Graph Convolutional Networks Reveal Network-Level Functional Dysconnectivity in Schizophrenia

Exploring Schizophrenia Classification Through Multimodal MRI and Deep Graph Neural Networks: Unveiling Brain Region-Specific Weight Discrepancies and Their Association With Cell-Type Specific Transcriptomic Features

BACKGROUND AND HYPOTHESIS

Schizophrenia (SZ) is a prevalent mental disorder that imposes significant health burdens. Diagnostic accuracy remains challenging due to clinical subjectivity. To address this issue, we explore magnetic resonance imaging (MRI) as a tool to enhance SZ diagnosis and provide objective references and biomarkers. Using deep learning with graph convolution, we represent MRI data as graphs, aligning with brain structure, and improving feature extraction, and classification. Integration of multiple modalities is expected to enhance classification.

STUDY DESIGN

Our study enrolled 683 SZ patients and 606 healthy controls from 7 hospitals, collecting structural MRI and functional MRI data. Both data types were represented as graphs, processed by 2 graph attention networks, and fused for classification. Grad-CAM with graph convolution ensured interpretability, and partial least squares analyzed gene expression in brain regions.

STUDY RESULTS

Our method excelled in the classification task, achieving 83.32% accuracy, 83.41% sensitivity, and 83.20% specificity in 10-fold cross-validation, surpassing traditional methods. And our multimodal approach outperformed unimodal methods. Grad-CAM identified potential brain biomarkers consistent with gene analysis and prior research.

CONCLUSIONS

Our study demonstrates the effectiveness of deep learning with graph attention networks, surpassing previous SZ diagnostic methods. Multimodal MRI's superiority over unimodal MRI confirms our initial hypothesis. Identifying potential brain biomarkers alongside gene biomarkers holds promise for advancing objective SZ diagnosis and research in SZ.

Supervised contrastive learning enhances graph convolutional networks for predicting neurodevelopmental deficits in very preterm infants using brain structural connectome

Very preterm (VPT) infants (born at less than 32 weeks gestational age) are at high risk for various adverse neurodevelopmental deficits. Unfortunately, most of these deficits cannot be accurately diagnosed until the age of 2-5 years old. Given the benefits of early interventions, accurate diagnosis and prediction soon after birth are urgently needed for VPT infants. Previous studies have applied deep learning models to learn the brain structural connectome (SC) to predict neurodevelopmental deficits in the preterm population. However, none of these models are specifically designed for graph-structured data, and thus may potentially miss certain topological information conveyed in the brain SC. In this study, we aim to develop deep learning models to learn the SC acquired at term-equivalent age for early prediction of neurodevelopmental deficits at 2 years corrected age in VPT infants. We directly treated the brain SC as a graph, and applied graph convolutional network (GCN) models to capture complex topological information of the SC. In addition, we applied the supervised contrastive learning (SCL) technique to mitigate the effects of the data scarcity problem, and enable robust training of GCN models. We hypothesize that SCL will enhance GCN models for early prediction of neurodevelopmental deficits in VPT infants using the SC. We used a regional prospective cohort of ∼280 VPT infants who underwent MRI examinations at term-equivalent age from the Cincinnati Infant Neurodevelopment Early Prediction Study (CINEPS). These VPT infants completed neurodevelopmental assessment at 2 years corrected age to evaluate cognition, language, and motor skills. Using the SCL technique, the GCN model achieved mean areas under the receiver operating characteristic curve (AUCs) in the range of 0.72∼0.75 for predicting three neurodevelopmental deficits, outperforming several competing models. Our results support our hypothesis that the SCL technique is able to enhance the GCN model in our prediction tasks.

