Trang Cao

Attention-deficit/hyperactivity disorder (ADHD) is thought to result from aberrant neurodevelopmental wiring of connectome architecture, yet the precise mechanisms underpinning this miswiring remain unclear. Here, we used a generative mathematical model to understand how physical and geometric constraints on brain network development might be altered in the disorder. Leveraging multisite diffusion magnetic resonance imaging (dMRI) datasets comprising 713 ADHD children and 304 neurotypical controls (aged 6-18 years), we fitted the recently developed geometric eigenmode model (GEM) to individual data describing connectivity between each pair of cortical regions, as quantified using probabilistic tractography. Replicating past work, the GEM was able to recover both topological and topographical properties of the empirical connectomes, but reconstruction accuracy was lower in ADHD children, suggesting a relative untethering of connectome architecture from geometric constraints. This untethering was particularly salient for long-range connections, which were disproportionately affected in the ADHD cohort. Regional variations in the degree to which connectivity was untethered from geometry were correlated with spatial maps of cortical developmental expansion, myelination, and transcriptomic organization, consistent with a neurodevelopmental basis for geometric untethering that is spatially patterned along the cortical hierarchy. Together, these findings suggest that aberrant neurodevelopment in ADHD may untether connectome development from typical geometric and physical constraints, resulting in subtle disruptions of long-range inter-regional connectivity.

Alzheimer’s disease (AD) patients suffer from consequential diagnostic delay due to the lack of accessible biomarkers. They also show different responses to treatments due to disease heterogeneity and progression. Here, we developed a novel framework to identify disease progression and subtypes by using geometric brain signatures derived from multiple neuroimaging modalities, including [ ¹⁸ F]-Florbetapir (AV45) Positron Emission Tomography (PET), [ ¹⁸ F]-Fludeoxyglucose (FDG) PET, and structural Magnetic Resonance Imaging (MRI). These signatures were derived by decomposing corresponding maps of amyloid-beta levels, metabolic activity, and cortical thickness in terms of the fundamental, resonant modes––eigenmodes––of cortical geometry, each tied to a specific spatial resolution scale. Our results showed that geometric eigenmode-based features identified trajectories of disease progression, quantified as pseudotime, in distinct subtypes. The disease progression trajectories and subtypes are identified with high stability and are highly related to biological and cognitive measures. These performances are superior to those obtained using conventional localised features and remain robust across datasets, indicating that geometric signatures of brain structure and function can be used to uncover new markers of AD diagnosis and prognosis that are missed by conventional localisation approaches.

Background: Hereditary cerebellar ataxias (HCAs) are rare neurodegenerative disorders characterised by progressive motor impairment and overlapping clinical phenotypes. Although genetic testing provides etiological diagnosis, diagnostic delays frequently arise before targeted testing, owing to non-specific presentation and limited clinician familiarity. Imaging-derived biomarkers that capture phenotypic expression and network-level consequences of disease could support earlier recognition of hereditary ataxia, guide appropriate genetic testing, and provide sensitive measures of disease evolution. Building on evidence that cortical geometry shapes functional organisation, we hypothesised that geometric signatures derived from structural magnetic resonance imaging (sMRI) could discriminate HCA subtypes and yield progression-sensitive biomarkers, while enabling scalable prediction of function. Methods: We decomposed sMRI and task-evoked functional MRI data from three independent cohorts using cortical geometric eigenmodes, intrinsic spatial patterns defined by cortical surface geometry, to obtain structural and functional geometric signatures. Structural signatures were used to train neural networks for disease classification and to derive biomarkers sensitive to annual progression. We further modelled structure-to-function mappings to predict functional geometric signatures from sMRI and evaluated their diagnostic and longitudinal utility. Findings: Our framework achieved high diagnostic performance, distinguishing healthy controls from Friedreich ataxia (FRDA) with a maximum AUC of 0.93 and separating FRDA from spinocerebellar ataxia type 1 (SCA1) and SCA3, with AUCs up to 0.81, showing cross-cohort generalisability. Structure-to-function-signature prediction achieved coefficient of determination up to 0.62 and correlation reaching 0.86 across health and disease, while predicted functional signatures improved classification beyond structural signatures alone and enabled partial reconstruction of individual task-activation map. Geometric brain signatures showed greater progression sensitivity than conventional volumetric MRI measures. Interpretation: This geometry-driven framework offers novel, objective, multiscale biomarkers for diagnostic-decision-support and monitoring HCAs and provides proof-of-concept for the feasibility of predicting fMRI-equivalent biomarkers in disease from routine sMRI, which is far more practical in movement-disorder populations. Funding: Friedreich Ataxia Research Alliance USA. Research in Context: Evidence before this study: We searched PubMed, Scopus, IEEE Xplore, and Google Scholar, for peer-reviewed studies published up to 2025 using combinations of terms related to hereditary cerebellar ataxia (HCAs), Friedreich ataxia (FRDA), spinocerebellar ataxia (SCA), diagnosis, progression, MRI biomarkers, structural MRI (sMRI), functional MRI (fMRI), machine learning (ML), deep learning (DL), task activation maps, prediction, and geometric eigenmodes. We found that while sMRI studies in HCAs consistently showed patterns of cerebellar, brainstem, and supratentorial atrophy, highlighting their potential diagnostic value as non-invasive biomarkers, existing studies on imaging-based diagnostic tools for HCAs typically used small sample size, single-site data and focused on narrow classification tasks rather than generalisable frameworks for differential diagnosis. Clinically meaningful objective biomarkers sensitive to disease progression are also limited, with most outcome measures relying on subjective clinical rating scales or conventional MRI metrics with restricted sensitivity and reproducibility. In addition, fMRI reveals important network-level abnormalities in disease, however, motion artefacts, task-performance difficulties and long acquisition times limit its applicability in movement-disorder populations. Recent work in healthy populations showed that structural data could predict task-fMRI activation using DL, yet disease-specific and clinically deployable investigations remain unexplored. In parallel, advances in brain geometric eigenmode research underscore that cortical geometry provides a principled structural basis that shapes multi-scale functional organisation. However, no study has investigated cortical geometric signatures as tools to address three major challenges in hereditary ataxia research and clinical care: diagnostic delay, lack of progression-sensitive objective biomarkers, and practical limitations of functional imaging acquisition. Added value of this study: Using a combined geometric and ML framework, we showed that cortical geometric signatures captured multiscale brain organisation that constitute novel, generalisable biomarkers for differential diagnosis across HCA subtypes. Our models reliably distinguished healthy controls from individuals with FRDA, demonstrated consistent performance across independent cohorts, and further separated FRDA from multiple SCA subtypes. Importantly, we provided proof-of-concept for structure-to-function prediction and showed that fMRI-equivalent functional signatures could be inferred from sMRI in both health and disease, enabling reliable approximation of individual task-activation maps without requiring fMRI acquisition. Incorporating these predicted functional signatures improved diagnostic accuracy beyond structural measures alone. Both structural and predicted functional biomarkers demonstrated greater sensitivity to annual disease progression than conventional volumetric metrics, with comparable performance to clinical scales. Implications of all the available evidence: Cortical geometric signatures pave the way for a clinically deployable neuroimaging diagnostic decision-support tool that could guide clinicians toward targeted genetic testing and potentially reduce diagnostic delay in HCAs. These biomarkers provide objective, rater-independent measures of disease evolution that are more...

