Brain Informatics

INTRODUCTION: The digital clock drawing test (dCDT) is a cognitive screening tool capturing high-resolution pen-movement data. Existing approaches for classifying cognitive impairment from dCDTs rely on selected features or image-based methods. We propose representing dCDTs as structural networks to classify cognitive impairment and characterize spatial organization in clock drawings. METHODS: Weighted networks were constructed from 3,362 Framingham Heart Study participants, with and without cognitive impairment. Network representations were generated separately for command and copy conditions, with nodes representing pre-defined clock components (digits, hands, clock face) and edges capturing pairwise spatial relationships between components. Random forest models and graph convolutional networks were trained to predict cognitive impairment from the networks. Models were benchmarked against demographics-only and established summary feature baselines. RESULTS: Network-based models achieved AUC values comparable to the established summary feature baseline across command-only and copy-only conditions. Differences in spatial organization between the cognitively intact and impaired groups were greatest in the upper-right clock region, involving digits 1 and 2, the clock hands, and the surrounding clock face. Differences between the command and copy networks were larger among cognitively impaired participants. DISCUSSION: The proposed network representations provide an objective, scalable, and flexible framework for representing clock drawings for use in classification models. dCDT networks matched the predictive performance of established summary feature-based approaches and identified clock regions where spatial organization differed by cognitive status. After external validation, this approach has the potential to be incorporated within a clinical workflow to support automated and interpretable cognitive screening.

The choroid plexus (ChP) is a highly vascularized structure located in the brain ventricles that plays important roles in multiple central nervous system functions. ChP enlargement has been implicated in many neuropsychiatric conditions such as multiple sclerosis, depression, schizophrenia, and Alzheimer's disease. However, limitation in the accuracy of the current segmentation tools highlight the need for more reliable methods to quantify ChP volume in healthy and clinical human populations. The shape and size of the ChP vary drastically from person to person, and currently there are no large ChP datasets which allow comprehensive training of segmentation models. Here we introduce DataPlex, a collection of 700 structural MRI brain images where the lateral ventricle ChP has been manually labeled and checked by expert human raters. DataPlex is balanced and representative across a wide range of scanning parameters, demographics, and clinical profiles, and subject-level information will be available for each labeled image. DataPlex will be open source to support the creation of reliable segmentation tools for the human lateral ventricle ChP.

Neuroimaging research focusing on subcortical brain regions showed that they play a role in a wide range of cognitive functions and associated their structural alterations to various psychiatric disorders. Methods to analyse such structures were developed to observe changes, not simply in terms of volumes, but in terms of surfaces, which are able to detect subtle and local changes in shape and geometrical configuration. Such advanced analyses have however been largely restricted to command line environments with relatively limited access to Python users or researchers with low technical expertise. We present SubCortexMesh as a user-friendly toolbox which covers automated surface estimation from popular subcortical volume segmentations (FreeSurfer and the Functional Magnetic Resonance Imaging of the Brain Software Library (FSL)), computes shape-related vertex-wise metrics (thickness, surface area and curvature) for whole cohorts and includes statistical analyses. This is all done inside Python, automatically computing all subjects of a given preprocessed directory, with the explicit intent to minimise steps and manual coding required from the users. SubCortexMesh includes statistical tools which allow conventional random field theory-based cluster analyses on native subject metrics, standardised within common surface templates, thus containing a whole workflow necessary to run an up-to-date shape-wise subcortical analyses in one package.

Neuroimaging studies that pool data across clinical sites often suffer from site effects --- variability in imaging measures that arises from technical heterogeneity across sites as opposed to true biological signal. While numerous harmonization methods have been proposed to remove site effects from imaging features, less attention has been placed on how to incorporate harmonization models into common machine-learning pipelines. We demonstrate that current approaches for integration either suffer from data leakage, fail to fully remove site effects from the target data, or attenuate true biological associations during the harmonization process. To address these issues, we propose Multiple Imputation for Removing Technical Heterogeneity (MIRTH), which uses imputed outcomes to harmonize the test data. We benchmark MIRTH's performance using both simulated and real-world volumetric data from the Alzheimer's Disease Neuroimaging Initiative and the Baltimore Longitudinal Study of Aging, demonstrating that MIRTH can achieve high predictive accuracy while avoiding inflated performance when the outcome is imbalanced across clinical sites.