Interpretable Cognitive Ability Prediction: A Comprehensive Gated Graph Transformer Framework for Analyzing Functional Brain Networks

Graph convolutional deep learning has emerged as a promising method to explore the functional organization of the human brain in neuroscience research. This paper presents a novel framework that utilizes the gated graph transformer (GGT) model to predict individuals' cognitive ability based on functional connectivity (FC) derived from fMRI. Our framework incorporates prior spatial knowledge and uses a random-walk diffusion strategy that captures the intricate structural and functional relationships between different brain regions. Specifically, our approach employs learnable structural and positional encodings (LSPE) in conjunction with a gating mechanism to efficiently disentangle the learning of positional encoding (PE) and graph embeddings. Additionally, we utilize the attention mechanism to derive multi-view node feature embeddings and dynamically distribute propagation weights between each node and its neighbors, which facilitates the identification of significant biomarkers from functional brain networks and thus enhances the interpretability of the findings. To evaluate our proposed model in cognitive ability prediction, we conduct experiments on two large-scale brain imaging datasets: the Philadelphia Neurodevelopmental Cohort (PNC) and the Human Connectome Project (HCP). The results show that our approach not only outperforms existing methods in prediction accuracy but also provides superior explainability, which can be used to identify important FCs underlying cognitive behaviors.

Elucidating salient site-specific functional connectivity features and site-invariant biomarkers in schizophrenia via deep neural networks

Schizophrenia is a highly heterogeneous disorder and salient functional connectivity (FC) features have been observed to vary across study sites, warranting the need for methods that can differentiate between site-invariant FC biomarkers and site-specific salient FC features. We propose a technique named Semi-supervised learning with data HaRmonisation via Encoder-Decoder-classifier (SHRED) to examine these features from resting state functional magnetic resonance imaging scans gathered from four sites. Our approach involves an encoder-decoder-classifier architecture that simultaneously performs data harmonisation and semi-supervised learning (SSL) to deal with site differences and labelling inconsistencies across sites respectively. The minimisation of reconstruction loss from SSL was shown to improve model performance even within small datasets whilst data harmonisation often led to lower model generalisability, which was unaffected using the SHRED technique. We show that our proposed model produces site-invariant biomarkers, most notably the connection between transverse temporal gyrus and paracentral lobule. Site-specific salient FC features were also elucidated, especially implicating the paracentral lobule for our local dataset. Our examination of these salient FC features demonstrates how site-specific features and site-invariant biomarkers can be differentiated, which can deepen our understanding of the neurobiology of schizophrenia.

Subject classification and cross-time prediction based on functional connectivity and white matter microstructure features in a rat model of Alzheimer's using machine learning

BACKGROUND

The pathological process of Alzheimer's disease (AD) typically takes decades from onset to clinical symptoms. Early brain changes in AD include MRI-measurable features such as altered functional connectivity (FC) and white matter degeneration. The ability of these features to discriminate between subjects without a diagnosis, or their prognostic value, is however not established.

METHODS

The main trigger mechanism of AD is still debated, although impaired brain glucose metabolism is taking an increasingly central role. Here, we used a rat model of sporadic AD, based on impaired brain glucose metabolism induced by an intracerebroventricular injection of streptozotocin (STZ). We characterized alterations in FC and white matter microstructure longitudinally using functional and diffusion MRI. Those MRI-derived measures were used to classify STZ from control rats using machine learning, and the importance of each individual measure was quantified using explainable artificial intelligence methods.

RESULTS

Overall, combining all the FC and white matter metrics in an ensemble way was the best strategy to discriminate STZ rats, with a consistent accuracy over 0.85. However, the best accuracy early on was achieved using white matter microstructure features, and later on using FC. This suggests that consistent damage in white matter in the STZ group might precede FC. For cross-timepoint prediction, microstructure features also had the highest performance while, in contrast, that of FC was reduced by its dynamic pattern which shifted from early hyperconnectivity to late hypoconnectivity.

CONCLUSIONS

Our study highlights the MRI-derived measures that best discriminate STZ vs control rats early in the course of the disease, with potential translation to humans.

