Biology subjects

Bethlehem, R. A. I.

Publications and source records attributed to Bethlehem, R. A. I..

54 records · Page 3Linked to original sources

bioRxiv2020

A Deep Graph Neural Network Architecture for Modelling Spatio-temporal Dynamics in resting-stating functional MRI Data

Resting-state functional magnetic resonance imaging (rs-fMRI) has been successfully employed to understand the organisation of the human brain. For rs-fMRI analysis, the brain is typically parcellated into regions of interest (ROIs) and modelled as a graph where each ROI is a node and pairwise correlation between ROI blood-oxygen-level-dependent (BOLD) time series are edges. Recently, graph neural networks (GNNs) have seen a surge in popularity due to their successes in modelling unstructured relational data. The latest developments with GNNs, however, have not yet been fully exploited for the analysis of rs-fMRI data, particularly with regards to its spatio-temporal dynamics. Herein we present a novel deep neural network architecture, combining both GNNs and temporal convolutional networks (TCNs), which is able to learn from the spatial and temporal components of rs-fMRI data in an end-to-end fashion. In particular, this corresponds to intra-feature learning (i.e., learning temporal dynamics with TCNs) as well as inter-feature learning (i.e., leveraging spatial interactions between ROIs with GNNs). We evaluate our model with an ablation study using 35,159 samples from the UK Biobank rs-fMRI database. We also demonstrate explainability features of our architecture which map to realistic neurobiological insights. We hope our model could lay the groundwork for future deep learning architectures focused on leveraging the inherently and inextricably spatio-temporal nature of rs-fMRI data.

neuroscience↗

bioRxiv2020

Assessing the relationship between cognitive empathy, age, and face perception using a sequential Bayesian analysis

The ability to correctly identify emotions in social stimuli such as faces is proposed to affect ones level of cognitive empathy. The Fusiform Face Area shows a heightened neural response during the perception of faces relative to objects. We tested whether neural responses to social stimuli were associated with performance in a measure of cognitive empathy, the Reading the Mind in the Eyes Task. To quantify face perception, participants were presented with images during a fast serial presentation task which elicited a steady state visual evoked potential, measured using electroencephalography. A Sequential Bayesian Analysis was used to assess if face specific neural responses were associated with either cognitive empathy or age. Data were collected from a participant group of both neurotypical individuals and individuals on the autistic spectrum. We found no significant relationship between the face-specific neural signature, cognitive empathy or age. This study highlights the efficiency of the Sequential Bayesian Analysis as an effective method of participant recruitment.

neuroscience↗

bioRxiv2020

Atypical genomic patterning of the cerebral cortex in autism with poor early language outcome

Cortical regional identities develop through anterior-posterior (A-P) and dorsal-ventral (D-V) prenatal genomic patterning gradients. Here we find that A-P and D-V genomic patterning of cortical surface area (SA) and thickness (CT) is intact in typically developing and autistic toddlers with good language outcome, but is absent in autistic toddlers with poor early language outcome. Genes driving this effect are prominent in midgestational A-P and D-V gene expression gradients and prenatal cell types driving SA and CT variation (e.g., progenitor cells versus excitatory neurons). These genes are also important for vocal learning, human-specific evolution, and prenatal co-expression networks enriched for high-penetrance autism risk genes. Autism with poor early language outcome may be linked to atypical genomic cortical patterning starting in prenatal periods and which impacts later development of regional functional specialization and circuit formation. One Sentence SummaryGenomic patterning of the cortex is atypical in autistic toddlers with poor early language outcome.

neuroscience↗

bioRxiv2020

Sexually dimorphic development of depression-related brain networks during healthy human adolescence

We hypothesized that there are sexual differences in human brain network development underlying the female > male divergence in adolescent depression. We tested for sex differences in parameters of brain network development (accelerated longitudinal fMRI, N=298 healthy adolescents, each scanned 1 to 3 times). Sexually divergent development of functional connectivity was located in default mode network (DMN), limbic cortex, and subcortical nuclei. Females had a more "disruptive" pattern of development, where weak functional connectivity at age 14 became stronger during adolescence. This fMRI-derived map of divergent adolescent development was co-located with (i) a map of functional dysconnectivity associated with adult major depressive disorder (MDD); and (ii) an adult brain gene expression pattern enriched for genes on the X chromosome, neurodevelopmental genes, and risk genes for MDD. Sexual divergence in disruptive development of DMN, limbic and subcortical functional networks is potentially relevant to the increased risk of depression in adolescent females.

