Human Brain Mapping

Analytical variability in neuroimaging pipelines contributes to concerns about reproducibility in the field. In structural MRI, different segmentation tools produce discrepant morphometric estimates that may influence downstream analyses, such as predictive modeling. We tested whether integrating several segmentation pipelines improves brain age prediction, and characterized the spatial and demographic structure of pipeline differences across several datasets. T1-weighted scans from five open-access datasets were processed with four widely-used structural segmentation pipelines. Brain age models were trained using single-pipeline features and compared with multi-pipeline aggregation strategies. Inter-pipeline variability was assessed across shared subcortical structures and examined in relation to age and sex. Integrating features across distinct segmentation frameworks improved predictive performance relative to individual pipelines, whereas aggregation within closely related software versions provided limited benefit. Variability was spatially structured and volumetric measures were often systematically associated with age and sex. These results suggest that segmentation differences reflect structured, demographically sensitive variation rather than random noise, and that multi-pipeline feature integration can enhance robustness in neuroimaging-based prediction.

Compared to traditional gross volumetrics, surface- based models provide greater spatial precision for understanding brain alterations related to developmental, neurological, and psychiatric disorders. Large-scale brain initiatives are combining data from around the world to discover and improve illness- related brain markers. Here, we present a toolkit for 3D brain geometry analysis aimed at addressing key challenges facing large- scale neuroimaging studies. Our framework incorporates scalable methods for multisite data integration, site-specific confound correction, accelerated statistical modeling, interpretable machine learning, and interactive results visualization. The toolkit was tested on data from 21 independently collected study samples participating in the ENIGMA Bipolar Disorder Working Group (N = 3,373). Compared to traditional volume features, we show how subcortical shape measures can be combined across study sites to capture spatially complex differences between diagnostic groups and associations with common treatments. Statistical modeling was accelerated using the Fast and Efficient Mixed- Effects Algorithm (FEMA) and achieved a 16-fold reduction in computation time compared to traditional approaches. Machine learning models showed shape features may provide greater predictive performance over traditional volumes for both diagnostic and treatment prediction tasks, with interpretable weight maps providing insights into the local features driving model performance.

Objective: Brain network controllability provides a framework for understanding how structural organization shapes brain dynamics, yet current models mainly rely on white-matter connectivity and may overlook the contribution of gray-matter architecture. Approach: We constructed a fusion network combining diffusion tensor imaging-derived white-matter connectivity with gray-matter morphological similarity and investigated its controllability, biological associations, heritability, phenotype prediction, and control energy. Main results: Controllability derived from the fusion network preserved key topological properties of the white-matter network and was associated with neurotransmitter systems and cerebral metabolism. Compared with the white-matter connectivity-based network, fusion-based controllability showed a systematic shift toward higher heritability, improved prediction of several individual characteristics and cognitive functions, and lower modeled control energy for activating resting-state networks. Significance: These findings suggest that incorporating gray-matter morphological information into a DTI-supported network provides a complementary structural representation for studying brain network controllability and state transitions. The lower control energy represents a model-derived transition cost and should not be interpreted as a direct measure of physiological energy expenditure.

BACKGROUND AND PURPOSENeighborhood-level socioeconomic disadvantage is associated with adverse brain morphometry, yet whether these associations differ by biological sex remain opaque. Here, we investigated sex-specific associations between the area deprivation index and brain morphometry derived from routine clinical MRI in a real-world clinical population. MATERIALS AND METHODSIntracranial volume-normalized regional brain volumes were extracted from T1-weighted MRI examinations performed in 2,863 consecutive clinical patients (median age 54 years [IQR 38-68]; 61.2% female) at a single academic medical center and associated community partners using an automated atlas-based segmentation pipeline. Exploratory factor analysis was applied to 131 regional brain volumes to identify latent neuroanatomical morphometric networks. Sex-stratified linear regression models examined associations between area deprivation index national percentile rank and each factor score, adjusting for age, with correction for multiple comparisons. RESULTSFactor analysis identified five neuroanatomical morphometric networks: cerebellar (ML1), frontal/executive (ML2), subcortical-ventricular (ML3), medial temporal/limbic (ML4), and posterior cortical/visual (ML5). In male patients (n = 1,112), linear regressions revealed that greater neighborhood-level socioeconomic disadvantage was significantly associated with lower factor scores on the cerebellar ({beta} = -0.006, 95% CI [-0.009, -0.003], P < .001), medial temporal/limbic ({beta} = -0.004, 95% CI [-0.007, -0.001], P = .01), and frontal/executive ({beta} = -0.004, 95% CI [-0.007, -0.0004], P = .04) networks. No significant associations were observed in female patients (all Ps [&ge;] .61). CONCLUSIONSIn a real-world clinical population, neighborhood-level socioeconomic disadvantage was associated with lower regional brain volumes across cerebellar, frontal/executive, and medial temporal/limbic neuroanatomical morphometric networks in male but not female patients. These findings suggest that the neuroanatomical correlates of neighborhood disadvantage may be sex-specific, and that sex-stratified analyses may be necessary to fully characterize the relationship between the social exposome and brain morphometry in clinical neuroimaging research.