Topological Data Analysis for Multivariate Time Series Data

Over the last two decades, topological data analysis (TDA) has emerged as a very powerful data analytic approach that can deal with various data modalities of varying complexities. One of the most commonly used tools in TDA is persistent homology (PH), which can extract topological properties from data at various scales. The aim of this article is to introduce TDA concepts to a statistical audience and provide an approach to analyzing multivariate time series data. The application's focus will be on multivariate brain signals and brain connectivity networks. Finally, this paper concludes with an overview of some open problems and potential application of TDA to modeling directionality in a brain network, as well as the casting of TDA in the context of mixed effect models to capture variations in the topological properties of data collected from multiple subjects.

The Combination of a Graph Neural Network Technique and Brain Imaging to Diagnose Neurological Disorders: A Review and Outlook

Neurological disorders (NDs), such as Alzheimer's disease, have been a threat to human health all over the world. It is of great importance to diagnose ND through combining artificial intelligence technology and brain imaging. A graph neural network (GNN) can model and analyze the brain, imaging from morphology, anatomical structure, function features, and other aspects, thus becoming one of the best deep learning models in the diagnosis of ND. Some researchers have investigated the application of GNN in the medical field, but the scope is broad, and its application to NDs is less frequent and not detailed enough. This review focuses on the research progress of GNNs in the diagnosis of ND. Firstly, we systematically investigated the GNN framework of ND, including graph construction, graph convolution, graph pooling, and graph prediction. Secondly, we investigated common NDs using the GNN diagnostic model in terms of data modality, number of subjects, and diagnostic accuracy. Thirdly, we discussed some research challenges and future research directions. The results of this review may be a valuable contribution to the ongoing intersection of artificial intelligence technology and brain imaging.

Blood-borne extracellular vesicles of bacteria and intestinal cells in patients with psychotic disorders

BACKGROUND

Human cells and bacteria secrete extracellular vesicles (EV) which play a role in intercellular communication. EV from the host intestinal epithelium are involved in the regulation of bacterial gene expression and growth. Bacterial EV (bactEV) produced in the intestine can pass to various tissues where they deliver biomolecules to many kinds of cells, including neurons. Emerging data indicate that gut microbiota is altered in patients with psychotic disorders. We hypothesized that the amount and content of blood-borne EV from intestinal cells and bactEV in psychotic patients would differ from healthy controls.

METHODS

We analyzed for human intestinal proteins by proteomics, for bactEV by metaproteomic analysis, and by measuring the level of lipopolysaccharide (LPS) in blood-borne EV from patients with psychotic disorders (n = 25), tested twice, in the acute phase of psychosis and after improvement, with age- and sex-matched healthy controls (n = 25).

RESULTS

Patients with psychotic disorders had lower LPS levels in their EV compared to healthy controls (p = .027). Metaproteome analyses confirmed LPS finding and identified Firmicutes and Bacteroidetes as dominating phyla. Total amounts of human intestine proteins in EV isolated from blood was lower in patients compared to controls (p = .02).

CONCLUSIONS

Our results suggest that bactEV and host intestinal EV are decreased in patients with psychosis and that this topic is worthy of further investigation given potential pathophysiological implications. Possible mechanisms involve dysregulation of the gut microbiota by host EV, altered translocation of bactEV to systemic circulation where bactEV can interact with both the brain and the immune system.

Bidirectional connectivity alterations in schizophrenia: a multivariate, machine-learning approach

Objective

It is well known that altered functional connectivity is a robust neuroimaging marker of schizophrenia. However, there is inconsistency in the direction of alterations, i.e., increased or decreased connectivity. In this study, we aimed to determine the direction of the connectivity alteration associated with schizophrenia using a multivariate, data-driven approach.

