Biology subjects

Fornito, A.

Publications and source records attributed to Fornito, A..

At least 19 recordsLinked to original sources

bioRxiv2026

Alignment with sex-typical brain architecture: Links to mood pathology, sex hormones and transcriptomics of neurodegenerative conditions

Understanding the mechanisms that underpin the pervasive sex bias in risk for specific psychiatric and neurodegenerative conditions could substantially advance efforts promoting personalized medicine. As a first step towards addressing this issue, we applied imaging transcriptomics and multivariate analysis techniques to data from an adolescent sample enriched in anxiety/depression cases, a child sample enriched in attention deficit hyperactivity disorder (ADHD) cases, and adult patients with treatment resistant major depressive disorder (MDD). We tested whether, across sexes, alignment with female-typical, rather than male-typical, patterns of neural functional connectivity (FC) would relate to a diagnosis (adolescent sample) or severity of symptoms (adult sample) associated with internalizing disorders, which tend to be more often diagnosed in females. To provide preliminary evidence on its functional specificity, we compared the sex-aligned FC profile that distinguished neurotypical adolescents from those diagnosed with depression/anxiety against the sex-aligned FC patterns that distinguished neurotypical children from those diagnosed with ADHD. All the analyses controlled for age within each sample and sex. Finally, we examined the spatial overlap between the sex-aligned FC profile linked to an internalizing diagnosis in adolescence and the transcriptomic maps of genes that show sex-dependent dysregulation in neurodegenerative conditions. Across sexes, increasing alignment with female-typical FC patterns, from sensory-perceptual to default mode regions, discriminated between typically developing and anxious/depressed teens and tracked longitudinal fluctuations in depressive symptoms within adult MDD patients. The identified internalizing-related gradient showed a negative correlation with risk for ADHD in the child sample. The brain regions most strongly aligned with female-, rather than male-, typical FC in depression/anxiety overlapped the spatial expression profiles of genes linked to gonadal hormone synthesis, as well as genes dysregulated in female autism spectrum disorder (ASD) and male neurodegenerative conditions, specifically, Alzheimer's Disease (AD), Huntington's Disease (HD) and Multiple Sclerosis (MS). The brain areas most strongly aligned with male-typical, rather than female-typical, FC patterns overlapped the transcriptomic maps of genes related to androgen and estrogen receptors, as well as of genes dysregulated in male ASD and female neurodegenerative conditions (i.e., AD, HD, MS). Our results raise the possibility that the link between mood pathology and neurodegeneration could reflect a partial, regional reversal of sex-normative functional brain profiles.

neuroscience↗

bioRxiv2026

Regional similarity to male-typical brain architecture links childhood ADHD to neurophysiological aging in later adulthood

Identifying early life biomarkers of suboptimal aging trajectories could help design interventions to delay or even prevent the onset of pathological conditions in older adulthood. To address this issue, we capitalize on evidence that the prevalence rates and presentation of neurodevelopmental and neurodegenerative conditions tend to vary between males and females, thereby raising the possibility that processes involved in sexual differentiation could help connect brain maturation and aging outcomes. Here, we focus on attention deficit hyperactivity disorder (ADHD), a neurodevelopmental condition with male-biased prevalence, and examine the relevance of its genetic risk and functional brain profile to aging outcomes. We analyzed multimodal data from a child sample enriched in ADHD cases and a neurologically intact adult sample. Among 6-to-12-year-olds (N =522), genetic vulnerability to pathological aging, indexed via polygenic risk scores (PRSs) for Alzheimer's Disease (AD), predicted ADHD diagnosis across sexes, regardless of psychiatric history and PRSs for multiple disorders (including ADHD). Irrespective of age, boys and girls diagnosed with ADHD showed preferential alignment with male-, rather than female-, typical functional connectivity patterns in metabolically active transmodal areas which are enriched in 5-HT2,4, D1, and mGLU5 receptors. In an independent sample aged 36-90 years (N = 345), the ADHD brain profile correlated with higher levels of testosterone, physiological markers of reproductive aging and poorer performance on AD-relevant cognitive tasks among older individuals of either sex. Our results imply that neurodevelopmental conditions could provide insights into vulnerability to suboptimal aging trajectories and could help guide early interventions targeting at-risk individuals.

