SynthStrip: skull-stripping for any brain image

MGA-Net: A novel mask-guided attention neural network for precision neonatal brain imaging

In this study, we introduce MGA-Net, a novel mask-guided attention neural network, which extends the U-net model for precision neonatal brain imaging. MGA-Net is designed to extract the brain from other structures and reconstruct high-quality brain images. The network employs a common encoder and two decoders: one for brain mask extraction and the other for brain region reconstruction. A key feature of MGA-Net is its high-level mask-guided attention module, which leverages features from the brain mask decoder to enhance image reconstruction. To enable the same encoder and decoder to process both MRI and ultrasound (US) images, MGA-Net integrates sinusoidal positional encoding. This encoding assigns distinct positional values to MRI and US images, allowing the model to effectively learn from both modalities. Consequently, features learned from a single modality can aid in learning a modality with less available data, such as US. We extensively validated the proposed MGA-Net on diverse and independent datasets from varied clinical settings and neonatal age groups. The metrics used for assessment included the DICE similarity coefficient, recall, and accuracy for image segmentation; structural similarity for image reconstruction; and root mean squared error for total brain volume estimation from 3D ultrasound images. Our results demonstrate that MGA-Net significantly outperforms traditional methods, offering superior performance in brain extraction and segmentation while achieving high precision in image reconstruction and volumetric analysis. Thus, MGA-Net represents a robust and effective preprocessing tool for MRI and 3D ultrasound images, marking a significant advance in neuroimaging that enhances both research and clinical diagnostics in the neonatal period and beyond.

Enhancing healthcare recommendation: transfer learning in deep convolutional neural networks for Alzheimer disease detection

Neurodegenerative disorders such as Alzheimer's Disease (AD) and Mild Cognitive Impairment (MCI) significantly impact brain function and cognition. Advanced neuroimaging techniques, particularly Magnetic Resonance Imaging (MRI), play a crucial role in diagnosing these conditions by detecting structural abnormalities. This study leverages the ADNI and OASIS datasets, renowned for their extensive MRI data, to develop effective models for detecting AD and MCI. The research conducted three sets of tests, comparing multiple groups: multi-class classification (AD vs. Cognitively Normal (CN) vs. MCI), binary classification (AD vs. CN, and MCI vs. CN), to evaluate the performance of models trained on ADNI and OASIS datasets. Key preprocessing techniques such as Gaussian filtering, contrast enhancement, and resizing were applied to both datasets. Additionally, skull stripping using U-Net was utilized to extract features by removing the skull. Several prominent deep learning architectures including DenseNet-201, EfficientNet-B0, ResNet-50, ResNet-101, and ResNet-152 were investigated to identify subtle patterns associated with AD and MCI. Transfer learning techniques were employed to enhance model performance, leveraging pre-trained datasets for improved Alzheimer's MCI detection. ResNet-101 exhibited superior performance compared to other models, achieving 98.21% accuracy on the ADNI dataset and 97.45% accuracy on the OASIS dataset in multi-class classification tasks encompassing AD, CN, and MCI. It also performed well in binary classification tasks distinguishing AD from CN. ResNet-152 excelled particularly in binary classification between MCI and CN on the OASIS dataset. These findings underscore the utility of deep learning models in accurately identifying and distinguishing neurodegenerative diseases, showcasing their potential for enhancing clinical diagnosis and treatment monitoring.

Brainchop: Providing an Edge Ecosystem for Deployment of Neuroimaging Artificial Intelligence Models

Deep learning has proven highly effective in various medical imaging scenarios, yet the lack of an efficient distribution platform hinders developers from sharing models with end-users. Here, we describe brainchop, a fully functional web application that allows users to apply deep learning models developed with Python to local neuroimaging data from within their browser. While training artificial intelligence models is computationally expensive, applying existing models to neuroimaging data can be very fast; brainchop harnesses the end user's graphics card such that brain extraction, tissue segmentation, and regional parcellation require only seconds and avoids privacy issues that impact cloud-based solutions. The integrated visualization allows users to validate the inferences, and includes tools to annotate and edit the resulting segmentations. Our pure JavaScript implementation includes optimized helper functions for conforming volumes and filtering connected components with minimal dependencies. Brainchop provides a simple mechanism for distributing models for additional image processing tasks, including registration and identification of abnormal tissue, including tumors, lesions and hyperintensities. We discuss considerations for other AI model developers to leverage this open-source resource.

APIS: a paired CT-MRI dataset for ischemic stroke segmentation - methods and challenges

Stroke, the second leading cause of mortality globally, predominantly results from ischemic conditions. Immediate attention and diagnosis, related to the characterization of brain lesions, play a crucial role in patient prognosis. Standard stroke protocols include an initial evaluation from a non-contrast CT to discriminate between hemorrhage and ischemia. However, non-contrast CTs lack sensitivity in detecting subtle ischemic changes in this phase. Alternatively, diffusion-weighted MRI studies provide enhanced capabilities, yet are constrained by limited availability and higher costs. Hence, we idealize new approaches that integrate ADC stroke lesion findings into CT, to enhance the analysis and accelerate stroke patient management. This study details a public challenge where scientists applied top computational strategies to delineate stroke lesions on CT scans, utilizing paired ADC information. Also, it constitutes the first effort to build a paired dataset with NCCT and ADC studies of acute ischemic stroke patients. Submitted algorithms were validated with respect to the references of two expert radiologists. The best achieved Dice score was 0.2 over a test study with 36 patient studies. Despite all the teams employing specialized deep learning tools, results reveal limitations of computational approaches to support the segmentation of small lesions with heterogeneous density.

deepbet: Fast brain extraction of T1-weighted MRI using Convolutional Neural Networks

BACKGROUND

Brain extraction in magnetic resonance imaging (MRI) data is an important segmentation step in many neuroimaging preprocessing pipelines. Image segmentation is one of the research fields in which deep learning had the biggest impact in recent years. Consequently, traditional brain extraction methods are now being replaced by deep learning-based methods.

METHOD

Here, we used a unique dataset compilation comprising 7837 T1-weighted (T1w) MR images from 191 different OpenNeuro datasets in combination with advanced deep learning methods to build a fast, high-precision brain extraction tool called deepbet.

RESULTS

deepbet sets a novel state-of-the-art performance during cross-dataset validation with a median Dice score (DSC) of 99.0 on unseen datasets, outperforming the current best performing deep learning (DSC=97.9) and classic (DSC=96.5) methods. While current methods are more sensitive to outliers, deepbet achieves a Dice score of >97.4 across all 7837 images from 191 different datasets. This robustness was additionally tested in 5 external datasets, which included challenging clinical MR images. During visual exploration of each method's output which resulted in the lowest Dice score, major errors could be found for all of the tested tools except deepbet. Finally, deepbet uses a compute efficient variant of the UNet architecture, which accelerates brain extraction by a factor of ≈10 compared to current methods, enabling the processing of one image in ≈2 s on low level hardware.

CONCLUSIONS

In conclusion, deepbet demonstrates superior performance and reliability in brain extraction across a wide range of T1w MR images of adults, outperforming existing top tools. Its high minimal Dice score and minimal objective errors, even in challenging conditions, validate deepbet as a highly dependable tool for accurate brain extraction. deepbet can be conveniently installed via "pip install deepbet" and is publicly accessible at https://github.com/wwu-mmll/deepbet.

Perfusion parameter map generation from TOF-MRA in stroke using generative adversarial networks

PURPOSE

To generate perfusion parameter maps from Time-of-flight magnetic resonance angiography (TOF-MRA) images using artificial intelligence to provide an alternative to traditional perfusion imaging techniques.

MATERIALS AND METHODS

This retrospective study included a total of 272 patients with cerebrovascular diseases; 200 with acute stroke (from 2010 to 2018), and 72 with steno-occlusive disease (from 2011 to 2014). For each patient the TOF MRA image and the corresponding Dynamic susceptibility contrast magnetic resonance imaging (DSC-MRI) were retrieved from the datasets. The authors propose an adapted generative adversarial network (GAN) architecture, 3D pix2pix GAN, that generates common perfusion maps (CBF, CBV, MTT, TTP, Tmax) from TOF-MRA images. The performance was evaluated by the structural similarity index measure (SSIM). For a subset of 20 patients from the acute stroke dataset, the Dice coefficient was calculated to measure the overlap between the generated and real hypoperfused lesions with a time-to-maximum (Tmax) > 6 s.

RESULTS

The GAN model exhibited high visual overlap and performance for all perfusion maps in both datasets: acute stroke (mean SSIM 0.88-0.92, mean PSNR 28.48-30.89, mean MAE 0.02-0.04 and mean NRMSE 0.14-0.37) and steno-occlusive disease patients (mean SSIM 0.83-0.98, mean PSNR 23.62-38.21, mean MAE 0.01-0.05 and mean NRMSE 0.03-0.15). For the overlap analysis for lesions with Tmax>6 s, the median Dice coefficient was 0.49.

CONCLUSION

Our AI model can successfully generate perfusion parameter maps from TOF-MRA images, paving the way for a non-invasive alternative for assessing cerebral hemodynamics in cerebrovascular disease patients. This method could impact the stratification of patients with cerebrovascular diseases. Our results warrant more extensive refinement and validation of the method.

Age-appropriate or delayed myelination? Scoring myelination in routine clinical MRI

BACKGROUND

Assessment of myelination is a core issue in paediatric neuroimaging and can be challenging, particularly in settings without dedicated paediatric neuroradiologists. Deep learning models have recently been shown to be able to estimate myelination age in children with normal MRI, but currently lack validation for patients with myelination delay and implementation including pre-processing suitable for local imaging is not trivial. Standardized myelination scores, which have been successfully used as biomarkers for myelination in hypomyelinating diseases, rely on visual, semiquantitative scoring of myelination on routine clinical MRI and may offer an easy-to-use alternative for assessment of myelination.

METHODS

Myelination was scored in 13 anatomic sites (items) on conventional T2w and T1w images in controls (n = 253, 0-2 years). Items for the score were selected based on inter-rater variability, practicability of scoring, and importance for correctly identifying validation scans.