Methods

Resting-state functional magnetic resonance imaging data were acquired from 109 individuals with schizophrenia and 120 controls across two openly available datasets. A whole-brain resting-state functional connectivity (rsFC) matrix was computed for each individual. A modified connectome-based predictive model (CPM) with a support vector machine (SVM) was used to classify patients and controls. We conducted a series of multivariate classification analyses using three different feature sets, increased, decreased, and both increased and decreased rsFC.

Results

For both datasets, combining information from both increased and decreased rsFC substantially improved prediction accuracy (Dataset 1: accuracy = 70.2%, permutation p = 0.001; Dataset 2: accuracy = 64.4%, permutation p = 0.003). When tested across datasets, the prediction model using decreased rsFC performed best. The identified predictive features of decreased rsFC were distributed mostly in the motor network for both datasets.

Conclusion

These findings suggest that bidirectional alterations in rsFC are distributed in schizophrenia patients, with the pattern of decreased rsFC being more similar across different populations.

Automatic recognition of schizophrenia from brain-network features using graph convolutional neural network

Schizophrenia is a severe mental illness that imposes considerable economic burden on families and society. However, its clinical diagnosis primarily relies on scales and doctors' clinical experience and lacks an objective and accurate diagnostic approach. In recent years, graph convolutional neural networks (GCN) have been used to assist in psychiatric diagnosis owing to their ability to learn spatial-association information. Therefore, this study proposes a schizophrenia automatic recognition model based on graph convolutional neural network. Herein, the resting-state electroencephalography (EEG) data of 103 first-episode schizophrenia patients and 92 normal controls (NCs) were obtained. The automatic recognition model was trained with a nodal feature matrix that comprised the time and frequency-domain features of the EEG signals and local features of the brain network. The most significant regions that contributed to the model classification were identified, and the correlation between the node topological features of each significant region and clinical evaluation metrics was explored. Experiments were conducted to evaluate the performance of the model using 10-fold cross-validation. The best performance in the theta frequency band with a 6 s epoch length and phase-locked value. The recognition accuracy was 90.01%. The most significant region for identifying with first-episode schizophrenia patients and NCs was located in the parietal lobe. The results of this study verify the applicability of the proposed novel method for the identification and diagnosis of schizophrenia.

Dissecting Psychiatric Heterogeneity and Comorbidity with Core Region-Based Machine Learning

Machine learning approaches are increasingly being applied to neuroimaging data from patients with psychiatric disorders to extract brain-based features for diagnosis and prognosis. The goal of this review is to discuss recent practices for evaluating machine learning applications to obsessive-compulsive and related disorders and to advance a novel strategy of building machine learning models based on a set of core brain regions for better performance, interpretability, and generalizability. Specifically, we argue that a core set of co-altered brain regions (namely 'core regions') comprising areas central to the underlying psychopathology enables the efficient construction of a predictive model to identify distinct symptom dimensions/clusters in individual patients. Hypothesis-driven and data-driven approaches are further introduced showing how core regions are identified from the entire brain. We demonstrate a broadly applicable roadmap for leveraging this core set-based strategy to accelerate the pursuit of neuroimaging-based markers for diagnosis and prognosis in a variety of psychiatric disorders.

A meta-analysis and systematic review of single vs. multimodal neuroimaging techniques in the classification of psychosis

BACKGROUND

Psychotic disorders are characterized by structural and functional abnormalities in brain networks. Neuroimaging techniques map and characterize such abnormalities using unique features (e.g., structural integrity, coactivation). However, it is unclear if a specific method, or a combination of modalities, is particularly effective in identifying differences in brain networks of someone with a psychotic disorder.