neuroscience↗

bioRxiv2026

From Geometry to Hierarchy: Charting the Dominant Modes of Human Cortical Function Through Development

Cortical function in the human cortex is spatially patterned along a macroscale information processing hierarchy. This hierarchy tracks a sensory-association axis anchored at one end by unimodal sensory regions and at the other by transmodal association areas. Biophysical models of brain dynamics suggest that this axis should align with the dominant, resonant modes of cortical anatomy and geometry, but they do not. Instead, the dominant anatomical modes correspond to simple spatial gradients extending along rostro-caudal, medio-lateral, and dorso-ventral axes. How does this divergence between anatomical and functional modes of the cortex arise? Here, using four independent functional magnetic resonance imaging datasets spanning infancy to young adulthood (n=1,512, aged 0-21 years), we show that this spatial divergence is attributable to the maturation of long-range cortico-cortical inter-regional functional coupling. Specifically, we find that the dominant functional mode of the human cortex during the first 3 to 4 years of life closely resembles the rostro-caudal mode of cortical geometry, then shifts to a rudimentary sensory-association-like mode between 4 and 6 years of age and ultimately matures into a prototypical sensory-association mode by age 13. In infancy and early childhood, a linear model predicting the topography of the dominant functional mode from modes of cortical geometry substantially outperforms a model relying on the adult sensory-association mode, whereas the reverse is true in later stages of development. Furthermore, targeted removal of long-range connections causes the adult functional mode to regress from a sensory-association topography to an infant-like rostro-caudal mode. Our findings indicate that the maturation of long-range cortico-cortical coupling supports the gradual emergence of hierarchical modes of cortical function from an organization initially dominated by geometric constraints.

neuroscience↗

bioRxiv2026

An Integrated EEG, TMS-EEG and Behavioural Dataset for Investigating the Neural Correlates of Working Memory

Working memory is a fundamental process that underlies cognition. Accordingly, the neural mechanisms that support working memory performance are of great interest in cognitive neuroscience. We present a publicly accessible dataset comprising 123 healthy adults (mean age = 28.58 years, SD = 7.56; age range = 18-46 years; 75 females) to facilitate investigation of the neural mechanisms underlying working memory performance in humans. Across two days of assessment, participants underwent a battery of cognitive tasks assessing working memory and other cognitive domains (day 1), followed by 62-channel EEG recordings (day 2). EEG data were acquired at rest, during a visual working memory task, and following single-pulse transcranial magnetic stimulation (TMS-EEG) targeting key brain regions involved in working memory. This article provides a detailed overview of the study design, methodology, and data characteristics.

neuroscience↗

bioRxiv2026

Cortical geometry constrains the unimodal anchors of sensory integration

The human cerebral cortex is organized along a unimodal-to-transmodal hierarchy, which provides a putative substrate for the integration of sensory signals from primary cortical fields with information from other modalities. Diverse structural and functional properties of the cortex, including myelination, gene expression, neurodevelopmental timing, and inter-regional functional connectivity are patterned along this hierarchy. One exception to this hierarchy are the dominant eigenmodes of cortical geometry, which are instead patterned along rostrocaudal, mediolateral, and dorsoventral axes, each anchored by a primary sensory area at one extreme. Recent work has reconstructed the unimodal-to-transmodal hierarchy by integrating seed-based functional connectivity from three primary sensory areas, suggesting hierarchical organization may be driven by converging sensory input. Although geometric eigenmodes do not directly express the unimodal-to-transmodal hierarchy, they may encode modality-specific sensory organization originating from primary areas. Using MRI data from the Human Connectome Project, we tested whether geometric eigenmodes encode sensory function by modelling multisensory integration directly from cortical geometry. Specifically, we substituted functional connectivity maps from each primary sensory area with the rostrocaudal, mediolateral, and dorsoventral geometric eigenmodes, following a previously validated sensory integration mapping framework. Each geometric eigenmode corresponded to a distinct sensory domain (rostrocaudal-visual: |r| = 0.516; mediolateral-somatosensory: |r| = 0.551; dorsoventral-auditory: |r| = 0.342). Together, the three geometric eigenmodes created a mapping space that differentiated between unimodal brain regions with similar accuracy as fMRI-based models ({delta} = 64.74{degrees}; p < .001); however, differences between geometric and functional maps were largest within the transmodal association cortex. Reproducing the full unimodal-to-transmodal hierarchy required additional higher-frequency geometric eigenmodes (15 eigenmodes: r2 = 0.64). These findings suggest that the unimodal anchors of sensory integration are shaped by cortical geometry, with low-frequency geometric eigenmodes providing a structural basis for sensory organization, while the transmodal apex requires additional structural information to emerge.