RESULTS

The resulting myelination score consisting of 7 T2- and 5 T1-items delineated myelination from term-equivalent to advanced, incomplete myelination which 50 % and 99 % of controls had reached by 19.1 and 32.7 months, respectively. It correctly identified 20/20 new control MRIs and 40/43 with myelination delay, missing one patient with borderline myelination delay at 8.6 months and 2 patients with incomplete T2-myelination of subcortical temporopolar white matter at 28 and 34 months.

CONCLUSIONS

The proposed myelination score provides an easy to use, standardized, and versatile tool to delineate myelination normally occurring during the first 1.5 years of life.

Next move in movement disorders: neuroimaging protocols for hyperkinetic movement disorders

Introduction

The Next Move in Movement Disorders (NEMO) study is an initiative aimed at advancing our understanding and the classification of hyperkinetic movement disorders, including tremor, myoclonus, dystonia, and myoclonus-dystonia. The study has two main objectives: (a) to develop a computer-aided tool for precise and consistent classification of these movement disorder phenotypes, and (b) to deepen our understanding of brain pathophysiology through advanced neuroimaging techniques. This protocol review details the neuroimaging data acquisition and preprocessing procedures employed by the NEMO team to achieve these goals.

Methods and analysis

To meet the study's objectives, NEMO utilizes multiple imaging techniques, including T1-weighted structural MRI, resting-state fMRI, motor task fMRI, and 18F-FDG PET scans. We will outline our efforts over the past 4 years to enhance the quality of our collected data, and address challenges such as head movements during image acquisition, choosing acquisition parameters and constructing data preprocessing pipelines. This study is the first to employ these neuroimaging modalities in a standardized approach contributing to more uniformity in the analyses of future studies comparing these patient groups. The data collected will contribute to the development of a machine learning-based classification tool and improve our understanding of disorder-specific neurobiological factors.

Ethics and dissemination

Ethical approval has been obtained from the relevant local ethics committee. The NEMO study is designed to pioneer the application of machine learning of movement disorders. We expect to publish articles in multiple related fields of research and patients will be informed of important results via patient associations and press releases.

Twinned neuroimaging analysis contributes to improving the classification of young people with autism spectrum disorder

Autism spectrum disorder (ASD) is diagnosed using comprehensive behavioral information. Neuroimaging offers additional information but lacks clinical utility for diagnosis. This study investigates whether multi-forms of magnetic resonance imaging (MRI) contrast can be used individually and in combination to produce a categorical classification of young individuals with ASD. MRI data were accessed from the Autism Brain Imaging Data Exchange (ABIDE). Young participants (ages 2-30) were selected, and two group cohorts consisted of 702 participants: 351 ASD and 351 controls. Image-based classification was performed using one-channel and two-channel inputs to 3D-DenseNet deep learning networks. The models were trained and tested using tenfold cross-validation. Two-channel models were twinned with combinations of structural MRI (sMRI) maps and amplitude of low-frequency fluctuations (ALFF) or fractional ALFF (fALFF) maps from resting-state functional MRI (rs-fMRI). All models produced classification accuracy that exceeded 65.1%. The two-channel ALFF-sMRI model achieved the highest mean accuracy of 76.9% ± 2.34. The one-channel ALFF-based model alone had mean accuracy of 72% ± 3.1. This study leveraged the ABIDE dataset to produce ASD classification results that are comparable and/or exceed literature values. The deep learning approach was conducive to diverse neuroimaging inputs. Findings reveal that the ALFF-sMRI two-channel model outperformed all others.

Exploring White Matter Microstructure with Symptom Severity and Outcomes Following Deep Brain Stimulation in Tremor Syndromes

Background

Essential tremor (ET) and dystonic tremor (DT) are movement disorders that cause debilitating symptoms, significantly impacting daily activities and quality of life. A poor understanding of their pathophysiology, as well as the mediators of clinical outcomes following deep brain stimulation (DBS), highlights the need for biomarkers to accurately characterise and optimally treat patients.

Objectives

We assessed the white matter microstructure of pathways implicated in the pathophysiology and therapeutic intervention in a retrospective cohort of patients with DT (n = 17) and ET (n = 19). We aimed to identity associations between white matter microstructure, upper limb tremor severity, and tremor improvement following DBS.

Methods

A fixel-based analysis pipeline was implemented to investigate white matter microstructural metrics in the whole brain, cerebello-thalamic pathways and tracts connected to stimulation volumes following DBS. Associations with preoperative and postoperative severity were analysed within each disorder group and across combined disorder groups.

Results

DBS led to significant improvements in both groups. No group differences in stimulation positions were identified. When white matter microstructural data was aligned according to the maximally affected upper limb, increased fiber density, and combined fiber density & cross-section of fixels in the left cerebellum were associated with greater tremor severity across DT and ET patients. White matter microstructure did not show associations with postoperative changes in cerebello-thalamic pathways, or tracts connected to stimulation volumes.

Discussion

Diffusion changes of the cerebellum are associated with the severity of upper limb tremor and appear to overlap in essential or dystonic tremor disorders.

Deep learning for detecting prenatal alcohol exposure in pediatric brain MRI: a transfer learning approach with explainability insights

Prenatal alcohol exposure (PAE) refers to the exposure of the developing fetus due to alcohol consumption during pregnancy and can have life-long consequences for learning, behavior, and health. Understanding the impact of PAE on the developing brain manifests challenges due to its complex structural and functional attributes, which can be addressed by leveraging machine learning (ML) and deep learning (DL) approaches. While most ML and DL models have been tailored for adult-centric problems, this work focuses on applying DL to detect PAE in the pediatric population. This study integrates the pre-trained simple fully convolutional network (SFCN) as a transfer learning approach for extracting features and a newly trained classifier to distinguish between unexposed and PAE participants based on T1-weighted structural brain magnetic resonance (MR) scans of individuals aged 2-8 years. Among several varying dataset sizes and augmentation strategy during training, the classifier secured the highest sensitivity of 88.47% with 85.04% average accuracy on testing data when considering a balanced dataset with augmentation for both classes. Moreover, we also preliminarily performed explainability analysis using the Grad-CAM method, highlighting various brain regions such as corpus callosum, cerebellum, pons, and white matter as the most important features in the model's decision-making process. Despite the challenges of constructing DL models for pediatric populations due to the brain's rapid development, motion artifacts, and insufficient data, this work highlights the potential of transfer learning in situations where data is limited. Furthermore, this study underscores the importance of preserving a balanced dataset for fair classification and clarifying the rationale behind the model's prediction using explainability analysis.

Iron changes within infarct tissue in ischemic stroke patients after successful reperfusion quantified using QSM

PURPOSE

For nearly half of patients who undergo Endovascular Thrombectomy following ischemic stroke, successful recanalisation does not guarantee a good outcome. Understanding the underlying tissue changes in the infarct tissue with the help of biomarkers specific to ischemic stroke could offer valuable insights for better treatment and patient management decisions. Using quantitative susceptibility mapping (QSM) MRI to measure cerebral iron concentration, this study aims to track the progression of iron within the infarct lesion after successful reperfusion.

METHODS

In a prospective study of 87 ischemic stroke patients, successfully reperfused patients underwent MRI scans at 24-to-72 h and 3 months after reperfusion. QSM maps were generated from gradient-echo MRI images. QSM values, measured in parts per billion (ppb), were extracted from ROIs defining the infarct and mirror homolog in the contralateral hemisphere and were compared cross-sectionally and longitudinally.

RESULTS

QSM values in the infarct ROIs matched those of the contralateral ROIs at 24-to-72 h, expressed as median (interquartile range) ppb [0.71(-7.67-10.09) vs. 2.20(-10.50-14.05) ppb, p = 0.55], but were higher at 3 months [10.68(-2.30-21.10) vs. -1.27(-12.98-9.82) ppb, p < 0.001]. The infarct QSM values at 3 months were significantly higher than those at 24-to-72 h [10.41(-2.50-18.27) ppb vs. 1.68(-10.36-12.25) ppb, p < 0.001]. Infarct QSM at 24-to-72 h and patient outcome measured at three months did not demonstrate a significant association.

CONCLUSION

Following successful endovascular reperfusion, iron concentration in infarct tissue, as measured by QSM increases over time compared to that in healthy tissue. However, its significance warrants further investigation.

A Deep Learning-Based Framework for Predicting Intracerebral Hemorrhage Hematoma Expansion Using Head Non-contrast CT Scan

RATIONALE AND OBJECTIVES

Hematoma expansion (HE) in intracerebral hemorrhage (ICH) is a critical factor affecting patient outcomes, yet effective clinical tools for predicting HE are currently lacking. We aim to develop a fully automated framework based on deep learning for predicting HE using only clinical non-contrast CT (NCCT) scans.

MATERIALS AND METHODS

A large retrospective dataset (n = 2484) was collected from 84 centers, while a prospective dataset (n = 500) was obtained from 26 additional centers. Baseline NCCT scans and follow-up NCCT scans were conducted within 6 h and 48 h from symptom onset, respectively. HE was defined as a volume increase of more than 6 mL on the follow-up NCCT. The retrospective dataset was divided into a training set (n = 1876) and a validation set (n = 608) by patient inclusion time. A two-stage framework was trained to predict HE, and its performance was evaluated on both the validation and prospective sets. Receiver operating characteristics area under the curve (AUC), sensitivity, and specificity were leveraged.

RESULTS

Our two-stage framework achieved an AUC of 0.760 (95% CI 0.724-0.799) on the retrospective validation set and 0.806 (95% CI 0.750-0.859) on the prospective set, outperforming the commonly used BAT score, which had AUCs of 0.582 and 0.699, respectively.

CONCLUSION

Our framework can automatically and robustly identify ICH patients at high risk of HE using admission head NCCT scans, providing more accurate predictions than the BAT score.