METHODS

A systematic meta-analysis evaluated machine learning classification of schizophrenia spectrum disorders in comparison to healthy control participants using various neuroimaging modalities (i.e., T1-weighted imaging (T1), diffusion tensor imaging (DTI), resting state functional connectivity (rs-FC), or some combination (multimodal)). Criteria for manuscript inclusion included whole-brain analyses and cross-validation to provide a complete picture regarding the predictive ability of large-scale brain systems in psychosis. For this meta-analysis, we searched Ovid MEDLINE, PubMed, PsychInfo, Google Scholar, and Web of Science published between inception and March 13th 2023. Prediction results were averaged for studies using the same dataset, but parallel analyses were run that included studies with pooled sample across many datasets. We assessed bias through funnel plot asymmetry. A bivariate regression model determined whether differences in imaging modality, demographics, and preprocessing methods moderated classification. Separate models were run for studies with internal prediction (via cross-validation) and external prediction.

RESULTS

93 studies were identified for quantitative review (30 T1, 9 DTI, 40 rs-FC, and 14 multimodal). As a whole, all modalities reliably differentiated those with schizophrenia spectrum disorders from controls (OR = 2.64 (95%CI = 2.33 to 2.95)). However, classification was relatively similar across modalities: no differences were seen across modalities in the classification of independent internal data, and a small advantage was seen for rs-FC studies relative to T1 studies in classification in external datasets. We found large amounts of heterogeneity across results resulting in significant signs of bias in funnel plots and Egger's tests. Results remained similar, however, when studies were restricted to those with less heterogeneity, with continued small advantages for rs-FC relative to structural measures. Notably, in all cases, no significant differences were seen between multimodal and unimodal approaches, with rs-FC and unimodal studies reporting largely overlapping classification performance. Differences in demographics and analysis or denoising were not associated with changes in classification scores.

CONCLUSIONS

The results of this study suggest that neuroimaging approaches have promise in the classification of psychosis. Interestingly, at present most modalities perform similarly in the classification of psychosis, with slight advantages for rs-FC relative to structural modalities in some specific cases. Notably, results differed substantially across studies, with suggestions of biased effect sizes, particularly highlighting the need for more studies using external prediction and large sample sizes. Adopting more rigorous and systematized standards will add significant value toward understanding and treating this critical population.

Discriminative analysis of schizophrenia patients using graph convolutional networks: A combined multimodal MRI and connectomics analysis

Introduction

Recent studies in human brain connectomics with multimodal magnetic resonance imaging (MRI) data have widely reported abnormalities in brain structure, function and connectivity associated with schizophrenia (SZ). However, most previous discriminative studies of SZ patients were based on MRI features of brain regions, ignoring the complex relationships within brain networks.

Methods

We applied a graph convolutional network (GCN) to discriminating SZ patients using the features of brain region and connectivity derived from a combined multimodal MRI and connectomics analysis. Structural magnetic resonance imaging (sMRI) and resting-state functional magnetic resonance imaging (rs-fMRI) data were acquired from 140 SZ patients and 205 normal controls. Eighteen types of brain graphs were constructed for each subject using 3 types of node features, 3 types of edge features, and 2 brain atlases. We investigated the performance of 18 brain graphs and used the TopK pooling layers to highlight salient brain regions (nodes in the graph).

Results

The GCN model, which used functional connectivity as edge features and multimodal features (sMRI + fMRI) of brain regions as node features, obtained the highest average accuracy of 95.8%, and outperformed other existing classification studies in SZ patients. In the explainability analysis, we reported that the top 10 salient brain regions, predominantly distributed in the prefrontal and occipital cortices, were mainly involved in the systems of emotion and visual processing.

Discussion

Our findings demonstrated that GCN with a combined multimodal MRI and connectomics analysis can effectively improve the classification of SZ at an individual level, indicating a promising direction for the diagnosis of SZ patients. The code is available at https://github.com/CXY-scut/GCN-SZ.git.