Background: Ultra-low field (ULF) magnetic resonance imaging (MRI) offers accessible point-of-care neuroimaging, yet its capability to support multivariate network analyses remains understudied. Methods: We systematically evaluated the feasibility and reliability of Morphometric Similarity Networks (MSN) and Morphometric Inverse Divergence (MIND) networks derived from paired high-field (HF, 3T/1.5T) and ULF (64mT) scans in a primary cohort(N=23) and an independent replication cohort (N=19). We tested the impact of super-resolution processing, cortical parcellation granularity (Desikan-Killiany vs. Destrieux), and mesh resolution (fsnative vs. fsaverage5) on edge-level cross-scanner agreement (Lin's concordance correlation coefficient, CCC) and subject-level identifiability. Results: MIND networks demonstrated higher edge-level HF-ULF agreement (mean CCC = 0.95) and ULF reproducibility (mean CCC = 0.98) compared to MSNs (mean CCC = 0.86-0.87 and 0.91-0.94, respectively). In contrast, MSNs yielded higher subject-level differential identifiability across field strengths due to a broader distinction between within- and between-subject similarity. Reconstructing MIND networks on native high-resolution meshes (fsnative) maximised cross-scanner agreement, whereas coarser meshes and finer cortical parcellations enhanced individual network identifiability. Conclusions: Structural similarity networks can be reliably constructed from ULF MRI data. Optimal analytic choices depend on study objectives: MIND on native meshes maximises cross-scanner stability, whereas MSNs excel at preserving individual uniqueness, paving the way for globally scalable connectomic research.

There is increasing demand for quantitative measurements in neuroimaging research, including absolute quantification of regional myelin density. Fast whole-brain macromolecular proton fraction (MPF) imaging enables acquisition and reconstruction of quantitative maps of MPF, a biomarker of myelin, and has been used in many research applications. To date, no standard protocols have been published for parcellating MPF maps. FreeSurfer is a widely used software package for automated brain parcellation. It is designed for use with high-resolution T1-weighted anatomical images, typically acquired with an MPRAGE sequence. MPF maps inherently provide high contrast between white and gray matter and can potentially be used as input images for parcellation, enabling simultaneous quantitative tissue characterization and volumetric assessment. However, it is unclear how the use of MPF maps as input images with FreeSurfer software affect reliability and repeatability of regional MPF and volume estimates. We evaluated three FreeSurfer parcellation workflows for this purpose. The workflows differed in the amount of neuroimaging data required and computational intensity. MPF and MPRAGE data were acquired in two separate sessions for 11 adults. Reliability was evaluated using mean relative differences in estimates, Bland-Altman analysis, and intra-class correlation coefficients (ICCs). Repeatability was assessed using within-subject coefficients of variation (CVws). All three workflows produced similarly high within-subject repeatability across scans. However, reliability in regional MPF estimates was lower in gray matter than in white matter for workflows that used MPF maps as input. Workflows that use MPF maps as FreeSurfer input may be adequate for estimating mean MPF in GM and WM of cortical parcels and in subcortical regions. However, the MPRAGE-based workflow is recommended when reliable individual-level estimates or volume estimates are required.

Quantitative analysis of biological shape is central to understanding how form encodes function. Despite rapid advances in volumetric imaging and segmentation, analysis and of three-dimensional morphology remains challenging due to complex geometries and topologies found in biology. Existing approaches either rely on handcrafted metrics that lack generalizability or employ data-driven neural models that sacrifice interpretability and robustness. Here, we introduce Minkowski Tensors (MT), a family of integral-geometric shape descriptors that extend classical scalar measures of volume, surface area, and curvature into tensor-valued quantities encoding direction and anisotropy. MT form a provably complete yet compact and interpretable feature set that captures volumetric and curvature-based information in a transformation- and scale-covariant, topology-aware manner. We benchmark MT on diverse segmented 3D biomedical images spanning vascular and adrenal structures from magnetic resonance angiography (MRA) and computed tomography (CT), and dividing cell nuclei imaged by confocal microscopy, demonstrating competitive or superior classification performance relative to classical geometric descriptors and neural baselines while retaining direct biological interpretability. Our results establish Minkowski tensors as a concise yet powerful framework for discriminative analysis of complex 3D biological morphology.