Laser Interstitial Thermal Therapy (LITT) is a minimally invasive neurosurgical technique in which a stereotactically-implanted fiber delivers thermal energy to ablate intracranial lesions. Existing computer-assisted planning systems optimize trajectories against a one-dimensional line abstraction, then approximate the ablation zone as a fixed-radius cylinder post-hoc to estimate coverage. Trajectories selected as optimal under this model are not guaranteed to remain optimal once the cylindrical extent is applied, which introduces a mismatch between predicted and true ablation coverage. This may also underestimate spillover into surrounding healthy tissue. We present OptiLITT, a treatment planning system that represents the laser probe as a cylindrical ablation volume from the onset of optimization, jointly solving dual-fiber placement, lesion coverage, and healthy-tissue spillover as a single coupled problem. All planning parameters are exposed through a user-configurable graphical user interface supporting intraoperative refinement between planning stages.

Objective: Accurate volumetric analysis of the brain and cerebrospinal fluid (CSF) is essential for monitoring hydrocephalus, a significant pediatric neurological condition. While computed tomography (CT) provides high-quality volumetric assessment, its associated ionizing radiation poses risks, especially for children. Low-field magnetic resonance imaging (LF-MRI) offers a safer and more accessible alternative, particularly in resource-constrained settings. However, its lower resolution and increased susceptibility to structural distortions make accurate segmentation challenging. This study aims to demonstrate that reliable volumetric measurements can be obtained from LF-MRI that are comparable to CT, enabling safer and more frequent monitoring of infants with hydrocephalus. Approach: We propose EnSegNet-Cross, a cross-modality, enhancement-aware segmentation network for brain volume analysis using LF-MRI. The framework leverages high-fidelity CT data during training but requires only LF-MRI during inference. At the core of the framework is a novel cross-modal topological penalty designed to minimize discrepancies between predicted LF-MRI and CT structures. A central contribution is the integration of a three-dimensional topological loss based on persistent homology, which penalizes topological discrepancies in CSF regions, specifically CSF holes formed by enclosed brain parenchyma, between CT and LF-MRI segmentations. Incorporating these structural priors facilitates generalization across heterogeneous clinical cases while eliminating the need for CT data during inference, resulting in more anatomically coherent and topologically faithful segmentations. Main Results: On a curated cohort of infants with hydrocephalus who had paired LF-MRI and CT scans, including cases with infectious and non-infectious causes, EnSegNet-Cross consistently outperformed state-of-the-art machine learning alternatives. It achieved the highest Dice score of 0.8532 plus/minus 0.03 and Volume Score of 0.9318 plus/minus 0.03. The method also demonstrated robust performance in challenging cases with confounding factors, achieving a Dice score of 0.8340 plus/minus 0.03 and a Volume Score of 0.9111 plus/minus 0.05. By leveraging CT-derived topological priors, EnSegNet-Cross successfully handled anatomically complex scenarios in which conventional models failed. Significance: EnSegNet-Cross provides a reliable and interpretable solution for brain and CSF segmentation, particularly in complex cases of hydrocephalus. This study demonstrates that high-fidelity volumetric estimates can be achieved using only LF-MRI, facilitating frequent, radiation-free monitoring. By bridging the fidelity gap between low-quality LF-MRI and high-resolution CT through clinically grounded enhancement and topological supervision, EnSegNet-Cross offers a robust clinical tool for brain volumetric analysis in infants with hydrocephalus using LF-MRI.