A Set of FMRI Quality Control Tools in AFNI: Systematic, in-depth, and interactive QC with afni_proc.py and more

Quality control (QC) assessment is a vital part of FMRI processing and analysis, and a typically underdiscussed aspect of reproducibility. This includes checking datasets at their very earliest stages (acquisition and conversion) through their processing steps (e.g., alignment and motion correction) to regression modeling (correct stimuli, no collinearity, valid fits, enough degrees of freedom, etc.) for each subject. There are a wide variety of features to verify throughout any single-subject processing pipeline, both quantitatively and qualitatively. We present several FMRI preprocessing QC features available in the AFNI toolbox, many of which are automatically generated by the pipeline-creation tool, afni_proc.py. These items include a modular HTML document that covers full single-subject processing from the raw data through statistical modeling, several review scripts in the results directory of processed data, and command line tools for identifying subjects with one or more quantitative properties across a group (such as triaging warnings, making exclusion criteria, or creating informational tables). The HTML itself contains several buttons that efficiently facilitate interactive investigations into the data, when deeper checks are needed beyond the systematic images. The pages are linkable, so that users can evaluate individual items across a group, for increased sensitivity to differences (e.g., in alignment or regression modeling images). Finally, the QC document contains rating buttons for each "QC block," as well as comment fields for each, to facilitate both saving and sharing the evaluations. This increases the specificity of QC, as well as its shareability, as these files can be shared with others and potentially uploaded into repositories, promoting transparency and open science. We describe the features and applications of these QC tools for FMRI.

Cerebrospinal fluid-based spatial statistics: towards quantitative analysis of cerebrospinal fluid pseudodiffusivity

BACKGROUND

Cerebrospinal fluid (CSF) circulation is essential in removing metabolic wastes from the brain and is an integral component of the glymphatic system. Abnormal CSF circulation is implicated in neurodegenerative diseases. Low b-value magnetic resonance imaging quantifies the variance of CSF motion, or pseudodiffusivity. However, few studies have investigated the relationship between the spatial patterns of CSF pseudodiffusivity and cognition.

METHODS

We introduced a novel technique, CSF-based spatial statistics (CBSS), to automatically quantify CSF pseudodiffusivity in each sulcus, cistern and ventricle. Using cortical regions as landmarks, we segmented each CSF region. We retrospectively analyzed a cohort of 93 participants with varying degrees of cognitive impairment.

RESULTS

We identified two groups of CSF regions whose pseudodiffusivity profiles were correlated with each other: one group displaying higher pseudodiffusivity and near large arteries and the other group displaying lower pseudodiffusivity and away from the large arteries. The pseudodiffusivity in the third ventricle positively correlated with short-term memory (standardized slope of linear regression = 0.38, adjusted p < 0.001) and long-term memory (slope = 0.37, adjusted p = 0.005). Fine mapping along the ventricles revealed that the pseudodiffusivity in the region closest to the start of the third ventricle demonstrated the highest correlation with cognitive performance.

CONCLUSIONS

CBSS enabled quantitative spatial analysis of CSF pseudodiffusivity and suggested the third ventricle pseudodiffusivity as a potential biomarker of cognitive impairment.

High-resolution computed tomographic (HRCT) image series from 413 canid and 18 felid skulls

Computed tomography (CT) is a non-invasive, three-dimensional imaging tool used in medical imaging, forensic science, industry and engineering, anthropology, and archaeology. The current study used high-resolution medical CT scanning of 431 animal skulls, including 399 dog skulls from 152 breeds, 14 cat skulls from 9 breeds, 14 skulls from 8 wild canid species (gray wolf, golden jackal, coyote, maned wolf, bush dog, red fox, Fennec fox, bat-eared fox), and 4 skulls from 4 wild felid species (wildcat, leopard, serval, caracal). This comprehensive and unique collection of CT image series of skulls can provide a solid foundation not only for comparative anatomical and evolutionary studies but also for the advancement of veterinary education, virtual surgery planning, and the facilitation of training in sophisticated machine learning methodologies.

Fetal-BET: Brain Extraction Tool for Fetal MRI

Goal: In this study, we address the critical challenge of fetal brain extraction from MRI sequences. Fetal MRI has played a crucial role in prenatal neurodevelopmental studies and in advancing our knowledge of fetal brain development in-utero. Fetal brain extraction is a necessary first step in most computational fetal brain MRI pipelines. However, it poses significant challenges due to 1) non-standard fetal head positioning, 2) fetal movements during examination, and 3) vastly heterogeneous appearance of the developing fetal brain and the neighboring fetal and maternal anatomy across gestation, and with various sequences and scanning conditions. Development of a machine learning method to effectively address this task requires a large and rich labeled dataset that has not been previously available. Currently, there is no method for accurate fetal brain extraction on various fetal MRI sequences. Methods: In this work, we first built a large annotated dataset of approximately 72,000 2D fetal brain MRI images. Our dataset covers the three common MRI sequences including T2-weighted, diffusion-weighted, and functional MRI acquired with different scanners. These data include images of normal and pathological brains. Using this dataset, we developed and validated deep learning methods, by exploiting the power of the U-Net style architectures, the attention mechanism, feature learning across multiple MRI modalities, and data augmentation for fast, accurate, and generalizable automatic fetal brain extraction. Results: Evaluations on independent test data, including data available from other centers, show that our method achieves accurate brain extraction on heterogeneous test data acquired with different scanners, on pathological brains, and at various gestational stages. Conclusions:By leveraging rich information from diverse multi-modality fetal MRI data, our proposed deep learning solution enables precise delineation of the fetal brain on various fetal MRI sequences. The robustness of our deep learning model underscores its potential utility for fetal brain imaging.

AI-based differential diagnosis of dementia etiologies on multimodal data

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an artificial intelligence (AI) model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a microaveraged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the microaveraged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in clinical settings and drug trials. Further prospective studies are needed to confirm its ability to improve patient care.

Cerebrospinal fluid volume as an early radiological factor for clinical course prediction after aneurysmal subarachnoid hemorrhage. A pilot study

BACKGROUND

The pathological mechanisms following aneurysmal subarachnoid hemorrhage (SAH) are poorly understood. Limited clinical evidence exists on the association between cerebrospinal fluid (CSF) volume and the risk of delayed cerebral ischemia (DCI) or cerebral vasospasm (CV). In this study, we raised the hypothesis that the amount of CSF or its ratio to hemorrhage blood volume, as determined from non-contrast Computed Tomography (NCCT) images taken on admission, could be a significant predictor for CV and DCI.

METHODS

The pilot study included a retrospective analysis of NCCT scans of 49 SAH patients taken shortly after an aneurysm rupture (33 males, 16 females, mean age 56.4 ± 15 years). The SynthStrip and Slicer3D software tools were used to extract radiological factors - CSF, brain, and hemorrhage volumes from the NCCT images. The "pure" CSF volume (VCSF) was estimated in the range of [-15, 15] Hounsfield units (HU).

RESULTS

VCSF was negatively associated with the risk of CV occurrence (p = 0.0049) and DCI (p = 0.0069), but was not associated with patients' outcomes. The hemorrhage volume (VSAH) was positively associated with an unfavorable outcome (p = 0.0032) but was not associated with CV/DCI. The ratio VSAH/VCSF was positively associated with, both, DCI (p = 0.031) and unfavorable outcome (p = 0.002). The CSF volume normalized by the brain volume showed the highest characteristics for DCI prediction (AUC = 0.791, sensitivity = 0.80, specificity = 0.812) and CV prediction (AUC = 0.769, sensitivity = 0.812, specificity = 0.70).

CONCLUSION

It was demonstrated that "pure" CSF volume retrieved from the initial NCCT images of SAH patients (including CV, Non-CV, DCI, Non-DCI groups) is a more significant predictor of DCI and CV compared to other routinely used radiological biomarkers. VCSF could be used to predict clinical course as well as to personalize the management of SAH patients. Larger multicenter clinical trials should be performed to test the added value of the proposed methodology.

Increased mean diffusivity of the caudal motor SNc identifies patients with REM sleep behaviour disorder and Parkinson's disease

Idiopathic rapid eye movement sleep behaviour disorder (iRBD)-a Parkinson's disease (PD) prodrome-might exhibit neural changes similar to those in PD. Substantia nigra pars compacta (SNc) degeneration underlies motor symptoms of PD. In iRBD and early PD (ePD), we measured diffusion MRI (dMRI) in the caudal motor SNc, which overlaps the nigrosome-1-the earliest-degenerating dopaminergic neurons in PD-and in the striatum. Nineteen iRBD, 26 ePD (1.7 ± 0.03 years), and 46 age-matched healthy controls (HCs) were scanned at Western University, and 47 iRBD, 115 ePD (0.9 ± 0.01 years), and 56 HCs were scanned through the Parkinson's Progression Markers Initiative, using 3T MRI. We segmented the SNc and striatum into subregions using automated probabilistic tractography to the cortex. We measured mean diffusivity (MD) and fractional anisotropy (FA) along white-matter bundles and subregional surfaces. We performed group-level and classification analyses. Increased caudal motor SNc surface MD was the only iRBD-HCs and ePD-HCs difference replicating across datasets (padj < 0.05). No iRBD-ePD differences emerged. Caudal motor SNc surface MD classified patient groups from HCs at the single-subject level with good-to-excellent balanced accuracy in an independent sample (0.91 iRBD and 0.86 iRBD and ePD combined), compared to fair performance for total SNc surface MD (0.72 iRBD and ePD). Caudal motor SNc surface MD correlated significantly with MDS-UPDRS-III scores in ePD patients. Using dMRI and automated segmentation, we detected changes suggesting altered microstructural integrity in iRBD and ePD in the nigrostriatal subregion known to degenerate first in PD. Surface MD of the caudal motor SNc presents a potential measure for inclusion in neuroimaging biomarkers of iRBD and PD.