Longitudinal clinical cohorts contain heterogeneous progression patterns that are obscured by coarse diagnostic labels. Discovering these latent progression phenotypes requires models that distinguish patients not only by their outcomes, but also by how their clinical trajectories evolve. This paper proposes TRACE, an interpretable temporal phenotyping method that enables the learning of outcome-informed progression phenotypes from longitudinal trajectories. TRACE combines a trajectory- and outcome-aware contrastive objective with a probabilistic predictive phenotype layer, encouraging phenotypes to capture progression structure while remaining predictive of diagnostic state. A dual-attention mechanism further supports interpretation by identifying the visits and clinical variables most associated with each phenotype assignment. TRACE is evaluated on the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort for cognitive disease modeling, showing improved performance in both diagnostic prediction and phenotype discovery compared with other temporal phenotyping baselines. Finally, by allowing the number of phenotypes to exceed the diagnostic labels, TRACE enables discovering additional progression patterns that are not captured by labels defined a priori.

Quantitative magnetic resonance imaging (qMRI) supports both basic neuroscience and clinical research, providing the foundation for longitudinal studies of brain development, aging, and neurological disease. However, its validity rests on a little-studied assumption: that physical relocation of imaging hardware leaves underlying neuroimaging biomarkers intact. When research facilities expand or modernize, an entire scanner undergoes radical disruption--field ramp-down, physical transport, shielding reconstruction, and re-shimming. Whether qMRI metrics remain stable following such invasive procedures without introducing systematic biases is largely unquantified. Here, we systematically evaluated the impact of relocating a 3T clinical scanner over a distance of 30 km on a broad, multimodal suite of neuroimaging markers in healthy volunteers. To disentangle variance components, we benchmarked within- and cross-site stability using percent variation (PV; measurement precision), intraclass correlation coefficients (ICC; absolute reliability), and Spearman rank correlations (R2; spatial and order agreement). We show that scanner relocation introduces no greater variance than baseline within-site assessments. High-resolution brain structural morphometry from different sequences (multi-echo MPRAGE, compressed sensing MPR2RAGE) and quantitative relaxometry (compressed sensing MP2RAGE T1 mapping) demonstrated excellent stability (PV<1%, ICC=0.89-0.99, R2>0.93). Diffusion tensor (mean diffusivity: MD, fractional anisotropy: FA) and mean signal kurtosis (MSK) metrics from multi-shell diffusion MRI maintained low variability across white and deep gray matter (global cross-site ICC>0.90; global PV: MD<1.7%, MSK<1.2%, FA<3.8%; global R2: MSK>0.93; MD>0.90; FA>0.77), confirming relocation robustness for diffusion-derived microstructural markers. Resting-state functional networks displayed high cross-site reproducibility (ICC>0.90), while task-based fMRI activation peaks retained acceptable spatial and magnitude agreement (ICC=0.30-0.60) despite intrinsic physiological variability. In the absence of software or hardware upgrades, physical scanner relocation leaves qMRI biomarkers remarkably intact--providing the empirical foundation required to preserve data integrity across major facility transitions and multi-site longitudinal cohorts. Key points that summarize the paperO_LIHardware transitions do not compromise qMRI data integrity: Physical relocation of a 3T scanner (including field ramp-down, 30 km transport, and re-shimming) introduces no greater variance in quantitative MRI metrics than baseline within-site test-retest assessments. C_LIO_LIMultimodal biomarker robustness: Brain morphometry, T1 relaxometry, diffusion tensor and kurtosis, resting-state functional connectivity display comparable within- and cross-site stability reflected by different metrics (percent variation, intraclass correlation coefficient and Spearman correlation) C_LIO_LIEmpirical backing for longitudinal research: In the absence of hardware or software upgrades, multimodal qMRI biomarkers remain remarkably stable, reassuring researchers that long-term cohort studies and multi-site harmonization can survive major facility relocations. C_LI Graphical Abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=175 SRC="FIGDIR/small/26363342v1_ufig1.gif" ALT="Figure 1"> View larger version (57K): org.highwire.dtl.DTLVardef@773d03org.highwire.dtl.DTLVardef@46a09borg.highwire.dtl.DTLVardef@da5bc0org.highwire.dtl.DTLVardef@1bdbf9d_HPS_FORMAT_FIGEXP M_FIG This study evaluates the impact of physically relocating a 3T MRI scanner on neuroimaging biomarkers. By comparing before and after relocation sessions on healthy volunteers, we conclude that the procedure does not introduce significant variability. Parameters for morphometry, T1 relaxometry, diffusion, and resting-state functional connectivity show high stability. Task-based fMRI showed greater within-site variability, but no increase in variability following relocation, consistent with the inherently higher variability of task-evoked responses.. The results confirm the integrity of qMRI data after infrastructure relocation, ensuring reliability for longitudinal studies. C_FIG