neuroscience↗

bioRxiv2026

Regional heterogeneity shapes macroscopic wave dynamics of the human and non-human primate cortex

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.

neuroscience↗

bioRxiv2025

Cumulative timing-dependent changes in corticospinal excitability during suprathreshold paired-pulse transcranial magnetic stimulation

Transcranial magnetic stimulation (TMS) is widely used to assess inhibitory and facilitatory circuits within the primary motor cortex. However, accumulating evidence suggests that even brief TMS paradigms may induce unintended changes in corticospinal excitability. Here, we examined whether suprathreshold paired-pulse TMS delivered at inter-stimulus intervals (ISIs) associated with intracortical facilitation or long-interval cortical inhibition (LICI) elicits cumulative changes in motor-evoked potential (MEP) amplitude. In experiment 1, we reanalysed data from 17 participants who received 20 suprathreshold paired-pulses at eight ISIs (10-200 ms) and 40 unconditioned single pulses. Stimulation was pseudo-randomised and distributed evenly across four blocks. A linear mixed-effects model assessed trial-wise changes in MEP amplitude across ISIs. We replicated the design in an independent sample (n=10, experiment 2). A significant trial-by-ISI interaction was observed in both cohorts. Specifically, MEP amplitudes increased across trials for ISIs of 20 and 30 ms (p<0.05), but remained stable at LICI-related ISIs (100-150 ms). A similar increase was also seen with single-pulse TMS. These findings demonstrate that suprathreshold paired-pulse TMS at short ISIs can cumulatively enhance corticospinal excitability during stimulation. Furthermore, the results suggest potential for using these protocols not just for probing cortical circuits, but also as interventions to modulate motor system excitability.

neuroscience↗

bioRxiv2025

Geometric constraints on the architecture of mammalian cortical connectomes

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.

neuroscience↗

bioRxiv2025

A unified model of cortico-hippocampal interactions through neural field theory

Numerous physical systems evolve on curved manifolds whose geometry constrains their dynamics. The human brain provides a canonical example, with large-scale neural activity evolving on distinct anatomical surfaces such as the cortex and hippocampus. Within each structure, intrinsic feedback loops and manifold geometry shape characteristic neural rhythms. However, core cognitive functions of the brain arise from reciprocal interactions between these spatially distinct neural structures. Here, we introduce a general framework for geometry-constrained coupling between spatially extended dynamical systems evolving on separate manifolds. Using quasi-conformal mapping, we construct neighbourhood-preserving interactions that enable continuous neural activity on distinct geometries to interact while preserving local neighbourhood structure. Applying this framework to neural activity evolving on cortical and hippocampal surfaces, we show that increasing inter-manifold coupling reorganises system dynamics, producing frequency shifts, mode interactions, and coupling-driven instabilities consistent with critical transitions. These effects reproduce key features of large-scale brain activity, including topographically organised synchronisation between cortex and hippocampus during healthy cognition and critical transitions to seizure-like spectral dynamics. These results identify inter-manifold coupling as a core influence on dynamics in the human brain and highlight the broader role of geometry-constrained coupling on emergent dynamics in complex spatially extended systems.

neuroscience↗

bioRxiv2025

An Evaluation of the Efficacy of Single-Echo and Multi-Echo fMRI Denoising Strategies

