Neuroinformatics

We propose a study protocol for routine clinical electroencephalograms (EEGs) from public hospitals, which represents a vast resource for neuroscience research. These non-invasive measures of brain function, paired with rich clinical annotations from large and diverse patient populations, are critical for developing robust artificial intelligence (AI) models and conducting population-level studies. This protocol presents a scalable methodology for curating and harmonizing extensive clinical EEG datasets, encompassing over 40,000 individual studies, to facilitate research applications. Key steps include: (i) integration of raw EEG recordings with corresponding clinical records, including neurological reports, diagnostic codes, and potentially medication data; and (ii) spatial standardization of EEG signals by mapping them to a common brain space defined by functional and anatomical landmarks. The resulting harmonized datasets enable the development of large-scale EEG foundation models, the discovery of novel EEG waveform representations, and the creation of normative "brain charts" for electrophysiological assessment across the lifespan. By enabling standardised, large-scale analyses of real-world clinical EEG data, this protocol supports data-intensive solutions for EEG applications and addresses the challenge of generalising AI models. Our approach promotes the translation of AI tools from research to diverse patient populations, advancing population neuroscience.

Abnormal states such as erroneous behaviors are generally difficult to represent from neural data. However, such states are also known to have specific spatiotemporal features, indicating a feasibility of developing a method to focus on them. If a method can highlight these spatiotemporal features, it may effectively represent such abnormal states, helping evaluate abnormal brain functions. In the present study, we proposed the hierarchy of supported modules (HSM) to highlight spatiotemporal features that can represent abnormal states. HSM spatiotemporally transforms multidimensional neural time-series based on their spatiotemporal context. We evaluated HSM through decoding and similarity analyses using multiple publicly available datasets. In the HSM results, decoding accuracies were higher for erroneous behaviors than for normal behaviors, and similarities were lower between erroneous behaviors and normal behaviors than between normal behaviors, demonstrating the ability of HSM to capture the spatiotemporal features of erroneous behaviors. Surprisingly, many parts of these results were also present even before HSM learning, showing the virtue of HSM as a simple-to-use method. The proposed HSM method may help elucidate the mechanisms underlying erroneous behaviors.

Introduction Understanding the mechanisms of human neuromaturation constitutes one of the fundamental questions of neuroscience. While it is well described that large-scale brain maturation is initiated within sensorimotor brain regions and progresses to associative cortex, the underlying developmental neurobiology remains to be fully characterized. Animal models have indicated that cortical inhibitory upregulation might be a driver of neurodevelopment. To investigate the hypothesis that cortical inhibitory upregulation plays a similar role in human neuromaturation, we developed a The Virtual Brain (TVB) based computational model (TVB-Child) to explore potential mechanisms of human neurodevelopment. Material and method We created neurodevelopmental dynamic brain network models capturing neurobiological maturation by using the large-scale brain simulator TVB and fitting brain network models to developmental functional MRI (fMRI) from the Human Connectome Project-Development (HCP-D) data set with 640 subjects with an age range of 6-21 years. Age-dependent trajectories in the fMRI data set were first analyzed by combined group-ICA/Dual Regression extracting subject-specific resting-state networks (RSN). Maturational topographical and topological redistribution of these networks were analyzed by linear and non-linear regression of RSN size and degree and strength centrality. Brain network models were fitted to the fMRI functional connectivity obtained from the HCP-D data set. Hypothesizing that cortical inhibition is a driver of neuromaturation, we analyzed spatiotemporal inhibition parameter gradients in the dynamic brain network model for the hypothesized significant correlations with fMRI RSN maturational trajectories. Results While during development frontoparietal (FP) and default mode network (DMN) grew and exhibited an increase in both degree and strength centrality, becoming dominant network hubs, the attention network underwent network pruning with a decrease in size and node degree. The primary sensory network changed little. For the fitted brain network models, we obtained a high degree of reproduction with correlation coefficients between empirical and simulated functional connectivities ranging between 0.80 and 0.95. Values of the feed forward inhibition model parameter wijFFI representing the strength of regional feedforward inhibitory input exhibited the most significant increase with age within the FP and DMN networks. A less pronounced, but significant, age-dependent increase of the inhibitory parameter values were seen in attention networks and no change within primary sensory networks. Conclusion Our study shows that high order (FP, DMN), attention and primary sensory networks exhibit distinct topographical and topological maturation trajectories. Moreover, brain network modeling revealed RSN-specific age-dependent inhibition trajectories, indicating that the model is able to reproduce and thus support candidate mechanisms of neurodevelopment.