Coarse to fine segmentation is a natural strategy for three dimensional lesion delineation becausefull resolution processing of an entire scan is expensive while only a small fraction of the volume contains pathology. Its main failure mode, however, is structural: a lesion missed by the coarse stage is usually absent from the region of interest and therefore cannot be recovered by the fine stage. This problem is most severe for small, low contrast, and multifocal lesions. We introduce RECALLREFINE, a coverage constrained coarse to fine framework that treats candidate-region selection as a recall critical learning problem rather than a deterministic crop around the coarse mask.A full-volume network first predicts coarse lesion probabilities and multi-scale features. A candidate score network then combines coarse probability, predictive uncertainty, feature novelty, and spatial context to score overlapping 3D blocks. During training, a differentiable selector is optimized with a component-wise coverage objective that requires every annotated lesion instance to remain covered by at least one selected block under a fixed refinement budget. Selected blocks are processed at native resolution by a local refiner with a context-ring branch, and overlapping predictions are merged back into the full volume. Across four public 3D lesion benchmarks, RECALLREFINE improves mean tumor Dice from 76.5% for the strongest matched coarse-to-fine baseline to 78.4%, while increasing sensitivity for lesions below 10 mm from 60.9% to 70.8%.At a 20% candidate budget, the selector covers 94.4% of lesion instances, substantially reducing irreversible coarse-stage misses. The results suggest that the central problem in cascaded lesion segmentation is not only how to refine a detected lesion, but how to guarantee that difficult lesions are still represented in the refinement set.

Three-dimensional imaging is rapidly increasing the scale of biological morphometric datasets, but establishing automated anatomical correspondence remains a major bottleneck. Conventional point-cloud registration methods transfer landmarks efficiently, yet poor initialization or high morphological variation can produce geometrically plausible but biologically implausible deformations. Here, we developed a statistical-shape-model (SSM)-regularized coherent point drift (CPD) estimator for automated morphometric landmark transfer. It replaces the unconstrained non-rigid CPD update with optimization in SSM coefficient space and regularizes coefficients by inverse population covariance, so deformations along low-variance biological modes are penalized more strongly. We compared pose-matched conventional CPD, SSM-CPD, and SSM-CPD with refinement in leave-one-out atlas experiments on laboratory mouse skulls (N=249), primate molars (N=115), and squamate vertebrae (N=254) under aligned, mildly perturbed, and severely perturbed initialization. We evaluated these formulations in terms of landmark accuracy, failure-tail behavior, runtime, and preservation of morphometric structure. Under severe perturbation, we also evaluated an integrated recovery procedure that optimizes the starting template, recovers global pose, and then applies SSM-CPD with refinement. Severe misalignment produced widespread failures in pose-matched conventional CPD, whereas SSM regularization reduced median error from 13 - 21% to approximately 1 - 3% of the target bounding-box diagonal. This improvement was highly consistent for skulls and molars. In the morphologically complex vertebral dataset, however, low median error concealed a substantial minority of catastrophic failures (28% of vertebrae above a 6% error threshold), which integrated recovery reduced to 4%. Across datasets, automated landmarks preserved the major axes and pairwise relationships of morphometric shape space, and local CPD refinement improved covariance preservation. Embedding learned biological variation directly into registration broadens the range of starting conditions and anatomies for which automated landmark transfer remains reliable. Our results also show that failure tails and downstream morphometric fidelity must be assessed alongside median accuracy when automated landmarking methods are evaluated.