Anatomy-aware and acquisition-agnostic joint registration with SynthMorph

Affine image registration is a cornerstone of medical-image analysis. While classical algorithms can achieve excellent accuracy, they solve a time-consuming optimization for every image pair. Deep-learning (DL) methods learn a function that maps an image pair to an output transform. Evaluating the function is fast, but capturing large transforms can be challenging, and networks tend to struggle if a test-image characteristic shifts from the training domain, such as the resolution. Most affine methods are agnostic to the anatomy the user wishes to align, meaning the registration will be inaccurate if algorithms consider all structures in the image. We address these shortcomings with SynthMorph, a fast, symmetric, diffeomorphic, and easy-to-use DL tool for joint affine-deformable registration of any brain image without preprocessing. First, we leverage a strategy that trains networks with widely varying images synthesized from label maps, yielding robust performance across acquisition specifics unseen at training. Second, we optimize the spatial overlap of select anatomical labels. This enables networks to distinguish anatomy of interest from irrelevant structures, removing the need for preprocessing that excludes content which would impinge on anatomy-specific registration. Third, we combine the affine model with a deformable hypernetwork that lets users choose the optimal deformation-field regularity for their specific data, at registration time, in a fraction of the time required by classical methods. This framework is applicable to learning anatomy-aware, acquisition-agnostic registration of any anatomy with any architecture, as long as label maps are available for training. We analyze how competing architectures learn affine transforms and compare state-of-the-art registration tools across an extremely diverse set of neuroimaging data, aiming to truly capture the behavior of methods in the real world. SynthMorph demonstrates high accuracy and is available at https://w3id.org/synthmorph, as a single complete end-to-end solution for registration of brain magnetic resonance imaging (MRI) data.

PyFaceWipe: a new defacing tool for almost any MRI contrast

RATIONALE AND OBJECTIVES

Defacing research MRI brain scans is often a mandatory step. With current defacing software, there are issues with Windows compatibility and researcher doubt regarding the adequacy of preservation of brain voxels in non-T1w scans. To address this, we developed PyFaceWipe, a multiplatform software for multiple MRI contrasts, which was evaluated based on its anonymisation ability and effect on downstream processing.

MATERIALS AND METHODS

Multiple MRI brain scan contrasts from the OASIS-3 dataset were defaced with PyFaceWipe and PyDeface and manually assessed for brain voxel preservation, remnant facial features and effect on automated face detection. Original and PyFaceWipe-defaced data from locally acquired T1w structural scans underwent volumetry with FastSurfer and brain atlas generation with ANTS.

RESULTS

214 MRI scans of several contrasts from OASIS-3 were successfully processed with both PyFaceWipe and PyDeface. PyFaceWipe maintained complete brain voxel preservation in all tested contrasts except ASL (45%) and DWI (90%), and PyDeface in all tested contrasts except ASL (95%), BOLD (25%), DWI (40%) and T2* (25%). Manual review of PyFaceWipe showed no failures of facial feature removal. Pinna removal was less successful (6% of T1 scans showed residual complete pinna). PyDeface achieved 5.1% failure rate. Automated detection found no faces in PyFaceWipe-defaced scans, 19 faces in PyDeface scans compared with 78 from the 224 original scans. Brain atlas generation showed no significant difference between atlases created from original and defaced data in both young adulthood and late elderly cohorts. Structural volumetry dice scores were ≥ 0.98 for all structures except for grey matter which had 0.93. PyFaceWipe output was identical across the tested operating systems.

CONCLUSION

PyFaceWipe is a promising multiplatform defacing tool, demonstrating excellent brain voxel preservation and competitive defacing in multiple MRI contrasts, performing favourably against PyDeface. ASL, BOLD, DWI and T2* scans did not produce recognisable 3D renders and hence should not require defacing. Structural volumetry dice scores (≥ 0.98) were higher than previously published FreeSurfer results, except for grey matter which were comparable. The effect is measurable and care should be exercised during studies. ANTS atlas creation showed no significant effect from PyFaceWipe defacing.

The Advanced BRain Imaging on ageing and Memory (ABRIM) data collection: Study design, data processing, and rationale

To understand the neurocognitive mechanisms that underlie heterogeneity in cognitive ageing, recent scientific efforts have led to a growing public availability of imaging cohort data. The Advanced BRain Imaging on ageing and Memory (ABRIM) project aims to add to these existing datasets by taking an adult lifespan approach to provide a cross-sectional, normative database with a particular focus on connectivity, myelinization and iron content of the brain in concurrence with cognitive functioning, mechanisms of reserve, and sleep-wake rhythms. ABRIM freely shares MRI and behavioural data from 295 participants between 18-80 years, stratified by age decade and sex (median age 52, IQR 36-66, 53.20% females). The ABRIM MRI collection consists of both the raw and pre-processed structural and functional MRI data to facilitate data usage among both expert and non-expert users. The ABRIM behavioural collection includes measures of cognitive functioning (i.e., global cognition, processing speed, executive functions, and memory), proxy measures of cognitive reserve (e.g., educational attainment, verbal intelligence, and occupational complexity), and various self-reported questionnaires (e.g., on depressive symptoms, pain, and the use of memory strategies in daily life and during a memory task). In a sub-sample (n = 120), we recorded sleep-wake rhythms using an actigraphy device (Actiwatch 2, Philips Respironics) for a period of 7 consecutive days. Here, we provide an in-depth description of our study protocol, pre-processing pipelines, and data availability. ABRIM provides a cross-sectional database on healthy participants throughout the adult lifespan, including numerous parameters relevant to improve our understanding of cognitive ageing. Therefore, ABRIM enables researchers to model the advanced imaging parameters and cognitive topologies as a function of age, identify the normal range of values of such parameters, and to further investigate the diverse mechanisms of reserve and resilience.

A Set of FMRI Quality Control Tools in AFNI: Systematic, in-depth and interactive QC with afni_proc.py and more

Quality control (QC) assessment is a vital part of FMRI processing and analysis, and a typically under-discussed aspect of reproducibility. This includes checking datasets at their very earliest stages (acquisition and conversion) through their processing steps (e.g., alignment and motion correction) to regression modeling (correct stimuli, no collinearity, valid fits, enough degrees of freedom, etc.) for each subject. There are a wide variety of features to verify throughout any single subject processing pipeline, both quantitatively and qualitatively. We present several FMRI preprocessing QC features available in the AFNI toolbox, many of which are automatically generated by the pipeline-creation tool, afni_proc.py. These items include: a modular HTML document that covers full single subject processing from the raw data through statistical modeling; several review scripts in the results directory of processed data; and command line tools for identifying subjects with one or more quantitative properties across a group (such as triaging warnings, making exclusion criteria or creating informational tables). The HTML itself contains several buttons that efficiently facilitate interactive investigations into the data, when deeper checks are needed beyond the systematic images. The pages are linkable, so that users can evaluate individual items across a group, for increased sensitivity to differences (e.g., in alignment or regression modeling images). Finally, the QC document contains rating buttons for each "QC block", as well as comment fields for each, to facilitate both saving and sharing the evaluations. This increases the specificity of QC, as well as its shareability, as these files can be shared with others and potentially uploaded into repositories, promoting transparency and open science. We describe the features and applications of these QC tools for FMRI.

Predicting Language Function Post-Stroke: A Model-Based Structural Connectivity Approach

BACKGROUND

The prediction of post-stroke language function is essential for the development of individualized treatment plans based on the personal recovery potential of aphasic stroke patients.

OBJECTIVE

To establish a framework for integrating information on connectivity disruption of the language network based on routinely collected clinical magnetic resonance (MR) images into Random Forest modeling to predict post-stroke language function.

METHODS

Language function was assessed in 76 stroke patients from the Non-Invasive Repeated Therapeutic Stimulation for Aphasia Recovery trial, using the Token Test (TT), Boston Naming Test (BNT), and Semantic Verbal Fluency (sVF) Test as primary outcome measures. Individual infarct masks were superimposed onto a diffusion tensor imaging tractogram reference set to calculate Change in Connectivity scores of language-relevant gray matter regions as estimates of structural connectivity disruption. Multivariable Random Forest models were derived to predict language function.

RESULTS

Random Forest models explained moderate to high amount of variance at baseline and follow-up for the TT (62.7% and 76.2%), BNT (47.0% and 84.3%), and sVF (52.2% and 61.1%). Initial language function and non-verbal cognitive ability were the most important variables to predict language function. Connectivity disruption explained additional variance, resulting in a prediction error increase of up to 12.8% with variable omission. Left middle temporal gyrus (12.8%) and supramarginal gyrus (9.8%) were identified as among the most important network nodes.

CONCLUSION

Connectivity disruption of the language network adds predictive value beyond lesion volume, initial language function, and non-verbal cognitive ability. Obtaining information on connectivity disruption based on routine clinical MR images constitutes a significant advancement toward practical clinical application.

Super resolution using sparse sampling at portable ultra-low field MR

Ultra-low field (ULF) magnetic resonance imaging (MRI) holds the potential to make MRI more accessible, given its cost-effectiveness, reduced power requirements, and portability. However, signal-to-noise ratio (SNR) drops with field strength, necessitating imaging with lower resolution and longer scan times. This study introduces a novel Fourier-based Super Resolution (FouSR) approach, designed to enhance the resolution of ULF MRI images with minimal increase in total scan time. FouSR combines spatial frequencies from two orthogonal ULF images of anisotropic resolution to create an isotropic T2-weighted fluid-attenuated inversion recovery (FLAIR) image. We hypothesized that FouSR could effectively recover information from under-sampled slice directions, thereby improving the delineation of multiple sclerosis (MS) lesions and other significant anatomical features. Importantly, the FouSR algorithm can be implemented on the scanner with changes to the k-space trajectory. Paired ULF (Hyperfine SWOOP, 0.064 tesla) and high field (Siemens, Skyra, 3 Tesla) FLAIR scans were collected on the same day from a phantom and a cohort of 10 participants with MS or suspected MS (6 female; mean ± SD age: 44.1 ± 4.1). ULF scans were acquired along both coronal and axial planes, featuring an in-plane resolution of 1.7 mm × 1.7 mm with a slice thickness of 5 mm. FouSR was evaluated against registered ULF coronal and axial scans, their average (ULF average) and a gold standard SR (ANTs SR). FouSR exhibited higher SNR (47.96 ± 12.6) compared to ULF coronal (36.7 ± 12.2) and higher lesion conspicuity (0.12 ± 0.06) compared to ULF axial (0.13 ± 0.07) but did not exhibit any significant differences contrast-to-noise-ratio (CNR) compared to other methods in patient scans. However, FouSR demonstrated superior image sharpness (0.025 ± 0.0040) compared to all other techniques (ULF coronal 0.021 ± 0.0037, q = 5.9, p-adj. = 0.011; ULF axial 0.018 ± 0.0026, q = 11.1, p-adj. = 0.0001; ULF average 0.019 ± 0.0034, q = 24.2, p-adj. < 0.0001) and higher lesion sharpness (-0.97 ± 0.31) when compared to the ULF average (-1.02 ± 0.37, t(543) = -10.174, p = <0.0001). Average blinded qualitative assessment by three experienced MS neurologists showed no significant difference in WML and sulci or gyri visualization between FouSR and other methods. FouSR can, in principle, be implemented on the scanner to produce clinically useful FLAIR images at higher resolution on the fly, providing a valuable tool for visualizing lesions and other anatomical structures in MS.