Resting-state functional MRI (rs-fMRI) studies in schizophrenia commonly rely on normalization to volumetric templates and fixed atlas parcellations derived from neurotypical populations. While these approaches enable group-level comparisons, they may obscure individual variation in cortical organization and intrinsic brain dynamics. In this study, we compared four preprocessing and parcellation strategies across two independent schizophrenia cohorts (MRI site 1, n=159; MRI site 2, n=255) to evaluate how analytic choices affect static functional connectivity and dynamic quasi-periodic pattern (QPP) measures, including default mode-dorsal attention network opposition, QPP component rank, explained variance, event rate, and associations with PANSS symptom severity. Across datasets, individualized surface-based parcellation (IndiPar) consistently detected more pronounced QPP dynamics, stronger default mode / dorsal attention network opposition, and greater explained variance of QPPs relative to atlas-based pipelines. IndiPar also produced larger and more reproducible patient-control differences in functional connectivity and QPP event-rate measures, suggesting improved sensitivity through preservation of subject-specific organization. IndiPar additionally detected a significantly increased QPP event rate and more symptom associations in patients in the larger dataset. However, associations between fMRI measures and symptom severity showed limited stability across cohorts. These findings extend previous reports of altered resting-state activity in schizophrenia, and demonstrate that preprocessing and parcellation choices substantially influence both static and dynamic rs-fMRI results. Individualized surface-based parcellation appears to better preserve subject-specific variability and improves detection of intrinsic brain dynamics. At the same time, the limited cross-dataset replication of symptom associations highlights the challenges of deriving stable brain-symptom relationships from heterogeneous psychiatric cohorts.

Structural MRI is routinely acquired in clinical practice, yet quantitative morphometry has had limited impact on clinical decision-making. Overlapping symptoms, trajectories and comorbidities remain difficult to interpret within disease-specific frameworks, leaving it unclear how individual patients relate to the broader organization of brain disease. Here, we construct a cross-disease morphological reference space from 110,591 T1w MRI scans of 78,794 participants, spanning four disease families, 19 diagnoses, and seven subtypes. To construct this space, we developed NeuroMorph, an AI framework deriving thirteen interpretable morphological descriptors and individual normative deviation profiles. The reference space reveals shared and distinct morphological signatures that distinguish conditions within a hierarchy of disease families, diagnoses, subtypes and individual profiles. It identifies overlapping and comorbid morphological profiles and captures longitudinal deviations that precede clinical diagnosis and track progression. Together, these findings establish a unified framework for mapping brain disease organization and positioning individual patients within its morphological landscape.

The asynchronous development of structural and functional brain networks in early childhood remains largely unexamined, primarily due to the scarcity of longitudinal neuroimaging data. Resolving this temporal dimension is critical, as it promises to reshape our understanding of structural-functional (S-F) coupling, revealing not only whether brain architecture supports function, but also when and over what timescale its influence emerges. However, while rapid neonatal brain maturation and logistical constraints continue to hinder longitudinal data collection, large-scale cross-sectional multimodal datasets are currently available to bridge this gap. Here, we propose a longitudinal subject-pairing framework that reconstructs developmental trajectories from cross-sectional data. It pairs the infants with a predefined age gap while maximizing their similarity in both structural and functional features, thereby approximating longitudinal trajectory of functional changes in relation to the structural maturation. As a case study, we applied this framework to the perisylvian region in a subset of 505 neonates from the multimodal dHCP brain dataset. The myelination index was derived as a structural feature from MRI, and the fractional amplitude of low-frequency fluctuations (fALFF) was derived as a functional feature from resting state functional MRI. A conventional cross-sectional analysis revealed a moderate S-F correlation magnitude (r = 0.34). In contrast, the proposed framework demonstrated a significant increase in S-F coupling to r = 0.46 when the structural maturation precedes functional maturation by approximately five days. These findings provide novel evidence of a functional maturation lag relative to structural brain development in neonates. Beyond elucidating S-F relationships in the early developing brain, this work establishes a framework for future longitudinal studies and advances in brain modeling across developmental trajectories, aging, and disease prediction.