Resting-state functional magnetic resonance imaging (rsfMRI) is commonly used to study brain-wide patterns of inter-regional functional coupling (FC). However, the resulting signals are vulnerable to multiple sources of noise, such as those related to non-neuronal physiological fluctuations and head motion, which can alter FC estimates and influence their associations with behavioral outcomes. The best strategy for acquiring and processing rsfMRI data to mitigate noise remains an open question. In this study of 358 healthy individuals, we compared the denoising efficacy of 60 multi-echo (ME) and 30 single-echo (SE) rsfMRI preprocessing pipelines across six distinct measures of data quality. We also evaluated how each pipeline influences the effect sizes of FC-based predictive models of personality and cognitive measures estimated via cross-validated kernel ridge regression. We found that ME pipelines generally showed superior denoising efficacy to SE pipelines, but that no single pipeline was associated with both superior denoising efficacy and behavioural prediction. Using a heuristic scheme to rank pipelines across benchmark evaluations, we found that an ME acquisition combined with Automatic Removal of Motion Artifacts Independent Component Analysis (ICA-AROMA) and Regressor Interpolation at Progressive Time Delays (RIPTiDe) offered a reasonable compromise between denoising efficacy and brain-behavior predictions for both ME and SE data. In general, ME pipelines ranked more highly than SE pipelines. These findings support the use of ME acquisitions in future work but suggest that no single denoising pipeline should be considered optimal for all purposes.

neuroscience↗

bioRxiv2025

The Motion Sensitivity and Predictive Utility of Different Estimates of Inter-regional Functional Coupling in Resting-state Functional MRI.

Numerous methods exist for quantifying statistical dependencies, termed functional coupling (FC), between regional brain activity recorded with resting-state functional magnetic resonance imaging (rs-fMRI). However, their efficacy in mitigating the effects of known sources of noise, such as those induced by participant head motion, and in augmenting effect sizes for brain-wide association studies (BWAS), remains unclear. Here we compared 10 different measures of FC, including correlations, partial correlations, coherence, mutual information, and partial information decomposition, and one measure of effective connectivity (EC; regression dynamic causal modelling), across two independent datasets comprising a total of 1,797 participants (867 males). Each method was evaluated for its ability to mitigate motion-related confounds in FC/EC estimates and for its utility in predicting 94 behavioural measures, as assessed via cross-validated kernel ridge regression. Our analyses showed that EC was most resistant to motion artifacts but had the weakest behavioral predictions. Conversely, traditional correlation-based methods showed the highest sensitivity to motion, but offered the strongest behavioral prediction across most domains and datasets. Nonetheless, relative differences in predictive accuracies were small, indicating that the use of different FC or EC metrics in rs-fMRI does not significantly impact BWAS effect sizes.

neuroscience↗

bioRxiv2025

Competing Programs Shape Cortical Sensorimotor-Association Axis Development

The neocortex is organized along a dominant sensorimotor-to-association (S-A) axis, anchored by modality-specific primary sensorimotor areas at one end and transmodal association areas that form distributed networks supporting abstract cognition at the other. The developmental mechanisms shaping this axis remain elusive. Here, we present converging multispecies evidence supporting the Multinodal Induction-Exclusion in Network Development (MIND) model, in which S-A patterning is governed by competing processes of induction and exclusion, driven by opposing transcriptomically-defined identity programs emerging from different nodes. Key molecular and connectional features of association cortices arise through pericentral programs, originating around fronto-temporal poles and partially regulated by retinoic acid. They progress inward toward central territories of the naive neocortex along fronto-temporally polarized trajectories. Central programs are induced through interactions between topographically separated first-order sensorimotor thalamocortical inputs and the neocortex, promoting the formation of primary areas while excluding pericentral programs. Influenced by SATB2 and ZBTB18, these evolutionarily conserved programs compete for the same territory and create spatial compartmentalization of axon guidance, cell-cell adhesion, retinoic acid signaling, synaptogenesis, Wnt signaling, and autism risk genes. Notably, PLXNC1 and SEMA7A exhibit anti-correlated expression and repulsive functions in shaping cortico-cortical connectivity along the S-A axis. These processes of induction and exclusion establish an S-A equilibrium and topography in which primary sensorimotor areas emerge as focal islands within the broader ocean of distributed associative networks. The MIND model provides a unifying framework for understanding experimental, evolutionary, and clinical phenomena, revealing induction and exclusion as antagonistic complementary principles shaping the S-A axis and processing hierarchies.