BOOSTING SKULL-STRIPPING PERFORMANCE FOR PEDIATRIC BRAIN IMAGES

Skull-stripping is the removal of background and non-brain anatomical features from brain images. While many skull-stripping tools exist, few target pediatric populations. With the emergence of multi-institutional pediatric data acquisition efforts to broaden the understanding of perinatal brain development, it is essential to develop robust and well-tested tools ready for the relevant data processing. However, the broad range of neuroanatomical variation in the developing brain, combined with additional challenges such as high motion levels, as well as shoulder and chest signal in the images, leaves many adult-specific tools ill-suited for pediatric skull-stripping. Building on an existing framework for robust and accurate skull-stripping, we propose developmental SynthStrip (d-SynthStrip), a skull-stripping model tailored to pediatric images. This framework exposes networks to highly variable images synthesized from label maps. Our model substantially outperforms pediatric baselines across scan types and age cohorts. In addition, the <1-minute runtime of our tool compares favorably to the fastest baselines. We distribute our model at https://w3id.org/synthstrip.

Enhanced Spatial Fuzzy C-Means Algorithm for Brain Tissue Segmentation in T1 Images

Magnetic Resonance Imaging (MRI) plays an important role in neurology, particularly in the precise segmentation of brain tissues. Accurate segmentation is crucial for diagnosing brain injuries and neurodegenerative conditions. We introduce an Enhanced Spatial Fuzzy C-means (esFCM) algorithm for 3D T1 MRI segmentation to three tissues, i.e. White Matter (WM), Gray Matter (GM), and Cerebrospinal Fluid (CSF). The esFCM employs a weighted least square algorithm utilizing the Structural Similarity Index (SSIM) for polynomial bias field correction. It also takes advantage of the information from the membership function of the last iteration to compute neighborhood impact. This strategic refinement enhances the algorithm's adaptability to complex image structures, effectively addressing challenges such as intensity irregularities and contributing to heightened segmentation accuracy. We compare the segmentation accuracy of esFCM against four variants of FCM, Gaussian Mixture Model (GMM) and FSL and ANTs algorithms using four various dataset, employing three measurement criteria. Comparative assessments underscore esFCM's superior performance, particularly in scenarios involving added noise and bias fields.The obtained results emphasize the significant potential of the proposed method in the segmentation of MRI images.

Characterization and Mitigation of a Simultaneous Multi-Slice fMRI Artifact: Multiband Artifact Regression in Simultaneous Slices

Simultaneous multi-slice (multiband) acceleration in fMRI has become widespread, but may be affected by novel forms of signal artifact. Here, we demonstrate a previously unreported artifact manifesting as a shared signal between simultaneously acquired slices in all resting-state and task-based multiband fMRI datasets we investigated, including publicly available consortium data from the Human Connectome Project (HCP) and Adolescent Brain Cognitive Development (ABCD) Study. We propose Multiband Artifact Regression in Simultaneous Slices (MARSS), a regression-based detection and correction technique that successfully mitigates this shared signal in unprocessed data. We demonstrate that the signal isolated by MARSS correction is likely non-neural, appearing stronger in neurovasculature than grey matter. Additionally, we evaluate MARSS both against and in tandem with sICA+FIX denoising, which is implemented in HCP resting-state data, to show that MARSS mitigates residual artifact signal that is not modeled by sICA+FIX. MARSS correction leads to study-wide increases in signal-to-noise ratio, decreases in cortical coefficient of variation, and mitigation of systematic artefactual spatial patterns in participant-level task betas. Finally, MARSS correction has substantive effects on second-level t-statistics in analyses of task-evoked activation. We recommend that investigators apply MARSS to multiband fMRI datasets with moderate or higher acceleration factors, in combination with established denoising methods.

Unconstrained quantitative magnetization transfer imaging: disentangling T1 of the free and semi-solid spin pools

Since the inception of magnetization transfer (MT) imaging, it has been widely assumed that Henkelman's two spin pools have similar longitudinal relaxation times, which motivated many researchers to constrain them to each other. However, several recent publications reported a T 1 s of the semi-solid spin pool that is much shorter than T 1 f of the free pool. While these studies tailored experiments for robust proofs-of-concept, we here aim to quantify the disentangled relaxation processes on a voxel-by-voxel basis in a clinical imaging setting, i.e., with an effective resolution of 1.24mm isotropic and full brain coverage in 12min. To this end, we optimized a hybrid-state pulse sequence for mapping the parameters of an unconstrained MT model. We scanned four people with relapsing-remitting multiple sclerosis (MS) and four healthy controls with this pulse sequence and estimated T 1 f ≈ 1.84 s and T 1 s ≈ 0.34 s in healthy white matter. Our results confirm the reports that T 1 s ≪ T 1 f and we argue that this finding identifies MT as an inherent driver of longitudinal relaxation in brain tissue. Moreover, we estimated a fractional size of the semi-solid spin pool of m 0 s ≈ 0.212 , which is larger than previously assumed. An analysis of T 1 f in normal-appearing white matter revealed statistically significant differences between individuals with MS and controls.

AI-based differential diagnosis of dementia etiologies on multimodal data

Differential diagnosis of dementia remains a challenge in neurology due to symptom overlap across etiologies, yet it is crucial for formulating early, personalized management strategies. Here, we present an AI model that harnesses a broad array of data, including demographics, individual and family medical history, medication use, neuropsychological assessments, functional evaluations, and multimodal neuroimaging, to identify the etiologies contributing to dementia in individuals. The study, drawing on 51,269 participants across 9 independent, geographically diverse datasets, facilitated the identification of 10 distinct dementia etiologies. It aligns diagnoses with similar management strategies, ensuring robust predictions even with incomplete data. Our model achieved a micro-averaged area under the receiver operating characteristic curve (AUROC) of 0.94 in classifying individuals with normal cognition, mild cognitive impairment and dementia. Also, the micro-averaged AUROC was 0.96 in differentiating the dementia etiologies. Our model demonstrated proficiency in addressing mixed dementia cases, with a mean AUROC of 0.78 for two co-occurring pathologies. In a randomly selected subset of 100 cases, the AUROC of neurologist assessments augmented by our AI model exceeded neurologist-only evaluations by 26.25%. Furthermore, our model predictions aligned with biomarker evidence and its associations with different proteinopathies were substantiated through postmortem findings. Our framework has the potential to be integrated as a screening tool for dementia in various clinical settings and drug trials, with promising implications for person-level management.

Greater white matter degeneration and lower structural connectivity in non-amnestic vs. amnestic Alzheimer's disease

Introduction

Multimodal evidence indicates Alzheimer's disease (AD) is characterized by early white matter (WM) changes that precede overt cognitive impairment. WM changes have overwhelmingly been investigated in typical, amnestic mild cognitive impairment and AD; fewer studies have addressed WM change in atypical, non-amnestic syndromes. We hypothesized each non-amnestic AD syndrome would exhibit WM differences from amnestic and other non-amnestic syndromes.

Materials and methods

Participants included 45 cognitively normal (CN) individuals; 41 amnestic AD patients; and 67 patients with non-amnestic AD syndromes including logopenic-variant primary progressive aphasia (lvPPA, n = 32), posterior cortical atrophy (PCA, n = 17), behavioral variant AD (bvAD, n = 10), and corticobasal syndrome (CBS, n = 8). All had T1-weighted MRI and 30-direction diffusion-weighted imaging (DWI). We performed whole-brain deterministic tractography between 148 cortical and subcortical regions; connection strength was quantified by tractwise mean generalized fractional anisotropy. Regression models assessed effects of group and phenotype as well as associations with grey matter volume. Topological analyses assessed differences in persistent homology (numbers of graph components and cycles). Additionally, we tested associations of topological metrics with global cognition, disease duration, and DWI microstructural metrics.

Results

Both amnestic and non-amnestic patients exhibited lower WM connection strength than CN participants in corpus callosum, cingulum, and inferior and superior longitudinal fasciculi. Overall, non-amnestic patients had more WM disease than amnestic patients. LvPPA patients had left-lateralized WM degeneration; PCA patients had reductions in connections to bilateral posterior parietal, occipital, and temporal areas. Topological analysis showed the non-amnestic but not the amnestic group had more connected components than controls, indicating persistently lower connectivity. Longer disease duration and cognitive impairment were associated with more connected components and fewer cycles in individuals' brain graphs.

Discussion

We have previously reported syndromic differences in GM degeneration and tau accumulation between AD syndromes; here we find corresponding differences in WM tracts connecting syndrome-specific epicenters. Determining the reasons for selective WM degeneration in non-amnestic AD is a research priority that will require integration of knowledge from neuroimaging, biomarker, autopsy, and functional genetic studies. Furthermore, longitudinal studies to determine the chronology of WM vs. GM degeneration will be key to assessing evidence for WM-mediated tau spread.

