-direction DWI scans using deep learning

PURPOSE

To develop and evaluate machine-learning methods that reconstruct fractional anisotropy (FA) values and mean diffusivities (MD) from 3-direction diffusion MRI (dMRI) acquisitions.

METHODS

Two machine-learning models were implemented to map undersampled dMRI signals with high-quality FA and MD maps that were reconstructed from fully sampled DTI scans. The first model was a previously described multilayer perceptron (MLP), which maps signals and FA/MD values from a single voxel. The second was a convolutional neural network U-Net model, which maps dMRI slices to full FA/MD maps. Each method was trained on dMRI brain scans (N = 46), and reconstruction accuracies were compared with conventional linear-least-squares (LLS) reconstructions.

RESULTS

In an independent testing cohort (N = 20), 3-direction U-Net reconstructions had significantly lower absolute FA error than both 3-direction MLP (U-Net_3-dir : 0.06 ± 0.01 vs. MLP_3-dir : 0.08 ± 0.01, P < 1 × 10^-5 ) and 6-direction LLS (LLS_6-dir : 0.09 ± 0.03, P = 1 × 10^-5 ). The MD errors were not significantly different among 3-direction MLP (0.06 ± 0.01 × 10^-3 mm² /s), 3-direction U-Net (0.06 ± 0.01 × 10^-3 mm² /s), and 6-direction LLS (0.07 ± 0.02 × 10^-3 mm² /s, P > .1).

CONCLUSION

The proposed U-Net model reconstructed FA from 3-direction dMRI scans with improved accuracy compared with both a previously described MLP approach and LLS fitting from 6-direction scans. The MD reconstruction accuracies did not differ significantly between reconstructions.

From a deep learning model back to the brain-Identifying regional predictors and their relation to aging

We present a Deep Learning framework for the prediction of chronological age from structural magnetic resonance imaging scans. Previous findings associate increased brain age with neurodegenerative diseases and higher mortality rates. However, the importance of brain age prediction goes beyond serving as biomarkers for neurological disorders. Specifically, utilizing convolutional neural network (CNN) analysis to identify brain regions contributing to the prediction can shed light on the complex multivariate process of brain aging. Previous work examined methods to attribute pixel/voxel-wise contributions to the prediction in a single image, resulting in "explanation maps" that were found noisy and unreliable. To address this problem, we developed an inference scheme for combining these maps across subjects, thus creating a population-based, rather than a subject-specific map. We applied this method to a CNN ensemble trained on predicting subjects' age from raw T1 brain images in a lifespan sample of 10,176 subjects. Evaluating the model on an untouched test set resulted in mean absolute error of 3.07 years and a correlation between chronological and predicted age of r = 0.98. Using the inference method, we revealed that cavities containing cerebrospinal fluid, previously found as general atrophy markers, had the highest contribution for age prediction. Comparing maps derived from different models within the ensemble allowed to assess differences and similarities in brain regions utilized by the model. We showed that this method substantially increased the replicability of explanation maps, converged with results from voxel-based morphometry age studies and highlighted brain regions whose volumetric variability correlated the most with the prediction error.

3D-MRI Brain Tumor Detection Model Using Modified Version of Level Set Segmentation Based on Dragonfly Algorithm

Accurate brain tumor segmentation from 3D Magnetic Resonance Imaging (3D-MRI) is an important method for obtaining information required for diagnosis and disease therapy planning. Variation in the brain tumor’s size, structure, and form is one of the main challenges in tumor segmentation, and selecting the initial contour plays a significant role in reducing the segmentation error and the number of iterations in the level set method. To overcome this issue, this paper suggests a two-step dragonfly algorithm (DA) clustering technique to extract initial contour points accurately. The brain is extracted from the head in the preprocessing step, then tumor edges are extracted using the two-step DA, and these extracted edges are used as an initial contour for the MRI sequence. Lastly, the tumor region is extracted from all volume slices using a level set segmentation method. The results of applying the proposed technique on 3D-MRI images from the multimodal brain tumor segmentation challenge (BRATS) 2017 dataset show that the proposed method for brain tumor segmentation is comparable to the state-of-the-art methods.

State-of-the-Art Traditional to the Machine- and Deep-Learning-Based Skull Stripping Techniques, Models, and Algorithms

Several neuroimaging processing applications consider skull stripping as a crucial pre-processing step. Due to complex anatomical brain structure and intensity variations in brain magnetic resonance imaging (MRI), an appropriate skull stripping is an important part. The process of skull stripping basically deals with the removal of the skull region for clinical analysis in brain segmentation tasks, and its accuracy and efficiency are quite crucial for diagnostic purposes. It requires more accurate and detailed methods for differentiating brain regions and the skull regions and is considered as a challenging task. This paper is focused on the transition of the conventional to the machine- and deep-learning-based automated skull stripping methods for brain MRI images. It is observed in this study that deep learning approaches have outperformed conventional and machine learning techniques in many ways, but they have their limitations. It also includes the comparative analysis of the current state-of-the-art skull stripping methods, a critical discussion of some challenges, model of quantifying parameters, and future work directions.

