OpenBHB: a Large-Scale Multi-Site Brain MRI Data-set for Age Prediction and Debiasing

Disentangled latent energy-based style translation: An image-level structural MRI harmonization framework

Brain magnetic resonance imaging (MRI) has been extensively employed across clinical and research fields, but often exhibits sensitivity to site effects arising from non-biological variations such as differences in field strength and scanner vendors. Numerous retrospective MRI harmonization techniques have demonstrated encouraging outcomes in reducing the site effects at image level. However, existing methods generally suffer from high computational requirements and limited generalizability, restricting their applicability to unseen MRIs. In this paper, we design a novel disentangled latent energy-based style translation (DLEST) framework for unpaired image-level MRI harmonization, consisting of (a) site-invariant image generation (SIG), (b) site-specific style translation (SST), and (c) site-specific MRI synthesis (SMS). Specifically, the SIG employs a latent autoencoder to encode MRIs into a low-dimensional latent space and reconstruct MRIs from latent codes. The SST utilizes an energy-based model to comprehend global latent distribution of a target domain and translate source latent codes towards the target domain, while SMS enables MRI synthesis with a target-specific style. By disentangling image generation and style translation in latent space, the DLEST can achieve efficient style translation. Our model was trained on T1-weighted MRIs from a public dataset (with 3,984 subjects across 58 acquisition sites/settings) and validated on an independent dataset (with 9 traveling subjects scanned in 11 sites/settings) in four tasks: histogram and feature visualization, site classification, brain tissue segmentation, and site-specific structural MRI synthesis. Qualitative and quantitative results demonstrate the superiority of our method over several state-of-the-arts.

Disparities in accelerated brain aging in recent-onset and chronic schizophrenia

BACKGROUND

Patients with schizophrenia experience accelerated aging, accompanied by abnormalities in biomarkers such as shorter telomere length. Brain age prediction using neuroimaging data has gained attention in schizophrenia research, with consistently reported increases in brain-predicted age difference (brain-PAD). However, its associations with clinical symptoms and illness duration remain unclear.

METHODS

We developed brain age prediction models using structural magnetic resonance imaging (MRI) data from 10,938 healthy individuals. The models were validated on an independent test dataset comprising 79 healthy controls, 57 patients with recent-onset schizophrenia, and 71 patients with chronic schizophrenia. Group comparisons and the clinical associations of brain-PAD were analyzed using multiple linear regression. SHapley Additive exPlanations (SHAP) values estimated feature contributions to the model, and between-group differences in SHAP values and group-by-SHAP value interactions were also examined.

RESULTS

Patients with recent-onset schizophrenia and chronic schizophrenia exhibited increased brain-PAD values of 1.2 and 0.9 years, respectively. Between-group differences in SHAP values were identified in the right lateral prefrontal area (false discovery rate [FDR] p = 0.022), with group-by-SHAP value interactions observed in the left prefrontal area (FDR p = 0.049). A negative association between brain-PAD and Full-scale Intelligence Quotient scores in chronic schizophrenia was noted, which did not remain significant after correction for multiple comparisons.

CONCLUSIONS

Brain-PAD increases were pronounced in the early phase of schizophrenia. Regional brain abnormalities contributing to brain-PAD likely vary with illness duration. Future longitudinal studies are required to overcome limitations related to sample size, heterogeneity, and the cross-sectional design of this study.

A survey on deep learning in medical image registration: New technologies, uncertainty, evaluation metrics, and beyond

Deep learning technologies have dramatically reshaped the field of medical image registration over the past decade. The initial developments, such as regression-based and U-Net-based networks, established the foundation for deep learning in image registration. Subsequent progress has been made in various aspects of deep learning-based registration, including similarity measures, deformation regularizations, network architectures, and uncertainty estimation. These advancements have not only enriched the field of image registration but have also facilitated its application in a wide range of tasks, including atlas construction, multi-atlas segmentation, motion estimation, and 2D-3D registration. In this paper, we present a comprehensive overview of the most recent advancements in deep learning-based image registration. We begin with a concise introduction to the core concepts of deep learning-based image registration. Then, we delve into innovative network architectures, loss functions specific to registration, and methods for estimating registration uncertainty. Additionally, this paper explores appropriate evaluation metrics for assessing the performance of deep learning models in registration tasks. Finally, we highlight the practical applications of these novel techniques in medical imaging and discuss the future prospects of deep learning-based image registration.