Study of multistep Dense U-Net-based automatic segmentation for head MRI scans

BACKGROUND

Despite extensive efforts to obtain accurate segmentation of magnetic resonance imaging (MRI) scans of a head, it remains challenging primarily due to variations in intensity distribution, which depend on the equipment and parameters used.

PURPOSE

The goal of this study is to evaluate the effectiveness of an automatic segmentation method for head MRI scans using a multistep Dense U-Net (MDU-Net) architecture.

METHODS

The MDU-Net-based method comprises two steps. The first step is to segment the scalp, skull, and whole brain from head MRI scans using a convolutional neural network (CNN). In the first step, a hybrid network is used to combine 2.5D Dense U-Net and 3D Dense U-Net structure. This hybrid network acquires logits in three orthogonal planes (axial, coronal, and sagittal) using 2.5D Dense U-Nets and fuses them by averaging. The resultant fused probability map with head MRI scans then serves as the input to a 3D Dense U-Net. In this process, different ratios of active contour loss and focal loss are applied. The second step is to segment the cerebrospinal fluid (CSF), white matter, and gray matter from extracted brain MRI scans using CNNs. In the second step, the histogram of the extracted brain MRI scans is standardized and then a 2.5D Dense U-Net is used to further segment the brain's specific tissues using the focal loss. A dataset of 100 head MRI scans from an OASIS-3 dataset was used for training, internal validation, and testing, with ratios of 80%, 10%, and 10%, respectively. Using the proposed approach, we segmented the head MRI scans into five areas (scalp, skull, CSF, white matter, and gray matter) and evaluated the segmentation results using the Dice similarity coefficient (DSC) score, Hausdorff distance (HD), and the average symmetric surface distance (ASSD) as evaluation metrics. We compared these results with those obtained using the Res-U-Net, Dense U-Net, U-Net++, Swin-Unet, and H-Dense U-Net models.

RESULTS

The MDU-Net model showed DSC values of 0.933, 0.830, 0.833, 0.953, and 0.917 in the scalp, skull, CSF, white matter, and gray matter, respectively. The corresponding HD values were 2.37, 2.89, 2.13, 1.52, and 1.53 mm, respectively. The ASSD values were 0.50, 1.63, 1.28, 0.26, and 0.27 mm, respectively. Comparing these results with other models revealed that the MDU-Net model demonstrated the best performance in terms of the DSC values for the scalp, CSF, white matter, and gray matter. When compared with the H-Dense U-Net model, which showed the highest performance among the other models, the MDU-Net model showed substantial improvements in the HD view, particularly in the gray matter region, with a difference of approximately 9%. In addition, in terms of the ASSD, the MDU-Net model outperformed the H-Dense U-Net model, showing an approximately 7% improvements in the white matter and approximately 9% improvements in the gray matter.

CONCLUSION

Compared with existing models in terms of DSC, HD, and ASSD, the proposed MDU-Net model demonstrated the best performance on average and showed its potential to enhance the accuracy of automatic segmentation for head MRI scans.

Automated ASPECTS calculation may equal the performance of experienced clinicians: a machine learning study based on a large cohort

OBJECTIVES

The Alberta Stroke Program Early CT Score (ASPECTS) is a semi-quantitative method to evaluate the severity of early ischemic change on non-contrast computed tomography (NCCT) in patients with acute ischemic stroke (AIS). In this work, we propose an automated ASPECTS method based on large cohort of data and machine learning.

METHODS

For this study, we collected 3626 NCCT cases from multiple centers and annotated directly on this dataset by neurologists. Based on image analysis and machine learning methods, we constructed a two-stage machine learning model. The validity and reliability of this automated ASPECTS method were tested on an independent external validation set of 300 cases. Statistical analyses on the total ASPECTS, dichotomized ASPECTS, and region-level ASPECTS were presented.

RESULTS

On an independent external validation set of 300 cases, for the total ASPECTS results, the intraclass correlation coefficient between automated ASPECTS and expert-rated was 0.842. The agreement between ASPECTS threshold of ≥ 6 versus < 6 using a dichotomized method was moderate (κ = 0.438, 0.391-0.477), and the detection rate (sensitivity) was 86.5% for patients with ASPECTS threshold of ≥ 6. Compared with the results of previous studies, our method achieved a slight lead in sensitivity (67.8%) and AUC (0.845), with comparable accuracy (78.9%) and specificity (81.2%).

CONCLUSION

The proposed automated ASPECTS method driven by a large cohort of NCCT images performed equally well compared with expert-rated ASPECTS. This work further demonstrates the validity and reliability of automated ASPECTS evaluation method.

CLINICAL RELEVANCE STATEMENT

The automated ASPECTS method proposed by this study may help AIS patients to receive rapid intervention, but should not be used as a stand-alone diagnostic basis.

KEY POINTS

NCCT-based manual ASPECTS scores were poorly consistent. Machine learning can automate the ASPECTS scoring process. Machine learning model design based on large cohort data can effectively improve the consistency of ASPECTS scores.

BOOSTING SKULL-STRIPPING PERFORMANCE FOR PEDIATRIC BRAIN IMAGES

Skull-stripping is the removal of background and non-brain anatomical features from brain images. While many skull-stripping tools exist, few target pediatric populations. With the emergence of multi-institutional pediatric data acquisition efforts to broaden the understanding of perinatal brain development, it is essential to develop robust and well-tested tools ready for the relevant data processing. However, the broad range of neuroanatomical variation in the developing brain, combined with additional challenges such as high motion levels, as well as shoulder and chest signal in the images, leaves many adult-specific tools ill-suited for pediatric skull-stripping. Building on an existing framework for robust and accurate skull-stripping, we propose developmental SynthStrip (d-SynthStrip), a skull-stripping model tailored to pediatric images. This framework exposes networks to highly variable images synthesized from label maps. Our model substantially outperforms pediatric baselines across scan types and age cohorts. In addition, the <1-minute runtime of our tool compares favorably to the fastest baselines. We distribute our model at https://w3id.org/synthstrip.

Multi-scale V-net architecture with deep feature CRF layers for brain extraction

BACKGROUND

Brain extraction is a computational necessity for researchers using brain imaging data. However, the complex structure of the interfaces between the brain, meninges and human skull have not allowed a highly robust solution to emerge. While previous methods have used machine learning with structural and geometric priors in mind, with the development of Deep Learning (DL), there has been an increase in Neural Network based methods. Most proposed DL models focus on improving the training data despite the clear gap between groups in the amount and quality of accessible training data between.

METHODS

We propose an architecture we call Efficient V-net with Additional Conditional Random Field Layers (EVAC+). EVAC+ has 3 major characteristics: (1) a smart augmentation strategy that improves training efficiency, (2) a unique way of using a Conditional Random Fields Recurrent Layer that improves accuracy and (3) an additional loss function that fine-tunes the segmentation output. We compare our model to state-of-the-art non-DL and DL methods.

RESULTS

Results show that even with limited training resources, EVAC+ outperforms in most cases, achieving a high and stable Dice Coefficient and Jaccard Index along with a desirable lower Surface (Hausdorff) Distance. More importantly, our approach accurately segmented clinical and pediatric data, despite the fact that the training dataset only contains healthy adults.

CONCLUSIONS

Ultimately, our model provides a reliable way of accurately reducing segmentation errors in complex multi-tissue interfacing areas of the brain. We expect our method, which is publicly available and open-source, to be beneficial to a wide range of researchers.

DCP: A pipeline toolbox for diffusion connectome

The brain structural network derived from diffusion magnetic resonance imaging (dMRI) reflects the white matter connections between brain regions, which can quantitatively describe the anatomical connection pattern of the entire brain. The development of structural brain connectome leads to the emergence of a large number of dMRI processing packages and network analysis toolboxes. However, the fully automated network analysis based on dMRI data remains challenging. In this study, we developed a cross-platform MATLAB toolbox named "Diffusion Connectome Pipeline" (DCP) for automatically constructing brain structural networks and calculating topological attributes of the networks. The toolbox integrates a few developed packages, including FSL, Diffusion Toolkit, SPM, Camino, MRtrix3, and MRIcron. It can process raw dMRI data collected from any number of participants, and it is also compatible with preprocessed files from public datasets such as HCP and UK Biobank. Moreover, a friendly graphical user interface allows users to configure their processing pipeline without any programming. To prove the capacity and validity of the DCP, two tests were conducted with using DCP. The results showed that DCP can reproduce the findings in our previous studies. However, there are some limitations of DCP, such as relying on MATLAB and being unable to fixel-based metrics weighted network. Despite these limitations, overall, the DCP software provides a standardized, fully automated computational workflow for white matter network construction and analysis, which is beneficial for advancing future human brain connectomics application research.

Two is better than one: longitudinal detection and volumetric evaluation of brain metastases after Stereotactic Radiosurgery with a deep learning pipeline

PURPOSE

Close MRI surveillance of patients with brain metastases following Stereotactic Radiosurgery (SRS) treatment is essential for assessing treatment response and the current disease status in the brain. This follow-up necessitates the comparison of target lesion sizes in pre- (prior) and post-SRS treatment (current) T1W-Gad MRI scans. Our aim was to evaluate SimU-Net, a novel deep-learning model for the detection and volumetric analysis of brain metastases and their temporal changes in paired prior and current scans.

METHODS

SimU-Net is a simultaneous multi-channel 3D U-Net model trained on pairs of registered prior and current scans of a patient. We evaluated its performance on 271 pairs of T1W-Gad MRI scans from 226 patients who underwent SRS. An expert oncological neurosurgeon manually delineated 1,889 brain metastases in all the MRI scans (1,368 with diameters > 5 mm, 834 > 10 mm). The SimU-Net model was trained/validated on 205 pairs from 169 patients (1,360 metastases) and tested on 66 pairs from 57 patients (529 metastases). The results were then compared to the ground truth delineations.