Integrated Automatic Detection, Classification and Imaging of High Frequency Oscillations With Stereoelectroencephalography

Objective

During presurgical evaluation for focal epilepsy patients, the evidence supporting the use of high frequency oscillations (HFOs) for delineating the epileptogenic zone (EZ) increased over the past decade. This study aims to develop and validate an integrated automatic detection, classification and imaging pipeline of HFOs with stereoelectroencephalography (SEEG) to narrow the gap between HFOs quantitative analysis and clinical application.

Methods

The proposed pipeline includes stages of channel inclusion, candidate HFOs detection and automatic labeling with four trained convolutional neural network (CNN) classifiers and HFOs sorting based on occurrence rate and imaging. We first evaluated the initial detector using an open simulated dataset. After that, we validated our full algorithm in a 20-patient cohort against three assumptions based on previous studies. Classified HFOs results were compared with seizure onset zone (SOZ) channels for their concordance. The receiver operating characteristic (ROC) curve and the corresponding area under the curve (AUC) were calculated representing the prediction ability of the labeled HFOs outputs for SOZ.

Results

The initial detector demonstrated satisfactory performance on the simulated dataset. The four CNN classifiers converged quickly during training, and the accuracies on the validation dataset were above 95%. The localization value of HFOs was significantly improved by HFOs classification. The AUC values of the 20 testing patients increased after HFO classification, indicating a satisfactory prediction value of the proposed algorithm for EZ identification.

Conclusion

Our detector can provide robust HFOs analysis results revealing EZ at the individual level, which may ultimately push forward the transitioning of HFOs analysis into a meaningful part of the presurgical evaluation and surgical planning.

A web-based brain metastases segmentation and labeling platform for stereotactic radiosurgery

PURPOSE

Stereotactic radiosurgery (SRS) has become a standard of care for patients' with brain metastases (BMs). However, the manual multiple BMs delineation can be time-consuming and could create an efficiency bottleneck in SRS workflow. There is a clinical need for automatic delineation and quantitative evaluation tools. In this study, building on our previous developed deep learning-based segmentation algorithms, we developed a web-based automated BMs segmentation and labeling platform to assist the SRS clinical workflow.

METHOD

This platform was developed based on the Django framework, including a web client and a back-end server. The web client enables interactions as database access, data import, and image viewing. The server performs the segmentation and labeling tasks including: skull stripping; deep learning-based BMs segmentation; and affine registration-based BMs labeling. Additionally, the client can display BMs contours with corresponding atlas labels, and allows further postprocessing tasks including: (a) adjusting window levels; (b) displaying/hiding specific contours; (c) removing false-positive contours; (d) exporting contours as DICOM RTStruct files; etc. RESULTS: We evaluated this platform on 10 clinical cases with BMs number varied from 12-81 per case. The overall operation took about 4-5 min per patient. The segmentation accuracy was evaluated between the manual contour and automatic segmentation with several metrics. The averaged center of mass shift was 1.55 ± 0.36 mm, the Hausdorff distance was 2.98 ± 0.63 mm, the mean of surface-to-surface distance (SSD) was 1.06 ± 0.31 mm, and the standard deviation of SSD was 0.80 ± 0.16 mm. In addition, the initial averaged false-positive over union (FPoU) and false-negative rate (FNR) were 0.43 ± 0.19 and 0.15 ± 0.10 respectively. After case-specific postprocessing, the averaged FPoU and FNR were 0.19 ± 0.10 and 0.15 ± 0.10 respectively.

CONCLUSION

The evaluated web-based BMs segmentation and labeling platform can substantially improve the clinical efficiency compared to manual contouring. This platform can be a useful tool for assisting SRS treatment planning and treatment follow-up.

Striatal bases of temporal discounting in early adolescents

Steeper rates of temporal discounting-the degree to which smaller-sooner (SS) rewards are preferred over larger-later (LL) ones-have been associated with impulsive and ill-advised behaviors in adolescence. Yet, the underlying neural systems remain poorly understood. Here we used a well-established temporal discounting paradigm and functional MRI (fMRI) to examine engagement of the striatum-including the caudate, putamen, and ventral striatum (VS)-in early adolescence (13-15 years; N = 27). Analyses provided evidence of enhanced activity in the caudate and VS during impulsive choice. Exploratory analyses revealed that trait impulsivity was associated with heightened putamen activity during impulsive choices. A more nuanced pattern was evident in the cortex, with the dorsolateral prefrontal cortex mirroring the putamen and posterior parietal cortex showing the reverse association. Taken together, these observations provide an important first glimpse at the distributed neural systems underlying economic choice and trait-like individual differences in impulsivity in the early years of adolescence, setting the stage for prospective-longitudinal and intervention research.