neuroscience↗

bioRxiv2020

Population GNNs for Brain Age Prediction

Many common neurological and neurodegenerative disorders, such as Alzheimers disease, dementia and multiple sclerosis, have been associated with abnormal patterns of apparent ageing of the brain. Discrepancies between the estimated brain age and the actual chronological age (brain age gaps) can be used to understand the biological pathways behind the ageing process, assess an individuals risk for various brain disorders and identify new personalised treatment strategies. By flexibly integrating minimally preprocessed neuroimaging and non-imaging modalities into a population graph data structure, we train two types of graph neural network (GNN) architectures to predict brain age in a clinically relevant fashion as well as investigate their robustness to noisy inputs and graph sparsity. The multimodal population graph approach has the potential to learn from the entire cohort of healthy and affected subjects of both sexes at once, capturing a wide range of con-founding effects and detecting variations in brain age trends between different sub-populations of subjects.

neuroscience↗

bioRxiv2020

Macroscale connectome manifold expansion in adolescence

AO_SCPLOWBSTRACTC_SCPLOWAdolescence is a critical time for the continued maturation of brain networks. Here, we assessed structural connectome development in a large longitudinal sample ranging from childhood to young adulthood. By projecting high-dimensional connectomes into compact manifold spaces, we identified a marked expansion of structural connectomes with the strongest effects in transmodal regions during adolescence. Findings reflected increased within-module connectivity together with increased segregation, indicating increasing differentiation of higher-order association networks from the rest of the brain. Projection of subcortico-cortical connectivity patterns into these manifolds showed parallel alterations in pathways centered on the caudate and thalamus. Connectome findings were contextualized via spatial transcriptome association analysis, highlighting genes enriched in cortex, thalamus, and striatum. Statistical learning of cortical and subcortical manifold features at baseline and their maturational change predicted measures of intelligence at follow-up. Our findings demonstrate that connectome manifold learning can bridge the conceptual and empirical gaps between macroscale network reconfigurations, microscale processes, and cognitive outcomes in adolescent development. IO_SCPLOWMPACTC_SCPLOWManifold learning of longitudinal brain network data provides novel insights into adolescent structural connectome maturation, and how multiple scales of cortical and subcortical organization interact in typical neurodevelopment.

neuroscience↗

bioRxiv2020

Connectome and microcircuit models implicate atypical subcortico-cortical interactions in autism pathophysiology

AO_SCPLOWBSTRACTC_SCPLOWBoth macroscale connectome miswiring and microcircuit anomalies have been suggested to play a role in the pathophysiology of autism. However, an overarching framework that consolidates these macro and microscale perspectives of the condition is lacking. Here, we combined connectome-wide manifold learning and biophysical simulation models to understand associations between global network perturbations and microcircuit dysfunctions in autism. Our analysis established that autism showed significant differences in structural connectome organization relative to neurotypical controls, with strong effects in low-level somatosensory regions and moderate effects in high-level association cortices. Computational models revealed that the degree of macroscale anomalies was related to atypical increases of subcortical inputs into cortical microcircuits, especially in sensory and motor areas. Transcriptomic decoding and developmental gene enrichment analyses provided biological context and pointed to genes expressed in cortical and thalamic areas during childhood and adolescence. Supervised machine learning showed the macroscale perturbations predicted socio-cognitive symptoms and repetitive behaviors. Our analyses provide convergent support that atypical subcortico-cortical interactions may contribute to both microcircuit and macroscale connectome anomalies in autism.

neuroscience↗

bioRxiv2020

Differential functional neural circuitry behind autism subtypes with marked imbalance between social-communicative and restricted repetitive behavior symptom domains