Mapping hierarchical brain networks within traditional Euclidean space causes significant structural distortion, undermining neuroimaging diagnostic frameworks. While hyperbolic models like the Poincare ball preserve these nested topologies, they demand heavy computational overhead due to intricate Mobius operations and curved geodesics. This paper introduces a highly efficient non-Euclidean framework for analyzing neurocognitive decline utilizing the Beltrami-Klein ball model. By projecting hyperbolic geodesics as Euclidean straight lines, this approach converts complex distance calculations into simple dot products, radically reducing processing demands. We validated our methodology against state-of-the-art Poincare and Lorentz baselines using datasets for Schizophrenia, Parkinsons Disease, and Alzheimers Disease. The Klein-based framework demonstrates superior performance, delivering both higher diagnostic precision and accelerated processing velocities across all three neurocognitive disorders.

Neurophysiology datasets are becoming increasingly complex, combining behavioral measurements with high-dimensional neuronal activity recordings coming from optical and/or electrophysiological measurements. The Neurodata Without Borders (NWB) standard has emerged in the community as the format of record. While standardized and widely used preprocessing tools generating NWB files have been developed, extensible frameworks for scientific analysis downstream of the NWB ecosystem are still under-represented. We present CICADA, a Python framework dedicated to analysis of neurophysiological data in the standardized NWB format. The toolbox is built as three hierarchically-organized packages: cicada-nwb (NWB access layer), cicada-analysis (plugin-based analysis engine and tool library), and cicada-gui (PyQt5 desktop application at the head of the pipeline). Beyond this architectural separation, CICADA is built around a central design principle: supporting a continuum from turnkey use to full modularity. Researchers can use the complete GUI-driven cicada-gui workflow without writing code, programmatically use existing analysis plugins from cicada-analysis, contribute to new analysis plugins, reuse utilities from cicada-tools, or build entirely custom pipelines on top of the cicada-nwb access layer alone. The same analysis plugin runs identically in interactive GUI and parameter-configured headless modes, enabling reproducible multi-session, multi-animal group analyses. We illustrate the versatility of CICADA with example analyses of behavioral, calcium imaging (two-photon and widefield) and extracellular electrophysiology datasets from rodent laboratories. CICADA is open source, actively maintained, and designed so that any laboratory can contribute at any level of the stack without modifying the core framework.

The corpus callosum is a major white matter bundle responsible for connecting both hemispheres. In mammals, due to a variety of causes, the development of the corpus callosum can be impaired - this brain malformation is known as corpus callosum dysgenesis (CCD). The clinical presentation of CCD varies, with patients exhibiting three morphological phenotypes: agenesis, partial dysgenesis, and hypoplasia. Although the first two presentations are easily detectable on MRI scans, the latter is more challenging, as the structure is fully formed but has a reduced area. In this study, we develop (1) a pipeline to generate synthetic MRI scans with apparent anatomical variation and (2) train a U-Net-based tool to automatically segment the corpus callosum of marmosets in both healthy and disease contexts. Methodologically, a custom script was devised to apply rotation and translation to T1-weighted MRI scans at the volume level. Because the slicing grid remains unchanged, these rigid transformations translate into apparent anatomical variations at the slice level. We compared corpus callosum measurements obtained from automatically segmented masks with those from manually delineated masks. The average Dice score was above 0.90, and the Hausdorff distance was below 0.4 mm. We also stratified our cohort according to phenotype (healthy controls and hypoplastic animals). The magnitude of the effect and the significance level observed between the voxel counts of healthy and hypoplastic animals using manually delineated masks were comparable to those obtained via automatic segmentations. These results show that our pipeline can generate a sufficiently varied training pool to build an accurate U-Net segmentation model with high diagnostic capability.