Background: MRI has increasingly been explored as a biomarker for detecting structural and functional brain changes. For clinical decision-making, it is crucial to validate observed changes in MRI indices at the individual level. The uncertainty in longitudinal MRI indices can be quantified using the repeatability coefficient (RC). Methods: Twenty healthy controls (10 males, 10 females) underwent two test-retest sessions of structural magnetic resonance imaging (MRI) and resting-state functional MRI (rs-fMRI) on the same day, separated by a 30-minute interval. RC values and their 95% confidence intervals (CI) were estimated for subcortical volumes, cortical thickness, and within-network functional connectivity. Additionally, 33 patients with mental health disorders underwent MRI before and after 20 sessions of transcranial magnetic stimulation (TMS). Percentage changes in MRI-derived indices were assessed at the individual level, with changes exceeding the RC threshold considered indicative of true change beyond measurement uncertainty. Results: The RC showed measurement variability in subcortical volumetric in the range of 8% to 17.5% for caudate and left amygdala, respectively. For cortical thickness, the RC was measured between 3.5% and 16.5% for the left occipital pole and the left temporal pole, respectively. The RC% for fractional anisotropy (FA) measures were variable between 11.3% (the left middle cingulum) and 62.9% (the right anterior cingulum). For within-network connectivity, the RC was measured in a range of 9.7% and 29.4% for sensorimotor and visual networks, respectively. TMS-treated patients exhibited no changes beyond the RC in almost all subcortical volumes and within-network connectivity. The most frequent changes beyond the uncertainty were observed in FA measures, particularly in the posterior cingulum, where 17 out of 23 patients exhibited clinically meaningful alterations. Conclusion: Structural brain features extracted from MRI demonstrated high reliability. Among all measures, FA, reflecting white matter integrity, was most sensitive in detecting neural changes following TMS, highlighting its potential utility as a treatment-responsive biomarker.

Background: Quantitative MRI (qMRI) provides voxel-wise measurements of tissue properties related to myelin, iron and water content, making it a powerful tool for studying brain aging and microstructural alterations in vivo. However, conventional spatial smoothing can introduce partial-volume effects and blur tissue boundaries, potentially affecting both statistical sensitivity and anatomical specificity. Several tissue-specific smoothing strategies have been proposed to address these limitations, yet their relative impact on voxel-wise statistical analyses remains insufficiently characterized. The present study aims (i) to systematically compare three tissue-specific smoothing strategies: a linear tissue-weighted compensated approach (TWS), a generalized version of nonlinear tissue-masked compensated smoothing approach (gTSPOON), and an intensity-weighted edge-preserving approach based on the Smallest Univalue Segment Assimilating Nucleus smoothing (SUSANs), and (ii) to investigate how smoothing approaches interact with statistical inference frameworks by comparing parametric and non-parametric voxel-wise analyse. Methods: Analyses were performed on a publicly available lifespan qMRI dataset comprising 138 healthy participants (19-75 years) and quantitative maps of MTsat, PD, R1, and R2*. The generalized TSPOON (gTSPOON) method was implemented using tissue-specific masks derived from probabilistic tissue segmentation. All three smoothing approaches (TWS, gTSPOON and SUSANs) were parameterized to achieve comparable nominal spatial smoothing. Age-related effects were investigated separately in GM and WM using voxel-wise general linear models following a previously published framework. Statistical inference was assessed using multiple complementary approaches, including parametric Random Field Theory (RFT), under both stationarity and non-stationarity assumptions, as well as non-parametric permutation-based inference. In addition to conventional thresholded statistical parametric maps, voxel-wise log-likelihood (LL) maps were computed to quantify general linear model (GLM) goodness-of-fit independently of statistical thresholding. Bland-Altman analyses and spatial agreement metrics were subsequently used to compare smoothing strategies. Results: TWS and gTSPOON produced highly similar spatial distributions of age-related effects across all qMRI parameters and tissue classes. However, TWS consistently yielded a larger number of significant voxels and clusters, reflecting slightly higher sensitivity, from slightly wider effective smoothness and reduced RESEL counts. By contrast, SUSANs generated substantially fewer significant voxels and clusters, associated with approximately half the effective smoothness and a markedly larger number of RESELs. Despite these differences in statistical sensitivity, voxel-wise LL analyses revealed distinct anatomical preferences for each smoothing strategy. TWS provided the best model fit predominantly within GM, whereas gTSPOON showed superior performance in homogeneous WM regions. Conversely, SUSANs achieved the highest LL values at GM-WM interfaces, particularly within sulcal and gyral transitions, indicating improved preservation of sharp anatomical gradients. These spatial patterns were consistently observed across MTsat, PD, R1 and R2* maps. Comparisons across stationary and non-stationary RFT assumptions revealed only minor differences, while non-parametric inference produced highly concordant results, indicating that the primary source of variability originated from the smoothing procedure itself rather than the inference framework. Conclusions: Tissue-specific smoothing strategies substantially influence both statistical sensitivity and voxel-wise model fitting in qMRI analyses. While TWS and gTSPOON provide highly consistent results, the edge-preserving SUSANs approach preferentially enhances model fit at tissue boundaries. Importantly, voxel-wise log-likelihood mapping revealed that no smoothing strategy is uniformly optimal throughout the brain; instead, each method exhibits anatomically preferential regions where model fit is maximized. These findings suggest that smoothing should be viewed as a region-dependent optimization problem and highlight voxel-wise LL mapping as a principled framework for selecting or developing adaptive smoothing strategies tailored to specific neuroanatomical structures and biological processes, including age-related brain changes.