RESULTS

SimU-Net yielded a mean (std) detection precision and recall of 1.00±0.00 and 0.99±0.06 for metastases > 10 mm, 0.90±0.22 and 0.97±0.12 for metastases > 5 mm of, and 0.76±0.27 and 0.94±0.16 for metastases of all sizes. It improves lesion detection precision by 8% for all metastases sizes and by 12.5% for metastases < 10 mm with respect to standalone 3D U-Net. The segmentation Dice scores were 0.90±0.10, 0.89±0.10 and 0.89±0.10 for the above metastases sizes, all above the observer variability of 0.80±0.13.

CONCLUSION

Automated detection and volumetric quantification of brain metastases following SRS have the potential to enhance the assessment of treatment response and alleviate the clinician workload.

Feasibility of high-resolution perfusion imaging using arterial spin labeling MRI at 3 Tesla

Cerebral blood flow (CBF) is a critical physiological parameter of brain health, and it can be non-invasively measured with arterial spin labeling (ASL) MRI. In this study, we evaluated and optimized whole-brain, high-resolution ASL as an alternative to the low-resolution ASL employed in the routine assessment of CBF in both healthy participants and patients. Two high-resolution protocols (i.e., pCASL and FAIR-Q2TIPS (PASL) with 2 mm isotropic voxels) were compared to a default clinical pCASL protocol (3.4 × 3.4 × 4 mm 3), all of whom had an acquisition time of ≈ 5 min. We assessed the impact of high-resolution acquisition on reducing partial voluming and improving sensitivity to the perfusion signal, and evaluated the effectiveness of z-deblurring on the ASL data. We compared the quality of whole-brain ASL acquired using three available head coils with differing number of receive channels (i.e., 20, 32, and 64ch). We found that using higher coil counts (32 and 64ch coils as compared to 20ch) offers improved signal-to-noise ratio (SNR) and acceleration capabilities that are beneficial for ASL imaging at 3 Tesla (3 T). The inherent reduction in partial voluming effects with higher resolution acquisitions improves the resolving power of perfusion without impacting the sensitivity. In conclusion, our results suggest that high-resolution ASL (2 to 2.5 mm isotropic voxels) has the potential to become a new standard for perfusion imaging at 3 T and increase its adoption into clinical research and cognitive neuroscience applications.

White matter microstructure of children with sensory over-responsivity is associated with affective behavior

BACKGROUND

Sensory processing dysfunction (SPD) is linked to altered white matter (WM) microstructure in school-age children. Sensory over-responsivity (SOR), a form of SPD, affects at least 2.5% of all children and has substantial deleterious impact on learning and mental health. However, SOR has not been well studied using microstructural imaging such as diffusion MRI (dMRI). Since SOR involves hypersensitivity to external stimuli, we test the hypothesis that children with SOR require compensatory neuroplasticity in the form of superior WM microstructural integrity to protect against internalizing behavior, leaving those with impaired WM microstructure vulnerable to somatization and depression.

METHODS

Children ages 8-12 years old with neurodevelopmental concerns were assessed for SOR using a comprehensive structured clinical evaluation, the Sensory Processing 3 Dimensions Assessment, and underwent 3 Tesla MRI with multishell multiband dMRI. Tract-based spatial statistics was used to measure diffusion tensor imaging (DTI) and neurite orientation dispersion and density imaging (NODDI) metrics from global WM and nineteen selected WM tracts. Correlations of DTI and NODDI measures with measures of somatization and emotional disturbance from the Behavioral Assessment System for Children, 3rd edition (BASC-3), were computed in the SOR group and in matched children with neurodevelopmental concerns but not SOR.

RESULTS

Global WM fractional anisotropy (FA) is negatively correlated with somatization and with emotional disturbance in the SOR group but not the non-SOR group. Also observed in children with SOR are positive correlations of radial diffusivity (RD) and free water fraction (FISO) with somatization and, in most cases, emotional disturbance. These effects are significant in boys with SOR, whereas the study is underpowered for girls. The most affected white matter are medial lemniscus and internal capsule sensory tracts, although effects of SOR are observed in many cerebral, cerebellar, and brainstem tracts.

CONCLUSION

White matter microstructure is related to affective behavior in children with SOR.

Cortical Network Disruption Is Minimal in Early Stages of Psychosis

Background and Hypothesis

Schizophrenia is associated with white matter disruption and topological reorganization of cortical connectivity but the trajectory of these changes, from the first psychotic episode to established illness, is poorly understood. Current studies in first-episode psychosis (FEP) patients using diffusion magnetic resonance imaging (dMRI) suggest such disruption may be detectable at the onset of psychosis, but specific results vary widely, and few reports have contextualized their findings with direct comparison to young adults with established illness.

Study Design

Diffusion and T1-weighted 7T MR scans were obtained from N = 112 individuals (58 with untreated FEP, 17 with established schizophrenia, 37 healthy controls) recruited from London, Ontario. Voxel- and network-based analyses were used to detect changes in diffusion microstructural parameters. Graph theory metrics were used to probe changes in the cortical network hierarchy and to assess the vulnerability of hub regions to disruption. The analysis was replicated with N = 111 (57 patients, 54 controls) from the Human Connectome Project-Early Psychosis (HCP-EP) dataset.

Study Results

Widespread microstructural changes were found in people with established illness, but changes in FEP patients were minimal. Unlike the established illness group, no appreciable topological changes in the cortical network were observed in FEP patients. These results were replicated in the early psychosis patients of the HCP-EP datasets, which were indistinguishable from controls in most metrics.

Conclusions

The white matter structural changes observed in established schizophrenia are not a prominent feature in the early stages of this illness.

Histopathologically confirmed radiation-induced damage of the brain - an in-depth analysis of radiation parameters and spatio-temporal occurrence

BACKGROUND

Radiation-induced damage (RID) after radiotherapy (RT) of primary brain tumors and metastases can be challenging to clinico-radiographically distinguish from tumor progression. RID includes pseudoprogression and radiation necrosis; the latter being irreversible and often associated with severe symptoms. While histopathology constitutes the diagnostic gold standard, biopsy-controlled clinical studies investigating RID remain limited. Whether certain brain areas are potentially more vulnerable to RID remains an area of active investigation. Here, we analyze histopathologically confirmed cases of RID in relation to the temporal and spatial dose distribution.

METHODS

Histopathologically confirmed cases of RID after photon-based RT for primary or secondary central nervous system malignancies were included. Demographic, clinical, and dosimetric data were collected from patient records and treatment planning systems. We calculated the equivalent dose in 2 Gy fractions (EQD22) and the biologically effective dose (BED2) for normal brain tissue (α/β ratio of 2 Gy) and analyzed the spatial and temporal distribution using frequency maps.

RESULTS

Thirty-three patients were identified. High-grade glioma patients (n = 18) mostly received one normofractionated RT series (median cumulative EQD22 60 Gy) to a large planning target volume (PTV) (median 203.9 ccm) before diagnosis of RID. Despite the low EQD22 and BED2, three patients with an accelerated hyperfractionated RT developed RID. In contrast, brain metastases patients (n = 15; 16 RID lesions) were often treated with two or more RT courses and with radiosurgery or fractionated stereotactic RT, resulting in a higher cumulative EQD22 (median 162.4 Gy), to a small PTV (median 6.7 ccm). All (n = 34) RID lesions occurred within the PTV of at least one of the preceding RT courses. RID in the high-grade glioma group showed a frontotemporal distribution pattern, whereas, in metastatic patients, RID was observed throughout the brain with highest density in the parietal lobe. The cumulative EQD22 was significantly lower in RID lesions that involved the subventricular zone (SVZ) than in lesions without SVZ involvement (median 60 Gy vs. 141 Gy, p = 0.01).

CONCLUSIONS

Accelerated hyperfractionated RT can lead to RID despite computationally low EQD22 and BED2 in high-grade glioma patients. The anatomical location of RID corresponded to the general tumor distribution of gliomas and metastases. The SVZ might be a particularly vulnerable area.

Cerebrovascular disease drives Alzheimer plasma biomarker concentrations in adults with Down syndrome

Importance

By age 40 years over 90% of adults with Down syndrome (DS) have Alzheimer's disease (AD) pathology and most progress to dementia. Despite having few systemic vascular risk factors, individuals with DS have elevated cerebrovascular disease (CVD) markers that track with the clinical progression of AD, suggesting a role for CVD that is hypothesized to be mediated by inflammatory factors.

Objective

To examine the pathways through which small vessel CVD contributes to AD-related pathophysiology and neurodegeneration in adults with DS.

Design

Cross sectional analysis of neuroimaging, plasma, and clinical data.

Setting

Participants were enrolled in Alzheimer's Biomarker Consortium - Down Syndrome (ABC-DS), a multisite study of AD in adults with DS.

Participants

One hundred eighty-five participants (mean [SD] age=45.2 [9.3] years) with available MRI and plasma biomarker data were included. White matter hyperintensity (WMH) volumes were derived from T2-weighted FLAIR MRI scans and plasma biomarker concentrations of amyloid beta (Aβ42/Aβ40), phosphorylated tau (p-tau217), astrocytosis (glial fibrillary acidic protein, GFAP), and neurodegeneration (neurofilament light chain, NfL) were measured with ultrasensitive immunoassays.

Main Outcomes and Measures

We examined the bivariate relationships of WMH, Aβ42/Aβ40, p-tau217, and GFAP with age-residualized NfL across AD diagnostic groups. A series of mediation and path analyses examined causal pathways linking WMH and AD pathophysiology to promote neurodegeneration in the total sample and groups stratified by clinical diagnosis.

Results

There was a direct and indirect bidirectional effect through GFAP of WMH on p-tau217 concentration, which was associated with NfL concentration in the entire sample. Among cognitively stable participants, WMH was directly and indirectly, through GFAP, associated with p-tau217 concentration, and in those with MCI, there was a direct effect of WMH on p-tau217 and NfL concentrations. There were no associations of WMH with biomarker concentrations among those diagnosed with dementia.

Conclusions and Relevance

The findings suggest that among individuals with DS, CVD promotes neurodegeneration by increasing astrocytosis and tau pathophysiology in the presymptomatic phases of AD. This work joins an emerging literature that implicates CVD and its interface with neuroinflammation as a core pathological feature of AD in adults with DS.