Background: Software tools can support MRI diagnosis of focal cortical dysplasias (FCDs) - subtle, surgically-treatable lesions that underlie drug-resistant focal epilepsy (DRFE). MELD-graph, a widely used open-source tool combining surface-based morphometry with a convolutional neural network, was developed using 1.5T and 3T MRI data. 7T MRI is increasingly used for lesion detection in DRFE. This study investigated whether MELD-graph is robust to domain shift at 7T. Methods: We collated paired 3T and 7T T1-weighted MRI, and FLAIR where available, from 23 patients with FCD (6 histologically-verified, 17 identified radiologically) and 39 healthy controls at four epilepsy centres. Manual lesion labels were used as ground truth. MELD-graph was run on 3T images, on 7T images with the default pipeline, and on 7T images with adaptations for 7T MRI (SPM bias-field correction, SynthStrip skull-stripping, sub-millimetric FreeSurfer reconstruction), with and without ComBat harmonisation. FCD detection performance and surface morphometry features were compared. Results: Sensitivity for FCD detection was 65% (15/23) at 3T, 48% (11/23) for 7T default and 52% (12/23) for 7T adapted, with Combat harmonisation. Among radiologically 3T-negative patients, 2/8 were detected at 3T and 3/8 at 7T; and among histologically confirmed FCD, 3/6 at 3T and 1/6 at 7T. Specificity was significantly lower at 7T (default 39%, adapted 45%) than 3T (76%). False-positives predominantly occurred in the basal temporal lobes. ComBat harmonisation improved specificity in all conditions and largely compensated for field-strength dependent differences in surface morphometry features. Adding FLAIR at 7T increased false-positive predictions. Conclusion: Sensitivity is comparable but specificity of MELD-graph is lower at 7T compared to 3T. Adaptations for 7T MRI can reduce segmentation errors but do not significantly improve FCD detection, while harmonisation improves performance. The adapted pipeline is released open-source. 7T MELD-graph outputs may inform radiologist review in challenging cases but require cautious interpretation.

An Ischemic stroke is one of the main causes of death and long-term disability around the world. There are very few treatment options available, especially for the cases where the short time period to give clot dissolving medicine has passed (Powers et al., 2019). While single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool for understanding cell-to-cell differences within at issue (cellular heterogeneity) (Tang et al., 2009). However translating these high dimensional datasets into clinically useful biomarkers remains challenging (Rajko-mar et al., 2019). A major problem with current biomarker studies is their reliance on a single machine learning model, where results may reflect model bias rather than true biological signals (Lopez-Rincon et al., 2020). Methods: We developed a multi-model consensus framework integrating eight different machine learning methods from three groups: traditional models (logistic regression, random forest, XGBoost), a hybrid deep learning model (CNN-XGBoost), and transformer-inspired models (scGPT, scBERT, sc-Former, Geneformer). The proposed framework was applied to scRNA-seq data from a mouse middle cerebral artery occlusion (MCAO) model (GSE174574; n=6 mice, 54,599 cells). To prevent data leakage, we implemented rigorous subject-wise train/test splitting, ensuring all cells from individual animals remained either in the training sets or in the testsets. Results: All eight models achieved excellent performance on the independent test data, with AUC scores between 0.9778 and 0.9975. From six models that allowed gene level importance extraction, we identified 63 consensus genes (threshold: [&ge;] 4/6 models), with 17 genes selected by all 6 models. Pathway enrichment analysis showed significant enrichment of interferon signaling, with a novel dual-interferon pattern. Twelve of the consensus genes respond to interferon, and five of them (Ifitm6, Ifitm1, Gbp2, Igtp, Tgtp1) respond to both Type I (IFN-/{beta}) and Type II (IFN-{gamma}) interferons. Notably the consensus included Il1rn (interleukin-1 receptor antagonist), the target of Anakinra -- an FDA-approved drug that has completed Phase II stroke trials with promising safety profiles (Smith et al., 2018; Emsley et al., 2005). Conclusions: The Multi-model consensus framework, proposed in this paper, provides a robust and reliable approach to biomarker discovery that goes beyond the limitations of any single method and reduces individual algorithm bias. The convergence of classical, ensemble, and deep learning methods on interferon and IL-1 pathway genes supports their biological validity as stroke biomarkers. Our findings provide a prioritized gene list for further clinical validation and highlight interferon signaling as a possible treatment strategy warranting further investigation. Keywords: ischemic stroke; machine learning; single-cell RNA sequencing; biomarker discovery; consensus analysis; interferon signaling; neuroinflammation