neuroscience↗

bioRxiv2025

Spectral imprint of structural embedding in effective connectivity

Neural fluctuations exhibit rich spectral profiles that reflects both local dynamics and structural (or anatomical) embedding. Yet, standard models of resting-state effective connectivity neglect structural embedding and assume uniformity in the timescales of regions endogenous fluctuations. We introduce a chromatic dynamic causal model (DCM) in which structural valency (or degree) modulates the spectral color of endogenous fluctuations. Specifically, we assume a linear mapping between regional structural valency and the spectral exponent of scale-free auto-spectra. Simulations show this mapping can emerge as a generic consequence of structural embedding under minimal coupling in a non-equilibrium regime. We show chromatic DCM reliably recovers ground-truth parameters across network sizes and noise conditions, outperforming standard spectral DCM. Applied to empirical data, chromatic DCM reveals that valency-exponent mappings vary across a cortical hierarchy, and that its parameters are conserved across a homologous network in humans, macaques, marmosets, and mice. These findings advance a generative account of structure-function coupling and expand the repertoire of biophysical mechanisms available for inference in effective connectivity modeling.

neuroscience↗

bioRxiv2025

Mechanisms of long term non-reinforced preference change: functional connectivity changes in a longitudinal functional MRI study

Behavioral change studies mostly focus on external reinforcements to modify preferences. Cue-approach training (CAT) is a paradigm that influences preferences by the mere association of stimuli, sensory cues, and a rapid motor response, without external reinforcements. The behavioral effect has been shown to last for months after less than one hour of training. Here, we used a modified version of CAT by changing the neutral-cue to a number that represented a monetary amount of reward that the participants accumulate (i.e. incentive-cue). After a single training session, we compared behavioral performance and functional connectivity (FC), as measured using functional magnetic resonance imaging, between two groups, one receiving a neutral-cue and the other receiving an incentive-cue, at 5 time points across one year. We replicated the maintenance of behavioral changes after 6-months for the non-reinforced neutral-cue participants, but not for the reinforced group. The reinforced training group showed higher FC within the limbic system, whereas the non-externally reinforced group showed higher functional connectivity within and between default-mode and dorsal-attention networks. Our findings offer putative neural correlates for both reinforced and non-reinforced preference changes that are maintained over time and which could be implemented in future behavioral change interventions. Significance StatementThe current work examines the neural mechanisms of non-externally reinforced preference change and its maintenance over time, using both a neutral cue and a modified version of the cue-approach training paradigm. While both groups initially exhibited preference shifts, only the non-reinforced group maintained preference changes over one year, suggesting an enduring internal reinforcement mechanism. We identified distinct patterns in functional connectivity related to behavioral maintenance, find that non-external reinforcement increases the connectivity of the default mode network while external reinforcement elevates connectivity between limbic areas. These findings enhance our understanding of sustainable behavior change and advocate for non-external reinforcement in behavioral interventions.

neuroscience↗

bioRxiv2025

Benchmarking overlapping community detection methods for applications in human connectomics

Brain networks exhibit non-trivial modular organization, with groups of densely connected areas participating in specialized functions. Traditional community detection algorithms assign each node to one module, but this representation cannot capture integrative, multi-functional nodes that span multiple communities. Despite the increasing availability of overlapping community detection algorithms (OCDAs) to capture such integrative nodes, there is no objective procedure for selecting the most appropriate method and its parameters for a given problem. Here we overcome this limitation by introducing a datadriven method for selecting an OCDA and its parameters from performance on a tailored ensemble of generated benchmark networks, assessing 22 unique algorithms and parameter settings. Applied to the human structural connectome, we find that the Order Statistics Local Optimization Method (OSLOM) best identifies ground-truth overlapping structure in the benchmark ensemble and yields a seven-network decomposition of the human cortex. These modules are bridged by fifteen overlapping regions that generally sit at the apex of the putative cortical hierarchy--suggesting integrative, higher-order function-- with network participation increasing along the cortical hierarchy, a finding not supported using a non-overlapping modular decomposition. This data-driven approach to selecting OCDAs is applicable across domains, opening new avenues to detecting and quantifying informative structures in complex real-world networks.