Improved motion correction for functional MRI using an omnibus regression model

Head motion during functional Magnetic Resonance Imaging acquisition can significantly contaminate the neural signal and introduce spurious, distance-dependent changes in signal correlations. This can heavily confound studies of development, aging, and disease. Previous approaches to suppress head motion artifacts have involved sequential regression of nuisance covariates, but this has been shown to reintroduce artifacts. We propose a new motion correction pipeline using an omnibus regression model that avoids this problem by simultaneously regressing out multiple artifacts using the best performing algorithms to estimate each artifact. We quantitatively evaluate its motion artifact suppression performance against sequential regression pipelines using a large heterogeneous dataset (n=151) which includes high-motion subjects and multiple disease phenotypes. The proposed concatenated regression pipeline significantly reduces the association between head motion and functional connectivity while significantly outperforming the traditional sequential regression pipelines in eliminating distance-dependent head motion artifacts.

Simultaneous Motion and Distortion Correction Using Dual-Echo Diffusion-Weighted MRI

BACKGROUND AND PURPOSE

Geometric distortions resulting from large pose changes reduce the accuracy of motion measurements and interfere with the ability to generate artifact-free information. Our goal is to develop an algorithm and pulse sequence to enable motion-compensated, geometric distortion compensated diffusion-weighted MRI, and to evaluate its efficacy in correcting for the field inhomogeneity and position changes, induced by large and frequent head motions.

METHODS

Dual echo planar imaging (EPI) with a blip-reversed phase encoding distortion correction technique was evaluated in five volunteers in two separate experiments and compared with static field map distortion correction. In the first experiment, dual-echo EPI images were acquired in two head positions designed to induce a large field inhomogeneity change. A field map and a distortion-free structural image were acquired at each position to assess the ability of dual-echo EPI to generate reliable field maps and enable geometric distortion correction in both positions. In the second experiment, volunteers were asked to move to multiple random positions during a diffusion scan. Images were reconstructed using the dual-echo correction and a slice-to-volume registration (SVR) registration algorithm. The accuracy of SVR motion estimates was compared to externally measured ground truth motion parameters.

RESULTS

Our results show that dual-echo EPI can produce slice-level field maps with comparable quality to field maps generated by the reference gold standard method. We also show that slice-level distortion correction improves the accuracy of SVR algorithms as slices acquired at different orientations have different levels of distortion, which can create errors in the registration process.

CONCLUSIONS

Dual-echo acquisitions with blip-reversed phase encoding can be used to generate slice-level distortion-free images, which is critical for motion-robust slice to volume registration. The distortion corrected images not only result in better motion estimates, but they also enable a more accurate final diffusion image reconstruction.

Brain Volume: An Important Determinant of Functional Outcome After Acute Ischemic Stroke

OBJECTIVE

To determine whether brain volume is associated with functional outcome after acute ischemic stroke (AIS).

PATIENTS AND METHODS

This study was conducted between July 1, 2014, and March 16, 2019. We analyzed cross-sectional data of the multisite, international hospital-based MRI-Genetics Interface Exploration study with clinical brain magnetic resonance imaging obtained on admission for index stroke and functional outcome assessment. Poststroke outcome was determined using the modified Rankin Scale score (0-6; 0 = asymptomatic; 6 = death) recorded between 60 and 190 days after stroke. Demographic characteristics and other clinical variables including acute stroke severity (measured as National Institutes of Health Stroke Scale score), vascular risk factors, and etiologic stroke subtypes (Causative Classification of Stroke system) were recorded during index admission.

RESULTS

Utilizing the data from 912 patients with AIS (mean ± SD age, 65.3±14.5 years; male, 532 [58.3%]; history of smoking, 519 [56.9%]; hypertension, 595 [65.2%]) in a generalized linear model, brain volume (per 155.1 cm³) was associated with age (β -0.3 [per 14.4 years]), male sex (β 1.0), and prior stroke (β -0.2). In the multivariable outcome model, brain volume was an independent predictor of modified Rankin Scale score (β -0.233), with reduced odds of worse long-term functional outcomes (odds ratio, 0.8; 95% CI, 0.7-0.9) in those with larger brain volumes.

CONCLUSION

Larger brain volume quantified on clinical magnetic resonance imaging of patients with AIS at the time of stroke purports a protective mechanism. The role of brain volume as a prognostic, protective biomarker has the potential to forge new areas of research and advance current knowledge of the mechanisms of poststroke recovery.