Social-communication (SC) and restricted repetitive behaviors (RRB) are autism diagnostic symptom domains. SC and RRB severity can markedly differ within and between individuals and may be underpinned by different neural circuitry and genetic mechanisms. Modeling SC-RRB balance could help identify how neural circuitry and genetic mechanisms map onto such phenotypic heterogeneity. Here we developed a phenotypic stratification model that makes highly accurate (97-99%) out-of-sample SC=RRB, SC>RRB, and RRB>SC subtype predictions. Applying this model to resting state fMRI data from the EU-AIMS LEAP dataset (n=509), we find that while the phenotypic subtypes share many commonalities in terms of intrinsic functional connectivity, they also show replicable differences within some networks compared to a typically-developing group (TD). Specifically, the somatomotor network is hypoconnected with perisylvian circuitry in SC>RRB and visual association circuitry in SC=RRB. The SC=RRB subtype show hyperconnectivity between medial motor and anterior salience circuitry. Genes that are highly expressed within these networks show a differential enrichment pattern with known autism-associated genes, indicating that such circuits are affected by differing autism-associated genomic mechanisms. These results suggest that SC-RRB imbalance subtypes share many commonalities, but also express subtle differences in functional neural circuitry and the genomic underpinnings behind such circuitry.

neuroscience↗

bioRxiv2020

Shared and distinct patterns of atypical cortical morphometry in children with autism and anxiety

Autism spectrum disorder (ASD) and anxiety disorders (ANX) are common neurodevelopmental conditions with several overlapping symptoms. Notably, many children and adolescents with ASD also have an ANX diagnosis, suggesting shared pathological mechanisms. Here, we leveraged structural imaging and phenotypic data from 82 closely matched children (28 ASD, 28 ANX, 26 typically developing controls). Our neuroimaging paradigm assessed cortical thickness and studied inter-regional structural covariance networks. Both ASD and ANX presented with atypical structural network organization relative to controls. Specifically, ASD presented with increased thickness in temporal and parietal midline cortices, while ANX was associated with increased cortical thickness in the left inferior frontal and precentral gyri. Despite the overall difference in the spatial distributions these clusters, unconstrained spin permutation testing showed that statistical maps from the ANX-vs-controls and ASD-vs-controls analyses were significantly correlated. The two diagnostic groups also presented with common decreases in structural covariance patterns, collectively pointing to decreased structural coupling between lateral frontal, lateral temporal, and temporo-parietal regions. Dimensional analysis of trait anxiety and social responsiveness partially recapitulated diagnosis-based findings. Collectively, our findings provide evidence for both shared as well as distinct effects of ASD and ANX on on regional and inter-regional structural network organization.

neuroscience↗

bioRxiv2020

Dispersion of functional gradients across the lifespan

Ageing is commonly associated with changes to segregation and integration of functional brain networks, but, in isolation, current network-based approaches struggle to elucidate changes across the many axes of functional organisation. However, the advent of gradient mapping techniques to neuroimaging provides a new means of studying functional organisation in a multi-dimensional connectivity space. Here, we studied ageing and behaviourally-relevant differences in a three-dimensional connectivity space using the Cambridge Centre for Ageing Neuroscience cohort (n=643). Building on gradient mapping techniques, we developed a set of measures to quantify the dispersion within and between functional communities. We detected a strong shift of the visual network across the lifespan from an extreme to a more central position in the 3D gradient space. In contrast, the dispersion distance of transmodal communities (dorsal attention, ventral attention, frontoparietal and default mode) did not change. However, these communities were increasingly dispersed with increasing age, reflecting more dissimilar functional connectivity profiles within each community. Increasing dispersion of frontoparietal and ventral attention networks, in particular, was associated negatively with cognition, measured by fluid intelligence. By using a technique that explicitly captures the ordering of functional systems in a multi-dimensional hierarchical framework, we identified behaviorally-relevant age-related differences of within and between network organisation. We propose that the study of functional gradients across the lifespan could provide insights that may facilitate the development of new strategies to maintain cognitive ability across the lifespan in health and disease.