Text-driven medical image segmentation demands unified models based on multi-modal learning, capable of processing diverse imaging modalities while providing spatially reliable prediction estimates for clinical practice. Nevertheless, existing vision-language model solutions suffer from poor domain generalization and exhibit limited capability for uncertainty estimation to model the cumulative unreliability of cross-modal alignment across network layers. In this paper, we propose UniMedSeg, a unified vision-language framework for multi-modal medical image segmentation built upon frozen pre-trained vision-language backbones. Our framework incorporates modality-aware adaptation, cross-layer uncertainty-guided encoding, and a frequency-spatial dual-prompt decoder, wherein the progressive uncertainty propagation mechanism of the encoder suppresses misleading cross-modal features, whereas the decoder disentangles anatomical structure and boundary details in the frequency domain under dual textual prompts. Auxiliary losses further sharpen boundary representations and calibrate pixel-wise uncertainty maps. Extensive experiments demonstrate that UniMedSeg preserves competitive in-domain segmentation performance, improves domain generalization, and yields clinically useful spatial risk estimation.

Lesion Network Mapping (LNM) is a method used to map symptom-related circuits in the human brain. Recent work has raised concerns about the method and the specificity of LNM-derived networks, which often show high spatial overlap across unrelated conditions. Null models have been proposed to address the raised limitations1. Useful statistical tools in network neuroscience, they are reference procedures; in LNM, the normative connectome matrix C remains the source of the derived networks also under null modelling. Their broader role as a general solution for LNM therefore warrants careful evaluation. Here, we examine the proposed null models step-by-step and test them empirically across multiple datasets. We show that the proposed null-model correction effectively only replaces C with a transformed matrix C' (e.g., C' ~ C - mean(C)), one that is directly derived from C and retains much of its low-dimensional structure. Consequently, null-model-corrected LNM maps remain shaped by the same standard set of connectome-derived patterns and continue to show repetition and non-specificity across unrelated conditions. We discuss the scope and limitations of null models for LNM, showing that they can statistically benchmark LNMs, but that their application does not resolve the core limitation of non-specificity of LNM networks.