Interpretable and integrative deep learning for discovering brain-behaviour associations

Recent advances highlight the limitations of classification strategies in machine learning that rely on a single data source for understanding, diagnosing and predicting psychiatric syndromes. Moreover, approaches based solely on clinician labels often fail to capture the complexity and variability of these conditions. Recent research underlines the importance of considering multiple dimensions that span across different psychiatric syndromes. These developments have led to more comprehensive approaches to studying psychiatric conditions that incorporate diverse data sources such as imaging, genetics, and symptom reports. Multi-view unsupervised learning frameworks, particularly deep learning models, present promising solutions for integrating and analysing complex datasets. Such models contain generative capabilities which facilitate the exploration of relationships between different data views. In this study, we propose a robust framework for interpreting these models that combines digital avatars with stability selection to assess these relationships. We apply this framework to the Healthy Brain Network cohort which includes clinical behavioural scores and brain imaging features, uncovering a consistent set of brain-behaviour interactions. These associations link cortical measurements obtained from structural MRI with clinical reports evaluating psychiatric symptoms. Our framework effectively identifies relevant and stable associations, even with incomplete datasets, while isolating variability of interest from confounding factors.

A review of artificial intelligence-based brain age estimation and its applications for related diseases

The study of brain age has emerged over the past decade, aiming to estimate a person's age based on brain imaging scans. Ideally, predicted brain age should match chronological age in healthy individuals. However, brain structure and function change in the presence of brain-related diseases. Consequently, brain age also changes in affected individuals, making the brain age gap (BAG)-the difference between brain age and chronological age-a potential biomarker for brain health, early screening, and identifying age-related cognitive decline and disorders. With the recent successes of artificial intelligence in healthcare, it is essential to track the latest advancements and highlight promising directions. This review paper presents recent machine learning techniques used in brain age estimation (BAE) studies. Typically, BAE models involve developing a machine learning regression model to capture age-related variations in brain structure from imaging scans of healthy individuals and automatically predict brain age for new subjects. The process also involves estimating BAG as a measure of brain health. While we discuss recent clinical applications of BAE methods, we also review studies of biological age that can be integrated into BAE research. Finally, we point out the current limitations of BAE's studies.

Explainable brain age prediction: a comparative evaluation of morphometric and deep learning pipelines

Brain age, a biomarker reflecting brain health relative to chronological age, is increasingly used in neuroimaging to detect early signs of neurodegenerative diseases and support personalized treatment plans. Two primary approaches for brain age prediction have emerged: morphometric feature extraction from MRI scans and deep learning (DL) applied to raw MRI data. However, a systematic comparison of these methods regarding performance, interpretability, and clinical utility has been limited. In this study, we present a comparative evaluation of two pipelines: one using morphometric features from FreeSurfer and the other employing 3D convolutional neural networks (CNNs). Using a multisite neuroimaging dataset, we assessed both model performance and the interpretability of predictions through eXplainable Artificial Intelligence (XAI) methods, applying SHAP to the feature-based pipeline and Grad-CAM and DeepSHAP to the CNN-based pipeline. Our results show comparable performance between the two pipelines in Leave-One-Site-Out (LOSO) validation, achieving state-of-the-art performance on the independent test set ( M A E = 3.21 with DNN and morphometric features and M A E = 3.08 with a DenseNet-121 architecture). SHAP provided the most consistent and interpretable results, while DeepSHAP exhibited greater variability. Further work is needed to assess the clinical utility of Grad-CAM. This study addresses a critical gap by systematically comparing the interpretability of multiple XAI methods across distinct brain age prediction pipelines. Our findings underscore the importance of integrating XAI into clinical practice, offering insights into how XAI outputs vary and their potential utility for clinicians.

A perspective on brain-age estimation and its clinical promise

Brain-age estimation has gained increased attention in the neuroscientific community owing to its potential use as a biomarker of brain health. The difference between estimated and chronological age based on neuroimaging data enables a unique perspective on brain development and aging, with multiple open questions still remaining in the brain-age research field. This Perspective presents an overview of current advancements in the field and envisions the future evolution of the brain-age framework before its potential deployment in hospital settings.