Magnetic resonance imaging scans are effective tools for unveiling brain structures and understanding pathology for complex neurodevelopmental and aging processes. The spatial correlations among various brain regions reveal critical information and insights into the mechanisms of brain functions and their associations with cognitive abilities. Large-scale neuroimaging studies that acquire or aggregate imaging scans from multiple sites have become increasingly popular. Doing so enhances the diversity of study samples and robustness of study findings, and increases the statistical power of any analysis conducted for the biological hypotheses of interest. However, collecting images across different sites introduces non-biological variability attributed to differences in imaging protocol and configurations, known as batch effects. While there are methods to perform this batch effect correction, there are limited methods that directly account for the spatial patterns found in images of the brain. We develop Covariance-Aware Multivariate (CAM) ComBat that accounts for such spatial correlation across high-dimensional features, which could be heterogeneous across batches. We also propose a computationally efficient alternative of CAM-ComBat, Spatially-Informed Iterative Block (SIB) ComBat, that is scalable for very high dimension of features. We show that these methods outperform existing batch effect correction methods through simulation studies and an application to real neuroimaging data.

Fairness audits of clinical AI models rarely make the evidentiary status of subgroup findings explicit: reassuring results may reflect insufficient statistical precision rather than true parity, and audit verdicts can easily reverse under equally defensible analytic choices. We introduce an evidence classification scheme that screens for sample size and precision, and integrates stability across design alternatives directly into the fairness claim. We demonstrate this scheme on the estimation of the brain-age gap (BAG), a potential clinical biomarker, from structural MRI using the Alzheimer's Disease Neuroimaging Initiative (ADNI) data. The male-female and Black-vs-White differences, along with the White-Male and Black-Female intersectional contrasts, are all classified as equivalence supported, stable across regressor choice (ridge vs. gradient-boosted trees) and feature representation (full feature set vs. cortical-thickness-only). The Asian-vs-White and Black-Male comparisons remain classified as insufficient data throughout, as neither meets the pre-specified minimum-sample threshold. The proposed scheme provides a path from raw fairness findings to justified fairness claims via pre-specified thresholds, minimum-information screening, and stability checks across declared design choices.

Background: White matter hyperintensities (WMH) represent the most visible manifestation of cerebral small vessel disease and of white matter pathology more broadly, yet empirical evidence points to a brain tissue injury extending beyond radiologically detectable lesions on fluid-attenuated inversion recovery (FLAIR) MRI. We present RADAR-WMH (Relaxometry And Diffusion Analysis for Radiological WMH), a multicontrast MRI machine learning framework that characterises white matter pathology through tissue microstructural information rather than lesion contrast alone. Methods: RADAR-WMH was trained on quantitative relaxometry and diffusion-weighted MRI acquired in community-dwelling participants (mean age 59.6 years [SD 22.4], 60.8% women, n=148) using a LightGBM classifier integrating local, textural, and anatomical features at the voxel level. Biological validity was assessed through longitudinal analyses and associations with age, cardio-vascular risk, and cognitive performance in independent cohorts. Results: RADAR-WMH achieved segmentation performance comparable to state-of-the-art FLAIR-based approaches without requiring FLAIR or T1-weighted data. Mean diffusivity was the most influential feature for lesion classification. Beyond FLAIR-defined WMH, RADAR-WMH identified tissue pathology extending outside lesion borders characterised by myelin loss, axonal injury, and increased extracellular water. These microstructural signatures persisted over follow-up and showed stronger association with age, systolic blood pressure, and cognitive performance than corresponding tissue properties restricted to FLAIR-defined WMH extent. Conclusions: RADAR-WMH reveals a significant burden of biologically meaningful white matter injury that remains invisible to FLAIR-defined WMH segmentation. By capturing microstructural pathology linked to vascular risk, cognitive decline, and lesion evolution, RADAR-WMH may provide more sensitive markers of cerebral small vessel disease than FLAIR-visible WMH alone.