Background: Traditional neuropathological assessment of Alzheimer's disease (AD) relies on manual, semi-quantitative scoring of amyloid-beta (A{beta}) burden. This approach is limited by high inter-rater variability and often qualitative, failing to capture subtle, continuous pathology changes. While 11C-PiB PET neuroimaging provides non-invasive longitudinal monitoring, it has a limited dynamic range and lacks the sub-millimeter spatial resolution necessary to track localized microvascular and parenchymal A{beta} trajectories. Methods: We introduce ViTamyloid, a generalizable modular computational pathology pipeline for high-throughput A{beta} segmentation and classification in postmortem tissue. In Stage 1, a pathologist-supervised hybrid framework combining an EfficientNet-B4 encoder with a Residual U-Net decoder, complement with a localization-aware Faster R-CNN detector, to generate binary segmentation masks and isolated lesion candidates. In Stage 2, classification model powered by a frozen, high-resolution UNI2-h computational pathology foundation backbone and a ViT-Adapter, classifies candidates into a fine-grained four-class taxonomy: classic cored plaques (CCP), coarse-grained plaques (CGP), capillary cerebral amyloid angiopathy (CAA Type 1), and parenchymal/leptomeningeal CAA (CAA Type 2). We bench-marked out-of-distribution (OOD) robustness on external datasets from different brain banks, and validated clinical scalability across 1,646 whole-slide images from three brain banks. Results: The ViTamyloid classifier achieved strong feature-separation metrics, yielding area under the receiver operating characteristic curve (AUROC) and precision-recall curve (AUPRC) values of 0.994/0.958 for CCP, 0.989/0.945 for CGP, 0.992/0.977 for CAA Type 1, and 0.997/0.978 for CAA Type 2. More importantly, ViTamyloid consistently outperforms other pipelines' models and frozen features in all external datasets with different antibodies, and scanner settings, and image quality. Conclusion: By replacing empirical, from-scratch architectures with a generalizable foundation model adapter framework, ViTamyloid provides a highly sensitive scalable digital pathomics toolkit. This pipeline expose continuous microscopic neuropathological dynamics beyond the spatial and sensitivity plateaus of in vivo neuroimaging, offering objective, quantitative endpoints for Alzheimer's disease clinical trials and disease subtyping.

Population-graph models aggregate a query subject with an unlabeled cohort, so their reported performance can depend on the evaluation regime rather than on predictive information alone. Evaluation context has two separable axes: the visibility of the cohort during training, and the context available at inference. We study the training-time axis under a protocol fixed before analysis, on an ABIDE-I cohort of 871 subjects (20 acquisition sites). Two population-graph implementations are compared: a learned/adaptive graph (EV-GCN) and a static weighted graph (Parisot-GCN). Within each implementation the cohort, connectivity features, outer splits and training and selection rules are held fixed, and only the information available to the graph varies. The primary estimand is the implementation-by-training-context interaction: the difference between the P0-to-P1 AUC changes of the two implementations ({Delta}{Delta}AUC). Across three preselected outer splits (666, 1024, 2024) the direction is consistent and positive, with observed interaction range 0.06285-0.13090; uncertainty is assessed separately within each split by paired subject and site-cluster bootstrap resampling and is reported per split. Splits are replication units: they are not pooled and no cross-split average is reported, so the result is not a claim of invariance to arbitrary splits. These findings tie reported population-graph performance to the evaluation information regime and motivate explicit reporting of cohort visibility. The contribution is methodological: a two-axis taxonomy of information regimes, preselected cross-split evidence that the sensitivity of reported performance to cohort context depends on the implementation, and a reusable reporting checklist.

Foundation models (FMs) have rapidly become dominant in artificial intelligence and are increasingly being adopted in computational pathology. Numerous pathology-specific FMs have been developed and evaluated for a variety of downstream tasks, most commonly using frozen image embeddings with linear probes. While a pathology FM, at least implicitly suggests broad reusability across tasks, the transferability of these models to neurodegenerative disease-related tasks remains largely unexplored. In this study, we evaluated fourteen frozen feature extractors, including general-vision and pathology FMs, across four pathology image classification datasets spanning two specialized neurodegenerative disease immuno-histochemistry (IHC) tasks (Tau neurofibrillary tangle (NFT) and amyloid-{beta} plaque classification) and two cancer-related hematoxylin and eosin (H&E) tasks (the breast tissue BACH dataset and the multi-tissue TIL dataset). The cancer-related H&E datasets represent tasks more closely aligned with the predominant pretraining domain of many current pathology FMs, whereas the neurodegenerative IHC datasets represent a more specialized domain with limited representation in existing FM pretraining cohorts. Frozen embedding linear probes were compared against a conventional supervised ResNet-50 convolutional neural network (CNN) trained directly on image tiles using identical train, validation, and test splits. Across both neurodegenerative IHC datasets, the supervised CNN substantially outperformed all frozen FM linear probes. In contrast, pathology FMs achieved performance comparable to, and in some cases exceeding, the supervised CNN across the two cancer-related H&E datasets, with UNI2-h achieving the highest performance on BACH and several pathology FMs performing on par with the CNN on the larger TIL dataset. Furthermore, a supervised CNN trained using only 1% of the Tau NFT training data (1,938 tiles) still exceeded the performance of the best frozen FM linear probe trained on the complete dataset. Together, these results suggest the transferability of frozen pathology FMs may depend strongly on how well the downstream task is represented by their pretraining domain. These findings demonstrate frozen pathology FMs transfer effectively to the cancer-related H&E tasks evaluated here but may be less effective for specialized neurodegenerative IHC tasks less well represented in current FM pretraining cohorts. Expanding pathology FM pretraining datasets to include a broader range of disease domains and staining modalities may therefore improve transferability to specialized pathology applications. Our findings also highlight the continued importance of conventional supervised learning and motivate future work investigating nonlinear probes and end-to-end FM fine-tuning.