A next-generation, histological atlas of the human brain and its application to automated brain MRI segmentation

Magnetic resonance imaging (MRI) is the standard tool to image the human brain in vivo. In this domain, digital brain atlases are essential for subject-specific segmentation of anatomical regions of interest (ROIs) and spatial comparison of neuroanatomy from different subjects in a common coordinate frame. High-resolution, digital atlases derived from histology (e.g., Allen atlas [7], BigBrain [13], Julich [15]), are currently the state of the art and provide exquisite 3D cytoarchitectural maps, but lack probabilistic labels throughout the whole brain. Here we present NextBrain, a next-generation probabilistic atlas of human brain anatomy built from serial 3D histology and corresponding highly granular delineations of five whole brain hemispheres. We developed AI techniques to align and reconstruct ~10,000 histological sections into coherent 3D volumes with joint geometric constraints (no overlap or gaps between sections), as well as to semi-automatically trace the boundaries of 333 distinct anatomical ROIs on all these sections. Comprehensive delineation on multiple cases enabled us to build the first probabilistic histological atlas of the whole human brain. Further, we created a companion Bayesian tool for automated segmentation of the 333 ROIs in any in vivo or ex vivo brain MRI scan using the NextBrain atlas. We showcase two applications of the atlas: automated segmentation of ultra-high-resolution ex vivo MRI and volumetric analysis of Alzheimer's disease and healthy brain ageing based on ~4,000 publicly available in vivo MRI scans. We publicly release: the raw and aligned data (including an online visualisation tool); the probabilistic atlas; the segmentation tool; and ground truth delineations for a 100 μm isotropic ex vivo hemisphere (that we use for quantitative evaluation of our segmentation method in this paper). By enabling researchers worldwide to analyse brain MRI scans at a superior level of granularity without manual effort or highly specific neuroanatomical knowledge, NextBrain holds promise to increase the specificity of MRI findings and ultimately accelerate our quest to understand the human brain in health and disease.

A multimodal neural network with gradient blending improves predictions of survival and metastasis in sarcoma

The objective of this study is to develop a multimodal neural network (MMNN) model that analyzes clinical variables and MRI images of a soft tissue sarcoma (STS) patient, to predict overall survival and risk of distant metastases. We compare the performance of this MMNN to models based on clinical variables alone, radiomics models, and an unimodal neural network. We include patients aged 18 or older with biopsy-proven STS who underwent primary resection between January 1st, 2005, and December 31st, 2020 with complete outcome data and a pre-treatment MRI with both a T1 post-contrast sequence and a T2 fat-sat sequence available. A total of 9380 MRI slices containing sarcomas from 287 patients are available. Our MMNN accepts the entire 3D sarcoma volume from T1 and T2 MRIs and clinical variables. Gradient blending allows the clinical and image sub-networks to optimally converge without overfitting. Heat maps were generated to visualize the salient image features. Our MMNN outperformed all other models in predicting overall survival and the risk of distant metastases. The C-Index of our MMNN for overall survival is 0.77 and the C-Index for risk of distant metastases is 0.70. The provided heat maps demonstrate areas of sarcomas deemed most salient for predictions. Our multimodal neural network with gradient blending improves predictions of overall survival and risk of distant metastases in patients with soft tissue sarcoma. Future work enabling accurate subtype-specific predictions will likely utilize similar end-to-end multimodal neural network architecture and require prospective curation of high-quality data, the inclusion of genomic data, and the involvement of multiple centers through federated learning.

Multivariate brain-cognition associations in euthymic bipolar disorder

BACKGROUND

People with bipolar disorder (BD) tend to show widespread cognitive impairment compared to healthy controls. Impairments in processing speed (PS), attention and executive function (EF) may represent 'core' impairments that have a role in wider cognitive dysfunction. Cognitive impairments appear to relate to structural brain abnormalities in BD, but whether core deficits are related to particular brain regions is unclear and much of the research on brain-cognition associations is limited by univariate analysis and small samples.

METHODS

Euthymic BD patients (n = 56) and matched healthy controls (n = 26) underwent T1-weighted MRI scans and completed neuropsychological tests of PS, attention and EF. We utilised public datasets to develop normative models of cortical thickness (n = 5977) to generate robust estimations of cortical abnormalities in patients. Canonical correlation analysis was used to assess multivariate brain-cognition associations in BD, controlling for age, sex and premorbid IQ.