neuroscience↗

bioRxiv2020

The cortical wiring scheme of hierarchical information processing

AO_SCPLOWBSTRACTC_SCPLOWThe vast net of fibres within and underneath the cortex is optimised to support the convergence of different levels of brain organisation. Here we propose a novel coordinate system of the human cortex based on an advanced model of its connectivity. Our approach is inspired by seminal, but so far largely neglected models of cortico-cortical wiring established by post mortem anatomical studies and capitalizes on cutting-edge neuroimaging and machine learning. The new model expands the currently prevailing diffusion MRI tractography approach by incorporation of additional features of cortical microstructure and cortico-cortical proximity. Studying several datasets, we could show that our coordinate system robustly recapitulates established sensory-limbic and anterior-posterior dimensions of brain organisation. A series of validation experiments showed that the new wiring space reflects cortical microcircuit features (including pyramidal neuron depth and glial expression) and allowed for competitive simulations of functional connectivity and dynamics across a broad range contexts (based on resting-state fMRI, task-based fMRI, and human intracranial EEG coherence). Our results advance our understanding of how cell-specific neurobiological gradients produce a hierarchical cortical wiring scheme that is concordant with increasing functional sophistication of human brain organisation. Our evaluations demonstrate the cortical wiring space bridges across scales of neural organisation and can be easily translated to single individuals. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=148 SRC="FIGDIR/small/899583v4_ufig1.gif" ALT="Figure 1"> View larger version (30K): org.highwire.dtl.DTLVardef@1b1c601org.highwire.dtl.DTLVardef@3e76c7org.highwire.dtl.DTLVardef@e828e5org.highwire.dtl.DTLVardef@1d4f8ce_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2019

Novel genome-wide associations for anhedonia, genetic correlation with psychiatric disorders, and polygenic association with brain structure.

Anhedonia is a core feature of several psychiatric disorders but its biological underpinnings are poorly understood. We performed a genome-wide association study of anhedonia in 375,275 UK Biobank participants and assessed for genetic correlation between anhedonia and neuropsychiatric conditions (major depressive disorder, schizophrenia, bipolar disorder, obsessive compulsive disorder and Parkinsons Disease). We then used a polygenic risk score approach to test for association between genetic loading for anhedonia and both brain structure and brain function. This included: magnetic resonance imaging (MRI) assessments of total grey matter volume, white matter volume, cerebrospinal fluid volume, and 15 cortical/subcortical regions of interest; diffusion tensor imaging (DTI) measures of white matter tract integrity; and functional MRI activity during an emotion processing task. We identified 11 novel loci associated at genome-wide significance with anhedonia, with a SNP heritability estimate (h2SNP) of 5.6%. Strong positive genetic correlations were found between anhedonia and major depressive disorder, schizophrenia and bipolar disorder; but not with obsessive compulsive disorder or Parkinsons Disease. Polygenic risk for anhedonia was associated with poorer brain white matter integrity, smaller total grey matter volume, and smaller volumes of brain regions linked to reward and pleasure processing, including nucleus accumbens, caudate and medial frontal cortex. In summary, the identification of novel anhedonia-associated loci substantially expands our current understanding of the biological basis of anhedonia and genetic correlations with several psychiatric disorders confirm the utility of this trait as a transdiagnostic marker of vulnerability to mental illness. We also provide the first evidence that genetic risk for anhedonia influences brain structure, particularly in regions associated with reward and pleasure processing.

genetics↗

bioRxiv2019

Transcriptomic and Cellular Decoding of Regional Brain Vulnerability to Neurodevelopmental Disorders

Neurodevelopmental disorders are highly heritable and associated with spatially-selective disruptions of brain anatomy. The logic that translates genetic risks into spatially patterned brain vulnerabilities remains unclear but is a fundamental question in disease pathogenesis. Here, we approach this question by integrating (i) in vivo neuroimaging data from patient subgroups with known causal genomic copy number variations (CNVs), and (ii) bulk and single-cell gene expression data from healthy cortex. First, for each of six different CNV disorders, we show that spatial patterns of cortical anatomy change in youth are correlated with spatial patterns of expression for CNV region genes in bulk cortical tissue from typically-developing adults. Next, by transforming normative bulk-tissue cortical expression data into cell-type expression maps, we further link each disorders anatomical change map to specific cell classes and specific CNV-region genes that these cells express. Finally, we establish convergent validity of this "transcriptional vulnerability model" by inter-relating patient neuroimaging data with measures of altered gene expression in both brain and blood-derived patient tissue. Our work clarifies general biological principles that govern the mapping of genetic risks onto regional brain disruption in neurodevelopmental disorders. We present new methods that can harness these principles to screen for potential cellular and molecular determinants of disease from readily available patient neuroimaging data.