Altered attention network activity has been consistently reported in recent neuroimaging literature of smartphone addiction and problematic smartphone use. However, many functional imaging studies have relied on traditional static functional connectivity analyses, which cannot capture time-varying or recurrent dynamics of attention networks. Given the importance of temporal and spatial dynamics for attentional processing, we applied two complementary dynamic functional connectivity approaches: functional connectivity-based attractor networks (fcANN), which characterize transitions between discrete attractor states related to internal-external context and action-perception dynamics, and cPCA-based quasi-periodic pattern (QPP) analysis, which captures phase relationships between cortical networks including the default mode network, dorsal attention network, and frontoparietal control network. In a resting-state fMRI dataset comparing participants with High and Low SAS-M scores (SAS-M: Smartphone addiction scale-Malay version questionnaire), we observed reduced DAN participation within subject-specific cPCA-derived QPP-like components, along with context-dependent differences in DMN-attention-network relationships across attractor states. The most consistent effects emerged during transitions between internal and external attractor states, where subjects with high smartphone addiction (SPA) scores showed greater DMN dominance relative to attention-related networks. Overall, these findings provide preliminary evidence that smartphone addiction may be associated with altered large-scale attentional dynamics at rest

Purpose: To investigate sex differences in cerebral blood flow through densely parcellated cortical and subcortical regions using explainable artificial intelligence methods and identify neurobiologically interpretable perfusion biomarkers. Methods: High-resolution pseudo-continuous arterial spin labelling (1.875 mm x 1.875 mm x 3 mm) and structural MRI data were curated from 215 healthy young adults (150 females, 95 males; age 18-30 years) from the publicly available I See your Brains (ISYB) dataset. Cerebral blood flow was quantified using atlas-based regional analysis with the Brainnetome Atlas (246 regions) and optimized registration procedures. Sex classification employed diverse machine learning paradigms including linear classifiers, ensemble methods, and kernel-based approaches for regional CBF features, with deep convolutional neural networks (CNN) applied to whole-brain 3D imaging data. Model interpretability was achieved using SHapley Additive exPlanations (SHAP), computed over an ensemble of 500 logistic regression models (100 iterations x 5-fold cross-validation). Regions appearing among the top 20% of discriminative features more than 289 times were considered statistically significant using binomial testing. GradCAM was used to obtain class-specific attribution maps from the CNN model. Results: Perfusion-based features demonstrated superior sex classification performance compared to structural morphometry. Regional CBF analysis using logistic regression achieved 91 +/- 2% balanced accuracy and 0.95 +/- 0.05 ROC-AUC, substantially outperforming morphometric features (85 +/- 8% balanced accuracy, 0.88 +/- 0.06 ROC-AUC). Deep learning classification of 3D CBF maps achieved a performance of 92 +/- 5% balanced accuracy, 0.92 +/- 0.05 ROC-AUC. SHAP analysis identified 30 statistically significant aggregation-agnostic CBF-based biomarker regions using regional CBF, predominantly involving frontoparietal control networks (27%) and default mode networks (17%). Grad-CAM revealed that the 3D CNN model primarily focused on regions within the frontal lobe. Morphometry-based analysis identified 28 discriminative regions with markedly different anatomical distribution (r = 0.21) emphasizing visual (32%) and default mode (14%) networks. Conclusion: Cerebral blood flow patterns provide highly sensitive and biologically interpretable markers of sex differences in young adult brain. The identification of robust perfusion biomarkers through explainable AI demonstrates the clinical potential of ASL imaging for precision medicine applications in neuroscience. We establish a methodological framework for investigating sex-specific brain physiology using non-invasive neuroimaging.