RESULTS

BD showed impairments on tests of PS, attention and EF, and abnormal cortical thickness in several brain regions compared to healthy controls. Impairments in tests of PS and EF were most strongly associated with cortical thickness in the left inferior temporal, right entorhinal and right temporal pole areas.

CONCLUSION

Impairments in PS, attention and EF can be observed in euthymic BD and may be related to abnormal cortical thickness in temporal regions. Future research should continue to leverage normative modelling and multivariate methods to examine complex brain-cognition associations in BD. Future research may benefit from exploring covariance between traditional brain structural morphological metrics such as cortical thickness, cortical volume and surface area.

BrainAgeNeXt: Advancing Brain Age Modeling for Individuals with Multiple Sclerosis

Aging is associated with structural brain changes, cognitive decline, and neurodegenerative diseases. Brain age, an imaging biomarker sensitive to deviations from healthy aging, offers insights into structural aging variations and is a potential prognostic biomarker in neurodegenerative conditions. This study introduces BrainAgeNeXt, a novel convolutional neural network inspired by the MedNeXt framework, designed to predict brain age from T1-weighted magnetic resonance imaging (MRI) scans. BrainAgeNeXt was trained and validated on 11,574 MRI scans from 33 private and publicly available datasets of healthy volunteers, aged 5 to 95 years, imaged with 3T and 7T MRI. Performance was compared against three state-of-the-art brain age prediction methods. BrainAgeNeXt achieved a mean absolute error (MAE) of 2.78 ± 3.64 years, lower than the compared methods (MAE = 3.55, 3.59, and 4.16 years, respectively). We tested all methods also across different levels of image quality, and BrainAgeNeXt performed well even with motion artifacts and less common 7T MRI data. In three longitudinal multiple sclerosis (MS) cohorts (273 individuals), brain age was, on average, 4.21 ± 6.51 years greater than chronological age. Longitudinal analysis indicated that brain age increased by 1.15 years per chronological year in individuals with MS (95% CI = [1.05, 1.26]). Moreover, in early MS, individuals with worsening disability had a higher annual increase in brain age compared to those with stable clinical assessments (1.24 vs. 0.75, p < 0.01). These findings suggest that brain age is a promising prognostic biomarker for MS progression and potentially a valuable endpoint for clinical trials.

Exploring the potential of representation and transfer learning for anatomical neuroimaging: Application to psychiatry

The perspective of personalized medicine for brain disorders requires efficient learning models for anatomical neuroimaging-based prediction of clinical conditions. There is now a consensus on the benefit of deep learning (DL) in addressing many medical imaging tasks, such as image segmentation. However, for single-subject prediction problems, recent studies yielded contradictory results when comparing DL with Standard Machine Learning (SML) on top of classical feature extraction. Most existing comparative studies were limited in predicting phenotypes of little clinical interest, such as sex and age, and using a single dataset. Moreover, they conducted a limited analysis of the employed image pre-processing and feature selection strategies. This paper extensively compares DL and SML prediction capacity on five multi-site problems, including three increasingly complex clinical applications in psychiatry namely schizophrenia, bipolar disorder, and Autism Spectrum Disorder (ASD) diagnosis. To compensate for the relative scarcity of neuroimaging data on these clinical datasets, we also evaluate three pre-training strategies for transfer learning from brain imaging of the general healthy population: self-supervised learning, generative modeling and supervised learning with age. Overall, we find similar performance between randomly initialized DL and SML for the three clinical tasks and a similar scaling trend for sex prediction. This was replicated on an external dataset. We also show highly correlated discriminative brain regions between DL and linear ML models in all problems. Nonetheless, we demonstrate that self-supervised pre-training on large-scale healthy population imaging datasets (N≈10k), along with Deep Ensemble, allows DL to learn robust and transferable representations to smaller-scale clinical datasets (N≤1k). It largely outperforms SML on 2 out of 3 clinical tasks both in internal and external test sets. These findings suggest that the improvement of DL over SML in anatomical neuroimaging mainly comes from its capacity to learn meaningful and useful abstract representations of the brain anatomy, and it sheds light on the potential of transfer learning for personalized medicine in psychiatry.