Wave recurrent neural networks (wRNNs) are biologically inspired recurrent architectures that use traveling-wave dynamics to support sequence learning and memory. However, their input-to-hidden pathway remains relatively simple compared with biological neurons, where dendrites perform nonlinear input integration. In this study, we introduce the Dendritic Wave Recurrent Neural Network (DWRNN), which augments the input pathway of the wRNN with nonlinear basal dendritic branches while preserving the original recurrent wave dynamics. We evaluate DW-RNN on a simple copy task, sequential MNIST (sMNIST), permuted sequential MNIST (psMNIST), and noisy sequential CIFAR-10 (nsCIFAR-10). On the copy task, DW-RNN shows learning behavior comparable to the standard wRNN, suggesting that dendritic input integration does not disrupt the recurrent wave-based memory mechanism. On the three sequential image-classification benchmarks, DW-RNN outperforms the standard wRNN, improving accuracy from 97.27 {+/-} 0.15% to 97.82 {+/-} 0.12% on sMNIST, from 96.74 {+/-} 0.17% to 96.92 {+/-} 0.10% on psMNIST, and from 54.30 {+/-} 0.79% to 55.65 {+/-} 0.55% on nsCIFAR-10. In addition to improving mean accuracy, DW-RNN exhibits lower across-seed variability on all three classification benchmarks, suggesting that dendritic input integration may improve the stability of wRNN training. Hidden-activity visualizations further show that DW-RNN preserves the characteristic traveling-wave patterns of the original wRNN. These results suggest that dendritic computation and traveling-wave recurrent dynamics provide complementary mechanisms for biologically inspired sequence learning.

Traumatic brain injury produces widespread axonal damage can be assessed histologically using amyloid precursor protein (APP) immunohistochemistry, which labels injured axonal profiles at cellular resolution [1, 2]. However, quantification of APP pathology remains a major bottleneck: annotation is manual, time-consuming, spatially localized, and variable across raters, limiting scalability and reproducibility. This limitation is particularly important in studies that use histology as a reference for neuroimaging or other tissue-level measurements, where cellular APP pathology must be quantified in a spatial form that can be aligned with imaging abnormalities. Here, we introduce PIGMENT, an annotation-efficient deep-learning framework for automated segmentation and quantification of APP-positive pathology in porcine white matter histology. PIGMENT uses a compact SegFormer-B0 architecture trained on 525 expert-annotated 512 x 512-pixel tiles from four APP-stained sections across three pigs. Because APP-positive profiles are sparse, fragmented, stain-variable, and morphologically diverse, PIGMENT combines limited expert labels with APP-specific augmentation designed to model variation in APP-positive intensity, size, continuity, fragmentation, and local tissue context. We evaluated PIGMENT using an instance-level detection rate that measures whether discrete APP-positive components are localized. Across held-out APP-stained data, PIGMENT achieved a mean instance-level detection rate of 0.86. Across the configurations tested, the highest mean detection rate was achieved by a training set that included sections from different animals, suggesting that annotation diversity may be an important factor under limited-label conditions. By extending limited high-confidence expert annotations into whole-section APP burden maps, PIGMENT provides a scalable framework for characterizing the extent and spatial distribution of traumatic axonal injury. These maps may support future studies that align histological injury burden with imaging-derived measures.

Converting medical images into anatomically detailed, subject-specific finite element (FE) models is a long-standing bottleneck in brain computational modelling. These models are used to predict brain tissue deformation, e.g. in traumatic brain injury, particle diffusion in brain drug delivery, and other biophysical phenomena across neurological disorders. However, existing model creation workflows depend on manual image segmentation, proprietary meshing software, and labour-intensive repair of meningeal and interface structures, limiting reproducibility and cohort analysis. Here we present PARS, a fully automated, open-source pipeline that converts a T1-weighted MRI scan into a simulation-ready FE head model. PARS combines anatomical parcellation with tissue maps and uses iterative neighbourhood-based reclassification, yielding a gap-free whole-head label volume. The volume is directly converted into a hexahedral mesh, augmented with algorithmically reconstructed falx, tentorium, pia and dura mater, and refined by Laplacian smoothing under a node-locking scheme that controls element quality and the explicit-solver stable timestep. We evaluated PARS on 23 subjects spanning cranial volumes of 832 to 1,329 cubic centimeter, at 1.0, 1.5 and 2.0 mm MRI resolutions. At 1 mm, meshes achieved a median Scaled Jacobian of 0.976, and total intracranial volume error of ~0.54; quality remained high at 1.5 mm (SJ of 0.933) and 2 mm (SJ of 0.921). Model creation runtime ranged from 9 to 38 minutes per subject. Models generated by PARS have been validated against cadaveric brain displacement data and demonstrated utility across traumatic brain injury and normal pressure hydrocephalus research. PARS provides an open-access, reproducible resource that substantially lowers the barriers to subject-specific brain modelling.