neuroscience↗

bioRxiv2018

Autistic traits, resting-state connectivity and absolute pitch in professional musicians: shared and distinct neural features

BackgroundRecent studies indicate increased autistic traits in musicians with absolute pitch and a higher incidence of absolute pitch in people with autism. Theoretical accounts connect both of these with shared neural principles of local hyper- and global hypoconnectivity, enhanced perceptual functioning and a detail-focused cognitive style. This is the first study to investigate absolute pitch proficiency, autistic traits and brain correlates in the same study.\n\nSample and MethodsGraph theoretical analysis was conducted on resting state (eyes closed and eyes open) EEG connectivity (wPLI, weighted Phase Lag Index) matrices obtained from 31 absolute pitch (AP) and 33 relative pitch (RP) professional musicians. Small Worldness, Global Clustering Coefficient and Average Path length were related to autistic traits, passive (tone identification) and active (pitch adjustment) absolute pitch proficiency and onset of musical training using Welch-two-sample-tests, correlations and general linear models.\n\nResultsAnalyses revealed increased Path length (delta 2-4 Hz), reduced Clustering (beta 13-18 Hz), reduced Small-Worldness (gamma 30-60 Hz) and increased autistic traits for AP compared to RP. Only Clustering values (beta 13-18 Hz) were predicted by both AP proficiency and autistic traits. Post-hoc single connection permutation tests among raw wPLI matrices in the beta band (13-18 Hz) revealed widely reduced interhemispheric connectivity between bilateral auditory related electrode positions along with higher connectivity between F7-F8 and F8-P9 for AP. Pitch naming ability and Pitch adjustment ability were predicted by Path length, Clustering, autistic traits and onset of musical training (for pitch adjustment) explaining 44% respectively 38% of variance.\n\nConclusionsResults show both shared and distinct neural features between AP and autistic traits. Differences in the beta range were associated with higher autistic traits in the same population. In general, AP musicians exhibit a widely underconnected brain with reduced functional integration and reduced small-world-property during resting state. This might be partly related to autism-specific brain connectivity, while differences in Path length and Small-Worldness reflect other ability-specific influences. This is further evidence for different pathways in the acquisition and development of absolute pitch, likely influenced by both genetic and environmental factors and their interaction.

neuroscience↗

bioRxiv2018

Using normative age modelling to isolate subsets of individuals with autism expressing highly age-atypical cortical thickness features

Understanding heterogeneity in neural phenotypes is an important goal on the path to precision medicine for autism spectrum disorders (ASD). Age is a critically important variable in normal structural brain development and examining structural features with respect to age-related norms could help to explain ASD heterogeneity in neural phenotypes. Here we examined how cortical thickness (CT) in ASD can be parameterized as an individualized metric of deviance relative to typically-developing (TD) age-related norms. Across a large sample (n=870 per group) and wide age range (5-40 years), we applied a normative modelling approach that provides individualized whole-brain maps of age-related CT deviance in ASD. This approach isolates a subgroup of ASD individuals with highly age-deviant CT. The median prevalence of this ASD subgroup across all brain regions is 7.6%, and can reach as high as 10% for some brain regions. This work showcases an individualized approach for understanding ASD heterogeneity that could potentially further prioritize work on a subset of individuals with significant cortical pathophysiology represented in age-related CT deviance. Rather than cortical thickness pathology being a widespread characteristic of most ASD patients, only a small subset of ASD individuals are actually highly deviant relative to age-norms. These individuals drive small on-average effects from case-control comparisons. Rather than sticking to the diagnostic label of autism, future research should pivot to focus on isolating subsets of autism patients with highly deviant phenotypes and better understand the underlying mechanisms that drive those phenotypes.

neuroscience↗

bioRxiv2017

Synaptic and transcriptionally downregulated genes are associated with cortical thickness differences in autism