Implementation and evaluation of a dynamic contrast-enhanced MR perfusion protocol for glioblastoma using a 0.35 T MRI-Linac system

PURPOSE

MRI-linear accelerator (MRI-Linac) systems allow for daily tracking of MRI changes during radiotherapy (RT). Since one common MRI-Linac operates at 0.35 T, there are efforts towards developing protocols at that field strength. In this study we demonstrate the implementation of a post-contrast 3DT1-weighted (3D-T1w) and dynamic contrast-enhancement (DCE) protocol to assess glioblastoma response to RT using a 0.35 T MRI-Linac.

METHODS AND MATERIALS

The protocol implemented was used to acquire 3D-T1w and DCE data from a flow phantom and two patients with glioblastoma (a responder and a non-responder) who underwent RT on a 0.35 T MRI-Linac. The detection of post-contrast-enhanced volumes was evaluated by comparing the 3DT1w images from the 0.35 T MRI-Linac to images obtained using a 3 T scanner. The DCE data were tested temporally and spatially using data from a flow phantom and patients. Ktrans maps were derived from DCE at three time points (a week before treatment-Pre RT, four weeks through treatment-Mid RT, and three weeks after treatment-Post RT) and were validated with patients' treatment outcomes.

RESULTS

The 3D-T1w contrast-enhancement volumes were visually and volumetrically similar between 0.35 T MRI-Linac and 3 T. DCE images showed temporal stability, and associated Ktrans maps were consistent with patient response to treatment. On average, Ktrans values showed a 54 % decrease and 8.6 % increase for a responder and non-responder respectively when Pre RT and Mid RT images were compared.

CONCLUSION

Our findings support the feasibility of obtaining post-contrast 3D-T1w and DCE data from patients with glioblastoma using a 0.35 T MRI-Linac system.

BASE: Brain Age Standardized Evaluation

Brain age, most commonly inferred from T1-weighted magnetic resonance images (T1w MRI), is a robust biomarker of brain health and related diseases. Superior accuracy in brain age prediction, often falling within a 2-3 year range, is achieved predominantly through deep neural networks. However, comparing study results is difficult due to differences in datasets, evaluation methodologies and metrics. Addressing this, we introduce Brain Age Standardized Evaluation (BASE), which includes (i) a standardized T1w MRI dataset including multi-site, new unseen site, test-retest and longitudinal data, and an associated (ii) evaluation protocol, including repeated model training and upon based comprehensive set of performance metrics measuring accuracy, robustness, reproducibility and consistency aspects of brain age predictions, and (iii) statistical evaluation framework based on linear mixed-effects models for rigorous performance assessment and cross-comparison. To showcase BASE, we comprehensively evaluate four deep learning based brain age models, appraising their performance in scenarios that utilize multi-site, test-retest, unseen site, and longitudinal T1w brain MRI datasets. Ensuring full reproducibility and application in future studies, we have made all associated data information and code publicly accessible at https://github.com/AralRalud/BASE.git.

Structural MRI Harmonization via Disentangled Latent Energy-Based Style Translation

Multi-site brain magnetic resonance imaging (MRI) has been widely used in clinical and research domains, but usually is sensitive to non-biological variations caused by site effects (e.g., field strengths and scanning protocols). Several retrospective data harmonization methods have shown promising results in removing these non-biological variations at feature or whole-image level. Most existing image-level harmonization methods are implemented through generative adversarial networks, which are generally computationally expensive and generalize poorly on independent data. To this end, this paper proposes a disentangled latent energy-based style translation (DLEST) framework for image-level structural MRI harmonization. Specifically, DLEST disentangles site-invariant image generation and site-specific style translation via a latent autoencoder and an energy-based model. The autoencoder learns to encode images into low-dimensional latent space, and generates faithful images from latent codes. The energy-based model is placed in between the encoding and generation steps, facilitating style translation from a source domain to a target domain implicitly. This allows highly generalizable image generation and efficient style translation through the latent space. We train our model on 4,092 T1-weighted MRIs in 3 tasks: histogram comparison, acquisition site classification, and brain tissue segmentation. Qualitative and quantitative results demonstrate the superiority of our approach, which generally outperforms several state-of-the-art methods.