Capturing naturalistic behavioral dynamics is essential for understanding social interaction in ecologically valid settings. Existing investigations of naturalistic social interaction rely on time-aggregated analysis methods better suited for task-based experiments, which lose the complex, moment-to-moment dynamics exhibited in naturalistic settings. The emerging field of topological data analysis (TDA) provides new tools to characterize fine-grained dynamics in time-series data that cannot be captured by time-averaged methods. The present work utilizes Temporal Mapper, a recently developed TDA specifically tailored to analyzing dynamical systems. Temporal Mapper characterizes complex temporal dynamics as transition networks, where nodes are stable states and edges are transitions between states. Originally designed for human neural time series analysis, here we demonstrate the utility of Temporal Mapper to capture rich animal postural dynamics during naturalistic social interaction. We utilized an existing dataset with 12 video recording sessions of two domestic ferrets (Mustela putorius furo) during naturalistic interaction and tracked the postures of animals during social interaction. Ferrets were chosen due to their strong social-cognitive skills and rich postural dynamics for investigating social behavior via posture estimation. Temporal Mapper was then used to represent the postural dynamics as transition networks for each recording session. Here, we found that posture states are significantly smaller and more widespread during active social interaction compared to non-social activities. Additionally, the number of sequential postural states before transitioning to new behaviors is more consistent during active social interaction than non-social activities. Together, our findings suggest that social activity has a broad range of unstable postural states arranged in consistent sequences. Our method, Temporal Mapper, allows for network structure analysis of complex naturalistic data, applicable for characterizing rich dynamics in different species, scales, and paradigms.

Deep brain stimulation has demonstrated its therapeutic potential in modulating pathological oscillations associated with Parkinsons disease and epilepsy. However, its efficacy in treating disrupted theta-gamma phase-amplitude coupling seen in memory-related disorders, such as Alzheimers disease, remains poorly understood. While recent studies have targeted the entorhinal-hippocampal circuit, results remain inconsistent. This discrepancy stems from a lack of mechanistic understanding regarding how stimulation protocols affect this circuit. In this work, we present a reduced multicompartment model of the hippocampal CA1 area that reproduces theta-nested gamma oscillations characteristic of healthy neural activity during memory performance. The model comprises pyramidal, basket and OLM cells with simplified morphologies. We also incorporated CA3-to-CA1 axonal projections, providing a foundational framework for studying how stimulation-induced recruitment of afferent pathways modulates CA1 dynamics. By balancing computational efficiency with anatomical accuracy, our model enables systematic investigation of the effects of electrode placement and orientation, as well as stimulation amplitude and frequency on CA1 neural activity. We demonstrate that the excitatory response in CA1 is primarily driven by the recruitment of Schaffer collateral projections. Overall, this work provides a computationally efficient template for exploring diverse stimulation configurations and could be expanded for developing neuromodulatory strategies to restore physiological network dynamics. Author summaryDeep brain stimulation has shown success in treating Parkinsons disease by suppressing abnormal neural activity responsible for movement disorders. However, when applied to memory-related pathologies, such as Alzheimers disease, the therapeutic outcomes remain unpredictable, ranging from cognitive improvement to impairment. This discrepancy highlights a critical gap in our understanding of how stimulation protocols interact with neural dynamics of the targeted circuits. To address this, we developed a computationally efficient model of the hippocampus, which is involved in memory processes, in order to understand how deep brain stimulation might influence its activity. Our model maintains enough biological accuracy to capture essential memory-related neural activity while remaining lightweight enough for rapid execution and systematic exploration of different protocols. This computational efficiency allowed us to conduct systematic investigations of several stimulation configurations to study their effects on hippocampal dynamics. Overall, this model could provide a useful and computationally cost-efficient tool for exploring the mechanisms of deep brain stimulation and help optimize stimulation protocols aimed at alleviating memory disorders.