Differences in cortical morphology - in particular, cortical volume, thickness and surface area - have been reported in individuals with autism. However, it is unclear what aspects of genetic and transcriptomic variation are associated with these differences. Here we investigate the genetic correlates of global cortical thickness differences ({Delta}CT) in children with autism. We used Partial Least Squares Regression (PLSR) on structural MRI data from 548 children (166 with autism, 295 neurotypical children and 87 children with ADHD) and cortical gene expression data from the Allen Institute for Brain Science to identify genetic correlates of {Delta}CT in autism.\n\nWe identify that these genes are enriched for synaptic transmission pathways and explain significant variation in {Delta}CT. These genes are also significantly enriched for genes dysregulated in the autism post-mortem cortex (Odd Ratio (OR) = 1.11, Pcorrected < 10-14), driven entirely by downregulated genes (OR = 1.87, Pcorrected < 10-15). We validated the enrichment for downregulated genes in two independent datasets: Validation 1 (OR = 1.44, Pcorrected = 0.004) and Validation 2 (OR = 1.30; Pcorrected = 0.001). We conclude that transcriptionally downregulated genes implicated in autism are robustly associated with global changes in cortical thickness variability in children with autism.

neuroscience↗

bioRxiv2017

Structural covariance networks in children with autism or ADHD

While autism and attention-deficit/hyperactivity disorder (ADHD) are considered distinct conditions from a diagnostic perspective, they share some phenotypic features and have high comorbidity. Taking a dual-condition approach might help elucidate shared and distinct neural characteristics.\n\nGraph theory was used to analyse properties of cortical thickness structural covariance networks across both conditions and relative to a neurotypical (NT; n=87) group using data from the ABIDE (autism; n=62) and ADHD-200 datasets (ADHD; n=69). This was analysed in a theoretical framework examining potential differences in long and short range connectivity.\n\nWe found convergence between autism and ADHD, where both conditions show an overall decrease in CT covariance with increased Euclidean distance compared to a neurotypical population. The two conditions also show divergence: less modular overlap between the two conditions than there is between each condition and the neurotypical group. Lastly, the ADHD group also showed reduced wiring costs compared to the autism groups.\n\nOur results indicate a need for taking an integrated approach when considering highly comorbid conditions such as autism and ADHD. Both groups show a distance-covariance relation that more strongly favours short-range over long-range. Thus, on some network features the groups seem to converge, yet on others there is divergence.

neuroscience↗

bioRxiv2016

Genetic overlap between educational attainment, schizophrenia and autism

ImportanceThe genetic relationship between cognition, autism, and schizophrenia is complex. It is unclear how genes that contribute to cognition also contribute to risk for autism and schizophrenia.\n\nObjectiveTo investigate the interaction between genes related to cognition (measured via proxy through educational attainment, which we call edu genes) and genes/biological pathways that are atypical in autism and schizophrenia.\n\nDesignGenetic correlation and enrichment analysis were conducted to identify the interaction between edu genes and risk genes and biological pathways for autism or schizophrenia.\n\nResultsFirst, edu genes are enriched in a specific developmental co-expression module that is also enriched for high confidence autism risk genes. Second, modules enriched for genes that are dysregulated in autism and schizophrenia are also enriched for edu genes. Finally, genes that overlap between the two above modules and educational attainment are significantly enriched for genes that flank human accelerated regions, suggesting increased positive selection for the overlapping gene sets.\n\nConclusionOur results identify distinct co-expression modules where risk genes for the two psychiatric conditions interact with edu genes. This suggests specific pathways that contribute to both cognitive deficits and cognitive talents, in individuals with schizophrenia or autism.\n\nKey PointsO_ST_ABSQuestionC_ST_ABSHow do genes for educational attainment interact with risk genes for autism and schizophrenia?\n\nFindingsWe show that genes for educational attainment (edu genes) are significantly likely to be mutated in autism and intellectual disability. We further show that edu genes also interact with co-expression modules that are associated with autism or schizophrenia and are enriched for differentially expressed genes in autism or schizophrenia. Finally, we identify that the enrichment between risk genes for autism and schizophrenia and human accelerated regions are driven, in part, by their overlap with edu genes.\n\nMeaningEdu genes interact with schizophrenia and autism risk genes in specific pathways, contributing to both cognitive deficits and talents.

genetics↗