Three-dimensional medical image segmentation is unusually heterogeneous in its computational needs: large homogeneous organ interiors are often resolved at coarse scale, whereas small organs, lesions, thin vessels, and ambiguous boundaries benefit from native-resolution processing. Yet modern 3D segmentation networks allocate nearly uniform computation across the entire volume. Existing dynamic-resolution or coarse-to fine methods reduce cost, but typically route regions using hand-crafted image complexity, prediction confidence, or a fixed crop, none of which directly estimates whether spending additional computation on a region will improve the segmentation. We introduce ERRORROUTE, a budgeted inference framework that learns the marginal refinement utility of candidate 3D blocks. A coarse branch first predicts global low-resolution semantics. During training, we estimate for each candidate block the loss reduction achievable by native resolution refinement and supervise a utility estimator to predict this gain before refinement. A differentiable budgeted selector then allocates a fixed compute budget to blocks with the highest expected utility per unit cost, while unselected regions remain on the efficient global path. Selected blocks are refined locally and merged through residual cross-scale fusion. Across six 3D CT/MRI benchmarks, ERRORROUTE achieves a mean Dice of 90.5% at 128 GFLOPs, improving the accuracy-compute frontier over dense 3D baselines and fixed coarse-to-fine systems. At a 30% routing budget, the learned utility estimator recovers 83% of the blocks selected by an oracle router, with the largest allocation concentrated on lesions, small structures, and complex boundaries. These results suggest that efficient medical segmentation benefits from learning where additional computation is useful, rather than merely where the model is uncertain.

Background and purpose: Accurate identification of cardiac MRI volume orientations is essential for reliable image interpretation and for enabling downstream automated analysis pipelines. However, orientation labels and commonly assigned manually, making the process time-consuming and prone to variability. Recent advances in vision-language models offers new opportunities for automated label generation by jointly modeling visual and semantic information. Methods: In this study, we systematically compare a conventional end-to-end transformer-based captioning model with a CLIPCap architecture for automated cardiac MRI view labelling. Both models were evaluated on a dataset of 1184 de-identified cardiac MR images spanning across six clinically relevant orientations: short axis, two-, three-, four-chamber views, left and right ventricular outflow tract views. The caption transformer was trained from scratch, whereas CLIPCap leveraged pretrained CLIP image embeddings combined with lightweight mapping network and a frozen language model. Results: Experimental results demonstrate a substantial performance advantage from CLIPCap, which achieved an overall accuracy of 98% compared to 68% for the caption transformer. CLIPCap consistently delivered high precision and recall across all orientation classes, including those of high clinical importance, while the end to end transformer exhibited unstable performance and complete failure in certain views. Conclusion: These findings highlight the benefits of leveraging large scale multimodal pretrained model for medical image labeling tasks, particularly in data limited settings. The study suggests that CLIPCap provides a more robust and clinically reliable solution for automated cardiac MRI orientation labeling, supporting its integration into efficient and scalable clinical workflows.