Deep learning-based algorithm for postoperative glioblastoma MRI segmentation: a promising new tool for tumor burden assessment

OBJECTIVE

Clinical and surgical decisions for glioblastoma patients depend on a tumor imaging-based evaluation. Artificial Intelligence (AI) can be applied to magnetic resonance imaging (MRI) assessment to support clinical practice, surgery planning and prognostic predictions. In a real-world context, the current obstacles for AI are low-quality imaging and postoperative reliability. The aim of this study is to train an automatic algorithm for glioblastoma segmentation on a clinical MRI dataset and to obtain reliable results both pre- and post-operatively.

METHODS

The dataset used for this study comprises 237 (71 preoperative and 166 postoperative) MRIs from 71 patients affected by a histologically confirmed Grade IV Glioma. The implemented U-Net architecture was trained by transfer learning to perform the segmentation task on postoperative MRIs. The training was carried out first on BraTS2021 dataset for preoperative segmentation. Performance is evaluated using DICE score (DS) and Hausdorff 95% (H95).

RESULTS

In preoperative scenario, overall DS is 91.09 (± 0.60) and H95 is 8.35 (± 1.12), considering tumor core, enhancing tumor and whole tumor (ET and edema). In postoperative context, overall DS is 72.31 (± 2.88) and H95 is 23.43 (± 7.24), considering resection cavity (RC), gross tumor volume (GTV) and whole tumor (WT). Remarkably, the RC segmentation obtained a mean DS of 63.52 (± 8.90) in postoperative MRIs.

CONCLUSIONS

The performances achieved by the algorithm are consistent with previous literature for both pre-operative and post-operative glioblastoma's MRI evaluation. Through the proposed algorithm, it is possible to reduce the impact of low-quality images and missing sequences.

Effects of MRI scanner manufacturers in classification tasks with deep learning models

Deep learning has become a leading subset of machine learning and has been successfully employed in diverse areas, ranging from natural language processing to medical image analysis. In medical imaging, researchers have progressively turned towards multi-center neuroimaging studies to address complex questions in neuroscience, leveraging larger sample sizes and aiming to enhance the accuracy of deep learning models. However, variations in image pixel/voxel characteristics can arise between centers due to factors including differences in magnetic resonance imaging scanners. Such variations create challenges, particularly inconsistent performance in machine learning-based approaches, often referred to as domain shift, where the trained models fail to achieve satisfactory or improved results when confronted with dissimilar test data. This study analyzes the performance of multiple disease classification tasks using multi-center MRI data obtained from three widely used scanner manufacturers (GE, Philips, and Siemens) across several deep learning-based networks. Furthermore, we investigate the efficacy of mitigating scanner vendor effects using ComBat-based harmonization techniques when applied to multi-center datasets of 3D structural MR images. Our experimental results reveal a substantial decline in classification performance when models trained on one type of scanner manufacturer are tested with data from different manufacturers. Moreover, despite applying ComBat-based harmonization, the harmonized images do not demonstrate any noticeable performance enhancement for disease classification tasks.

The Río Hortega University Hospital Glioblastoma dataset: A comprehensive collection of preoperative, early postoperative and recurrence MRI scans (RHUH-GBM)

Glioblastoma, a highly aggressive primary brain tumor, is associated with poor patient outcomes. Although magnetic resonance imaging (MRI) plays a critical role in diagnosing, characterizing, and forecasting glioblastoma progression, public MRI repositories present significant drawbacks, including insufficient postoperative and follow-up studies as well as expert tumor segmentations. To address these issues, we present the "Río Hortega University Hospital Glioblastoma Dataset (RHUH-GBM)," a collection of multiparametric MRI images, volumetric assessments, molecular data, and survival details for glioblastoma patients who underwent total or near-total enhancing tumor resection. The dataset features expert-corrected segmentations of tumor subregions, offering valuable ground truth data for developing algorithms for postoperative and follow-up MRI scans.

Subregional analysis of striatum iron in Parkinson's disease and rapid eye movement sleep behaviour disorder

The loss of dopamine in the striatum underlies motor symptoms of Parkinson's disease (PD). Rapid eye movement sleep behaviour disorder (RBD) is considered prodromal PD and has shown similar neural changes in the striatum. Alterations in brain iron suggest neurodegeneration; however, the literature on striatal iron has been inconsistent in PD and scant in RBD. Toward clarifying pathophysiological changes in PD and RBD, and uncovering possible biomarkers, we imaged 26 early-stage PD patients, 16 RBD patients, and 39 age-matched healthy controls with 3 T MRI. We compared mean susceptibility using quantitative susceptibility mapping (QSM) in the standard striatum (caudate, putamen, and nucleus accumbens) and tractography-parcellated striatum. Diffusion MRI permitted parcellation of the striatum into seven subregions based on the cortical areas of maximal connectivity from the Tziortzi atlas. No significant differences in mean susceptibility were found in the standard striatum anatomy. For the parcellated striatum, the caudal motor subregion, the most affected region in PD, showed lower iron levels compared to healthy controls. Receiver operating characteristic curves using mean susceptibility in the caudal motor striatum showed a good diagnostic accuracy of 0.80 when classifying early-stage PD from healthy controls. This study highlights that tractography-based parcellation of the striatum could enhance sensitivity to changes in iron levels, which have not been consistent in the PD literature. The decreased caudal motor striatum iron was sufficiently sensitive to PD, but not RBD. QSM in the striatum could contribute to development of a multivariate or multimodal biomarker of early-stage PD, but further work in larger datasets is needed to confirm its utility in prodromal groups.

Review on deep learning fetal brain segmentation from Magnetic Resonance images

Brain segmentation is often the first and most critical step in quantitative analysis of the brain for many clinical applications, including fetal imaging. Different aspects challenge the segmentation of the fetal brain in magnetic resonance imaging (MRI), such as the non-standard position of the fetus owing to his/her movements during the examination, rapid brain development, and the limited availability of imaging data. In recent years, several segmentation methods have been proposed for automatically partitioning the fetal brain from MR images. These algorithms aim to define regions of interest with different shapes and intensities, encompassing the entire brain, or isolating specific structures. Deep learning techniques, particularly convolutional neural networks (CNNs), have become a state-of-the-art approach in the field because they can provide reliable segmentation results over heterogeneous datasets. Here, we review the deep learning algorithms developed in the field of fetal brain segmentation and categorize them according to their target structures. Finally, we discuss the perceived research gaps in the literature of the fetal domain, suggesting possible future research directions that could impact the management of fetal MR images.

Development and Evaluation of Deep Learning Models for Automated Estimation of Myelin Maturation Using Pediatric Brain MRI Scans

Purpose

To predict the corresponding age of myelin maturation from brain MRI scans in infants and young children by using a deep learning algorithm and to build upon previously published models.

Materials and Methods

Brain MRI scans acquired between January 1, 2011, and March 17, 2021, in our institution in patients aged 0-3 years were retrospectively retrieved from the archive. An ensemble of two-dimensional (2D) and three-dimensional (3D) convolutional neural network models was trained and internally validated in 710 patients to predict myelin maturation age on the basis of radiologist-generated labels. The model ensemble was tested on an internal dataset of 123 patients and two external datasets of 226 (0-25 months of age) and 383 (0-2 months of age) healthy children and infants, respectively. Mean absolute error (MAE) and Pearson correlation coefficients were used to assess model performance.

Results

The 2D, 3D, and 2D-plus-3D ensemble models showed MAE values of 1.43, 2.55, and 1.77 months, respectively, on the internal test set, values of 2.26, 2.27, and 1.22 months on the first external test set, and values of 0.44, 0.27, and 0.31 months on the second external test set. The ensemble model outperformed the previous state-of-the-art model on the same external test set (MAE = 1.22 vs 2.09 months).

Conclusion

The proposed deep learning model accurately predicted myelin maturation age using pediatric brain MRI scans and may help reduce the time needed to complete this task, as well as interobserver variability in radiologist predictions.Keywords: Pediatrics, MR Imaging, CNS, Brain/Brain Stem, Convolutional Neural Network (CNN), Artificial Intelligence, Pediatric Imaging, Myelin Maturation, Brain MRI, Neuroradiology Supplemental material is available for this article. © RSNA, 2023.

Fully Automated Skull Stripping from Brain Magnetic Resonance Images Using Mask RCNN-Based Deep Learning Neural Networks

This research comprises experiments with a deep learning framework for fully automating the skull stripping from brain magnetic resonance (MR) images. Conventional techniques for segmentation have progressed to the extent of Convolutional Neural Networks (CNN). We proposed and experimented with a contemporary variant of the deep learning framework based on mask region convolutional neural network (Mask-RCNN) for all anatomical orientations of brain MR images. We trained the system from scratch to build a model for classification, detection, and segmentation. It is validated by images taken from three different datasets: BrainWeb; NAMIC, and a local hospital. We opted for purposive sampling to select 2000 images of T1 modality from data volumes followed by a multi-stage random sampling technique to segregate the dataset into three batches for training (75%), validation (15%), and testing (10%) respectively. We utilized a robust backbone architecture, namely ResNet-101 and Functional Pyramid Network (FPN), to achieve optimal performance with higher accuracy. We subjected the same data to two traditional methods, namely Brain Extraction Tools (BET) and Brain Surface Extraction (BSE), to compare their performance results. Our proposed method had higher mean average precision (mAP) = 93% and content validity index (CVI) = 0.95%, which were better than comparable methods. We contributed by training Mask-RCNN from scratch for generating reusable learning weights known as transfer learning. We contributed to methodological novelty by applying a pragmatic research lens, and used a mixed method triangulation technique to validate results on all anatomical modalities of brain MR images. Our proposed method improved the accuracy and precision of skull stripping by fully automating it and reducing its processing time and operational cost and reliance on technicians. This research study has also provided grounds for extending the work to the scale of explainable artificial intelligence (XAI).