Acquired Pedophilia: international Delphi-method-based consensus guidelines

Idiopathic and acquired pedophilia are two different disorders with two different etiologies. However, the differential diagnosis is still very difficult, as the behavioral indicators used to discriminate the two forms of pedophilia are underexplored, and clinicians are still devoid of clear guidelines describing the clinical and neuroscientific investigations suggested to help them with this difficult task. Furthermore, the consequences of misdiagnosis are not known, and a consensus regarding the legal consequences for the two kinds of offenders is still lacking. The present study used the Delphi method to reach a global consensus on the following six topics: behavioral indicators/red flags helpful for differential diagnosis; neurological conditions potentially leading to acquired pedophilia; neuroscientific investigations important for a correct understanding of the case; consequences of misdiagnosis; legal consequences; and issues and future perspectives. An international and multidisciplinary board of scientists and clinicians took part in the consensus statements as Delphi members. The Delphi panel comprised 52 raters with interdisciplinary competencies, including neurologists, psychiatrists, neuropsychologists, forensic psychologists, expert in ethics, etc. The final recommendations consisted of 63 statements covering the six different topics. The current study is the first expert consensus on a delicate topic such as pedophilia. Important exploitable consensual recommendations that can ultimately be of immediate use by clinicians to help with differential diagnosis and plan and guide therapeutic interventions are described, as well as future perspectives for researchers.

The inferior frontal gyrus and familial risk for bipolar disorder

Bipolar disorder (BD) is a familial disorder with high heritability. Genetic factors have been linked to the pathogenesis of BD. Relatives of probands with BD who are at familial risk can exhibit brain abnormalities prior to illness onset. Given its involvement in prefrontal cognitive control and in frontolimbic circuitry that regulates emotional reactivity, the inferior frontal gyrus (IFG) has been a focus of research in studies of BD-related pathology and BD-risk mechanism. In this review, we discuss multimodal neuroimaging findings of the IFG based on studies comparing at-risk relatives and low-risk controls. Review of these studies in at-risk cases suggests the presence of both risk and resilience markers related to the IFG. At-risk individuals exhibited larger gray matter volume and increased functional activities in IFG compared with low-risk controls, which might result from an adaptive brain compensation to support emotion regulation as an aspect of psychological resilience. Functional connectivity between IFG and downstream limbic or striatal areas was typically decreased in at-risk individuals relative to controls, which could contribute to risk-related problems of cognitive and emotional control. Large-scale and longitudinal investigations on at-risk individuals will further elucidate the role of IFG and other brain regions in relation to familial risk for BD, and together guide identification of at-risk individuals for primary prevention.

Abnormal regional homogeneity in right caudate as a potential neuroimaging biomarker for mild cognitive impairment: A resting-state fMRI study and support vector machine analysis

Objective

Mild cognitive impairment (MCI) is a heterogeneous syndrome characterized by cognitive impairment on neurocognitive tests but accompanied by relatively intact daily activities. Due to high variation and no objective methods for diagnosing and treating MCI, guidance on neuroimaging is needed. The study has explored the neuroimaging biomarkers using the support vector machine (SVM) method to predict MCI.

Methods

In total, 53 patients with MCI and 68 healthy controls were involved in scanning resting-state functional magnetic resonance imaging (rs-fMRI). Neurocognitive testing and Structured Clinical Interview, such as Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) test, Activity of Daily Living (ADL) Scale, Hachinski Ischemic Score (HIS), Clinical Dementia Rating (CDR), Montreal Cognitive Assessment (MoCA), and Hamilton Rating Scale for Depression (HRSD), were utilized to assess participants' cognitive state. Neuroimaging data were analyzed with the regional homogeneity (ReHo) and SVM methods.

Results

Compared with healthy comparisons (HCs), ReHo of patients with MCI was decreased in the right caudate. In addition, the SVM classification achieved an overall accuracy of 68.6%, sensitivity of 62.26%, and specificity of 58.82%.

Conclusion

The results suggest that abnormal neural activity in the right cerebrum may play a vital role in the pathophysiological process of MCI. Moreover, the ReHo in the right caudate may serve as a neuroimaging biomarker for MCI, which can provide objective guidance on diagnosing and managing MCI in the future.