Cortico-thalamo-cortical circuits entail extensive trans-thalamic connectivity between cortical areas, yet their structural organization and function remain poorly understood. Here, the thalamocortical projections of several higher-order thalamic nuclei were characterized by retrograde tracing from two cortical areas, the primary somatosensory (S1) and motor (M1) cortices. Cholera toxin B conjugated with different fluorophores allowed for simultaneous detection of projection neurons targeting S1 and M1. A cell detection pipeline based on neural networks was developed to allow semi-automated analysis of large thalamic imaging volumes to quantitatively infer the spatial distribution of projection neurons in the posterior complex (PO) and the adjacent ethmoid nucleus (Eth), nucleus centrolateralis (CL), nucleus paracentralis (PCN), and the nucleus parafascicularis (PF). The arrangement of neurons projecting to both, primary somatosensory and motor cortices, occurs at different connection strengths and was topographically organized in all nuclei studied. Co-injections into both cortical areas revealed projection neurons with axons branching into both S1 and M1 cortices. Our work introduces a pipeline for semi-automated quantitative analysis of thalamic projection patterns that could be useful for connectivity analyses in general. This approach revealed repetitive anatomical patterns in different thalamic nuclei with regard to projection strength, spatial organization and fraction of projection neurons targeting two cortical areas simultaneously.

golgi is an open-source platform that takes a peripheral nerve from image to stimulated fiber population through a single graphical interface, with an equivalent scriptable Python API and command-line interface for batch and high-performance use. It integrates promptable image segmentation, automated multi-region tetrahedral meshing, anisotropic finite-element solution of the extracellular field with an explicit perineurium contact impedance, generation of realistic fiber populations and their three-dimensional trajectories, and biophysical activation thresholds through interchangeable backends-- NEURON (via PyFibers) and a GPU-accelerated surrogate (AxonML). Every study exports as an integrity-hashed bundle whose image-to-recruitment provenance is verifiable byte-for-byte. golgi lowers the barrier to in-silico peripheral nerve stimulation modeling for experimentalists and clinicians, using a fully open finite-element stack with no commercial dependencies.

In patients with epilepsy, seizures are associated with pathological neural synchronization. However, the preictal period preceding a seizure often exhibits reduced spatial synchronization compared to normal cognition. This observation aligns with the concept of the brain as a complex dynamical system, where a reduction in dimensionality and resilience can precede a phase transition. The Critical Brain Hypothesis suggests a connection between the loss of healthy scale-free behavior and various disorders, including epilepsy. Our study investigates preictal changes by utilizing network features, such as mean node degree and mean clustering coefficient, derived from thresholded correlation matrices of patient intracranial electrocorticographic electrode data. We observed a suppression of intermittent high-synchronization periods within the feature space during the minutes leading up to seizure onset. This constriction of the explored hypervolume in the preictal state indicates a breakdown in the brains ability to maintain normal coherence. We use these preictal changes to predict the probability of seizure onset using a Support Vector Machine algorithm. These discrete predictions can then be combined into real-time continuous seizure risk forecasts via Bayesian updating. This innovative and computationally lightweight approach has the potential to significantly improve upon static predictions, providing opportunities for more adaptable, quantitative, and interpretable tools for managing seizures.

Transformer-based deep learning models have shown great potential for decoding visual EEG signals. However, their internal attention mechanisms are often evaluated primarily on optimization objectives, leaving their alignment with biological brain connectivity an open question. This study empirically evaluates how variations in EEG preprocessing strategies affect these attention representations using the Adaptive Thinking Mapper (ATM) model as a framework. We compared a baseline pipeline (MVNN only) against a comprehensive cleaning pipeline integrating ICA and notch filtering. The models were evaluated through cross-generalization, noise robustness, and spectral-temporal ablation analyses. Furthermore, we investigated the structural correspondence between the model's data-driven attention weights and neurophysiological reference networks (GPDC, PDC, and DTF) using Node Strength Correlation and Representational Similarity Analysis (RSA). The results show that the comprehensive preprocessing successfully suppresses non-neural artifacts, such as frontal noise and electrical interference, while maintaining comparable decoding accuracy and baseline robustness. Alignment analyses revealed that the broad spatial organization of the learned attention patterns remains highly stable across pipelines, capturing key directed connectivity dynamics with subtle, metric-dependent variations in global representational geometry. This work provides an empirical exploration into bridging data-driven attention weights with neurophysiological consistency, offering insights toward more transparent brain-computer interfaces.

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.