Growing evidence indicates that macroscopic cortical activity is dominated by propagating waves of excitation. However, many computational models of such wave dynamics assume that the cortex is a spatially homogeneous medium, ignoring the rich regional variations in cellular, molecular, and physiological properties that are known to shape how brain activity evolves through both space and time. Here, we develop a general framework grounded in neural field theory to model how regional heterogeneities in diverse cortical properties shape spatiotemporal brain activity evolving on cortical surface meshes. This enables high resolution, vertex-level simulations without requiring predefined parcellations. The model requires only a standard mesh model of the cortical manifold and a spatial heterogeneity map, providing a biologically grounded and computationally efficient framework that can be generalised to human and non-human species. Using multiple cellular, molecular, and physiological maps in humans—and analogous maps in macaques and marmosets—we show that our model can consistently recapitulate known relationships between regional heterogeneities and variations in cortical wave speed. In particular, we find that models parameterised by heterogeneities in intracortical myelin and excitation-inhibition balance yield the largest performance improvements relative to spatially homogeneous models. Our results identify a key role for regional variations in myelin, receptor, and genetic architecture in shaping the spatial patterning of macroscale cortex-wide activity that is conserved across primate species with diverse cortical geometries.

The intricate network of axonal fibres that forms the mammalian cortical connectome has a complex topology, being organized in a way that is neither completely regular nor random, as well as a characteristic topography, in which specific anatomical locations are imbued with distinctive connectivity profiles. The mechanisms that give rise to such properties remain a mystery. Here, we formulate a simple analytic model derived from neural field theory that prioritizes physical constraints on connectome architecture by assuming that connectivity is preferentially concentrated between pairs of cortical locations that facilitate the excitation of resonant geometric modes of the cortex. We show that the model outperforms existing approaches in reproducing multiple topological and topographical properties of cortical connectomes mapped at spatial scales spanning orders of magnitude in humans, chimpanzees, macaques, marmosets, and mice, as mapped with either non-invasive diffusion magnetic resonance imaging or invasive viral tract-tracing. Our findings thus point to a fundamental role of geometry in shaping the multiscale architecture of cortical connectomes that has been conserved across 90 million years of evolution.

Decades of structural magnetic resonance imaging (MRI) studies have documented alterations of grey matter morphometry in psychiatric disorders, but the field has failed to identify any consensus disease phenotypes. Here, we examine whether current approaches will ever converge on such phenotypes by evaluating the consistency of brain-wide maps of grey matter volume and cortical thickness differences obtained for each of 59 study sites of five psychiatric disorders (schizophrenia, schizoaffective disorder, autism spectrum disorder, major depressive disorder, and bipolar disorder), totaling 2437 patients and 2065 controls. We find that cross-site consistency is low (median r ≤ 0.16); markedly reduced compared to Alzheimer’s Disease (r = 0.54); unexplained by demographic, clinical, or scanner differences; and robust to analytic choices. Using bootstrapping, we observe that consistency may improve for sample sizes ≥200 per group for schizophrenia, but that other disorders may require much larger samples. Our findings indicate that current widespread practices in structural MRI are unlikely to identify robust morphometric phenotypes for psychiatric disorders.