As neuroimaging analysis shifts toward large-scale, multi-site studies, managing the unwanted variability introduced by combining heterogeneous datasets has become a critical challenge. Although tools such as ComBat and its neuroimaging extensions are widely used to address this variability, they only permit the modeling of categorical site effects and cannot account for continuous sources of confounding, such as image quality, head motion, and acquisition parameters. We introduce ComCat, an extension of the ComBat framework that preserves biologically relevant covariates while removing the effects of categorical site indicators and continuous nuisance variables. The latter are modeled as smooth nonlinear functions via B-spline basis expansion. ComCat is applicable to a broad range of brain analysis tasks, including voxel- and surface-based morphometry, normative modeling, and machine learning-based prediction. To demonstrate its capabilities, we evaluated ComCat on brain age prediction across five datasets covering complementary multi-site harmonization scenarios: ON-Harmony (10 subjects x 6 scanners; n = 80); the Buchert traveling-phantom dataset (1 subject x 116 scanners; n = 531); the Tohoku single-scanner, varying-acquisition dataset (n = 121); MR-ART (148 subjects with varying motion levels); and an ABIDE subset comprising 229 control subjects and 208 individuals with autism spectrum disorder across 14 scanners. Using image quality measures derived from CAT12 as continuous nuisance variables, ComCat reduced the mean absolute error (MAE) in brain age prediction relative to ComBat-GAM in all five datasets, including the two scenarios where site information was unavailable or uninformative. In the ABIDE dataset, ComCat improved harmonization while preserving the difference between the control and ASD groups, demonstrating that scanner-related variance can be removed without affecting biologically meaningful signals. ComCat can operate with or without site labels and is agnostic to the source of image quality metrics.

Brain-age estimates from grey matter structure provide a promising tool to examine the neurobiology of mental health. However, whether existing models generalise to longitudinal adolescent data remains unclear. The CentileBrain Global-BrainAGE Lifespan Model, was applied to N=138 participants (74 females, 64 males) from the Longitudinal Adolescent Brain Study. Participants completed 2-14 MRI scans between the ages of 12-17 years (752 datapoints). Model fit, prediction accuracy and longitudinal consistency were examined. Results showed moderate-to-good longitudinal consistency and reliability, consistent with high-performing cross-sectional age-to-brain-age correlations in youth cohorts. However, the model systematically overestimated brain-age changes relative to chronological changes, and prediction accuracy was unstable, with the mean absolute error increasing with age. Despite sex-specific brain-age calculations, on average, females showed older brain-ages compared with males. Further, younger adolescents were underpredicted, while older adolescents were overpredicted, indicating that standard model adjustments and assumptions may not be applicable.

Optically pumped magnetometer magnetoencephalography (OPM-MEG) offers a wearable, movement-tolerant alternative to conventional cryogenic MEG, placing sensors closer to the scalp and, in principle, improving sensitivity to deep sources. This is advantageous for examining subcortical structures that are affected by ageing, disorders and disease, such as the hippocampus. However, it remains unclear whether well-established activity (such as the attenuation of theta oscillations during the imagination of novel scenes) can be recovered from the medial temporal lobe (MTL) with OPM-MEG, and whether an individual structural MRI is required. Here, fifteen adults completed a scene imagination task. Initially we applied a 12-parameter template warping coregistration pipeline to the full sample. Following source reconstruction, we recovered the expected attenuation of theta (4-8 Hz) power during scene imagination compared to a counting baseline, with a significant cluster of activity peaking in the left parahippocampal gyrus. The clusters centre of mass was localised to the left hippocampus (t = -3.55, p = 0.048, whole-brain FWE-corrected) and was mostly confined to the left medial temporal lobe. We further supported our findings by using an individual T1-weighted MRI reconstruction pipeline in six participants who had these scans available. The two approaches produced similar whole-brain topographies and localised the peak MTL theta effect to left hippocampus, with temporal-lobe conjunction centroids 3-mm apart. These findings provide evidence that the theta attenuation of the scene construction network can be recovered at the group level with OPM-MEG, without an individual MRI.

Whole-brain functional connectivity, estimated from magnetoencephalography (MEG) data, provides a compact representation of long-range neuronal communication, making it suitable for predictive biomarker discovery. In this work, we propose a deep learning framework (FC-CNN) for predicting brain states from frequency-resolved functional connectivity estimates derived from resting-state MEG recordings. We systematically compare the performance of FC-CNN to that of conventional regression methods using amplitude and phase-based functional connectivity in the well-studied age-prediction task on the Cam-CAN cohort (n=576). We show that FC-CNN outperforms conventional approaches, and that, compared to phase synchronization, amplitude envelope correlation consistently leads to higher prediction performance. Moreover, we present quantitative evidence that the weights of a trained deep learning model can enable neurophysiological interpretation of the activity patterns that inform successful predictions. Our work demonstrates that the proposed approach successfully decodes brain states from MEG functional connectivity and is promising for discovery of predictive biomarkers for brain disorders.