A recent study (Manjarrez et al., 2026) showed that the classification of cortical dendritic spines into stubby, thin, and mushroom subtypes is unstable under rotation. That result criticizes the categorical scheme but leaves an open question. What is the actual structure of spine morphology once the viewing angle is controlled? Here we answer it. We analyzed 228 spines from layer II pyramidal neurons in the H01 nanometer-resolution reconstruction of human temporal cortex. We first quantified the source of instability. We found that rotating dendritic segments by 90 degrees about their axes shifted the apparent spine height and head width in opposite directions across the population, thereby confirming orientation-dependent measurement error. Furthermore, to obtain measurements free of this artifact, we developed the Spine Morphometry Hub (SMH), a 12-point anatomical landmark framework that characterizes each spine in all three orthogonal planes and extracts geometric, voxel-based, and mesh-based metrics. All morphometric distributions were unimodal and right-skewed. Density-based clustering assigned most spines to noise, and a Monte-Carlo test against a discrete two-type null model confirmed that this pattern is incompatible with categorical subtypes. We also confirmed that apical and basal spines were statistically indistinguishable. Unlike previous reports of a spine continuum, all based on orientation-dependent measurements, our framework removes the viewing-angle confound itself, so the continuum we observe cannot be attributed to a projection artifact. Hence, our framework will be useful to quantify dendritic-spine remodeling in neurological disorders, in which spine shape has long been observed but never measured against an orientation-invariant morphometric standard. HighlightsO_LISpine Morphometry Hub (SMH) measures spines free of viewing-angle error C_LIO_LISMH was validated as an orientation-invariant morphometry framework C_LIO_LIRotating dendrites by 90{degrees} shifts spine height and head width oppositely C_LIO_LIAll morphometric distributions are unimodal and right-skewed, not categorical C_LIO_LISMH could be used to quantify dendritic-spine remodeling in neurological disorders C_LI

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.

Transcranial Electric stimulation (tES) is a safe and noninvasive technique increasingly used in treating brain disorders. Despite many studies on tES, there is still a lack of understanding of its mechanisms at the microscale network level. This is crucial for optimized parameter selection in therapy approaches such as the treatment of pharmacoresistant epilepsy. In this study, we made use of a recently published neuroinspired microscale model of the neocortex, known as NeoCoMM, and integrated a "Lambda E"-based model of tES. This updated version was used to investigate the acute effects of tES (tDCS and tACS), on the neural activity of various neuron types in both healthy and epileptic brain states. Results showed that in the case of healthy alpha and gamma rhythms, tACS induced electric field entrainment at the peak power frequency of the network oscillations as measured by the Local field Potentials (LFPs). This resonance-like entrainment was independent from the individual firing rate of cell types. For epileptic activity, tACS did not provide consistent results. Cathodal tDCS resulted in a promising decrease in hyperexcitable activity throughout simulations. These results advance our understanding of the impact of tES on network dynamics at both the extracellular and intracellular activity levels. Author summary

Personalized whole-brain modeling aims to transform treatment planning for neurological disorders by enabling patient-specific simulations of brain network dynamics. Neural mass models (NMMs) offer a tractable compromise between biophysical detail and computational cost and can be directly linked to macroscopic observables such as EEG. However, scaling NMMs to whole-brain networks with realistic connectivity, conduction delays, and cortical surface resolution--and fitting them to individual patient data--imposes computational demands that existing frameworks cannot meet at clinically relevant timescales. Here we introduce BraiNN, a JAX-based Python framework for large-scale neural mass modeling that achieves speedups of up to two to three orders of magnitude over existing tools by leveraging GPU/TPU-accelerated, XLA-compiled array computation. BraiNN combines a region-level Jansen-Rit network with a subject-specific cortical surface mesh of coupled neural mass models and biophysically grounded EEG forward modeling via reciprocity-based lead fields. Its fully differentiable computational graph enables a hybrid personalization pipeline that pairs Bayesian optimization for global parameter exploration with gradient-based refinement, completing EEG-driven spectral fitting of an eight-dimensional parameter space in approximately 2-3 hours on a single consumer GPU--compared to multiple days with conventional neural mass modeling software. Numerical verification against established benchmarks confirms that BraiNN faithfully reproduces canonical synchronization and bifurcation dynamics of Jansen-Rit networks. By reducing the time requirements for personalizing a high-detail whole-brain surface model from days to a few hours on consumer-grade hardware, BraiNN brings personalized brain network modeling closer to practical use in clinical contexts. We anticipate that BraiNN will serve as a foundation for patient-specific digital twins and EEG-guided neuromodulation planning.