neuroscience↗

bioRxiv2025

Geometric influences on the regional organization of the mammalian brain

The mammalian brain comprises anatomically and functionally distinct regions, yet the principles governing their organization across scales remain unclear. Efforts to map regional boundaries use expert judgement or data-driven clustering of functional, connectional, and/or architectonic properties, but these approaches are often descriptive, have limited generalizability, and do not elucidate generative mechanisms. Here, we develop a geometrically constrained framework for the hierarchical, multiscale regional organization of neocortical and non-neocortical structures. The framework yields parcels at any resolution with high internal homogeneity across hundreds of anatomical, functional, cellular, and molecular properties in humans, macaques, marmosets, and mice, and it generalizes to understudied mammalian species. Finally, we demonstrate that the framework captures the essence of a reaction-diffusion mechanism in which brain geometry shapes the spatial expression of putative patterning molecules that establish a blueprint of regional identity during development. Our findings point to a conserved, universal influence of geometry on mammalian brain organization.

neuroscience↗

bioRxiv2024

The efficacy of resting-state fMRI denoising pipelines for motion correction and behavioural prediction.

Resting-state functional magnetic resonance imaging (rs-fMRI) is a pivotal tool for mapping the functional organization of the brain and its relation to individual differences in behaviour. One challenge for the field is that rs-fMRI signals are contaminated by multiple sources of noise that can contaminate these rs-fMRI signals, affecting the reliability and validity of any derivative phenotypes and attenuating their correlations with behaviour. Here, we investigate the efficacy of different noise mitigation pipelines, including white-matter and cerebrospinal fluid regression, independent component analysis (ICA) - based artefact removal, volume censoring, global signal regression (GSR), and diffuse cluster estimation and regression (DiCER), in simultaneously achieving two objectives: mitigating motion-related artifacts and augmenting brain-behaviour associations. Our analysis, which employed three distinct quality control metrics to evaluate motion influence and a kernel ridge regression for behavioural predictions of 81 different behavioural variables across two independent datasets, revealed that no single pipeline universally excels at achieving both objectives consistently across different cohorts. Pipelines combining ICA-FIX and GSR demonstrate a reasonable trade-off between motion reduction and behavioural prediction performance, but inter-pipeline variations in predictive performance are modest.

neuroscience↗

bioRxiv2024

Coming up short: generative network models fail to accurately capture long-range connectivity

Generative network models (GNMs) have been proposed to identify the mechanisms/constraints that shape the organisation of the connectome. These models parameterise the formation of inter-regional connections using a trade-off between connection cost and topological complexity or biophysical similarity. Despite their simplicity, GNMs can generate synthetic networks that capture many topological properties of empirical brain networks. However, current models often fail to capture the topography (i.e., spatial embedding) of many such properties, such as the anatomical location of network hubs. In this study, we investigate a diverse array of generative network model formulations and find that none can accurately capture empirical patterns of long-range connectivity. We demonstrate that the spatial embedding of longer-range connections is critical in defining hub locations and that it is precisely these connections that are poorly captured by extant models. We further show how standard measures used for model optimisation and evaluation mask these and other differences between synthetic and empirical brain networks, highlighting the need for care when interpreting generative network models and metrics. Overall, our findings demonstrate common failure modes of GNMs, identify why these models do not fully capture brain network organisation, and suggest ways the field can move forward to address these challenges. Author summaryGenerative network models aim to explain the organisation of connectomes using simple wiring rules. While these models replicate topological features of brain networks, they do not capture key topographical properties, like the anatomical location of network hubs. We show that this failure occurs because the models are unable to accurately capture the spatial position of long-range inter-regional connections. Moreover, standard evaluation measures fail to accurately quantify the similarity between model and empirical networks. This study identifies how and why limitations of current generative models occur and suggests ways forward for improved practices.

neuroscience↗