Optimizing graph neural network architectures for schizophrenia spectrum disorder prediction using evolutionary algorithms

Thalamocortical dysrhythmia (TCD) is a model characterized by abnormal resting-state thalamic oscillatory patterns where the alpha rhythm is replaced by cross-frequency coupling of low- and high-frequency rhythms. Although disrupted thalamic function is a suggested important pathophysiological mechanism underlying schizophrenia, knowledge regarding the TCD model in schizophrenia spectrum disorder (SSD) patients and individuals at clinical high risk (CHR) for psychosis is limited. A total of 169 SSD patients, 106 individuals at CHR for psychosis, and 105 healthy controls (HCs) underwent resting-state electroencephalography recordings. We performed mean global field power (MGFP) spectral analysis between 1 and 49 Hz as well as source-level theta phase-gamma amplitude coupling (TGC) analysis and compared resting-state oscillatory patterns across groups. Correlations between altered TGC values and psychotic symptom severity in the patient group were investigated. Spectral MGFP of low- and high-frequencies was larger in the SSD and CHR groups than in the HC group. The TGC of SSD patients was greater than that of HCs in the right frontal, right parietal, and left and right limbic lobes. Greater TGC in the right frontal and limbic lobes was associated with positive symptom severity in SSD patients. However, TGC in the CHR group was comparable to that in the HCs and was smaller than that in the SSD group in widespread cortical regions. The TCD pattern may be apparent after frank psychotic disorder onset in tandem with overt positive symptoms. A psychosis-risk state without overt psychotic symptoms could be characterized by abnormally increased low- and high-frequency activities with relatively preserved TGC.

Sex classification is a major benchmark of previous work in learning on the structural connectome, a naturally occurring brain graph that has proven useful for studying cognitive function and impairment. While graph neural networks (GNNs), specifically graph convolutional networks (GCNs), have gained popularity lately for their effectiveness in learning on graph data, achieving strong performance in adult sex classification tasks, their application to pediatric populations remains unexplored. We seek to characterize the capacity for GNN models to learn connectomic patterns on pediatric data through an exploration of training techniques and architectural design choices. Two datasets comprising an adult BRIGHT dataset (N = 147 Hodgkin's lymphoma survivors and N = 162 age similar controls) and a pediatric Human Connectome Project in Development (HCP-D) dataset (N = 135 healthy subjects) were utilized. Two GNN models (GCN simple and GCN residual), a deep neural network (multi-layer perceptron), and two standard machine learning models (random forest and support vector machine) were trained. Architecture exploration experiments were conducted to evaluate the impact of network depth, pooling techniques, and skip connections on the ability of GNN models to capture connectomic patterns. Models were assessed across a range of metrics including accuracy, AUC score, and adversarial robustness. GNNs outperformed other models across both populations. Notably, adult GNN models achieved 85.1% accuracy in sex classification on unseen adult participants, consistent with prior studies. The extension of the adult models to the pediatric dataset and training on the smaller pediatric dataset were sub-optimal in their performance. Using adult data to augment pediatric models, the best GNN achieved comparable accuracy across unseen pediatric (83.0%) and adult (81.3%) participants. Adversarial sensitivity experiments showed that the simple GCN remained the most robust to perturbations, followed by the multi-layer perceptron and the residual GCN. These findings underscore the potential of GNNs in advancing our understanding of sex-specific neurological development and disorders and highlight the importance of data augmentation in overcoming challenges associated with small pediatric datasets. Further, they highlight relevant tradeoffs in the design landscape of connectomic GNNs. For example, while the simpler GNN model tested exhibits marginally worse accuracy and AUC scores in comparison to the more complex residual GNN, it demonstrates a higher degree of adversarial robustness.

Developing fast and accurate computational models to simulate intricate physical phenomena has been a persistent research challenge. Recent studies have demonstrated remarkable capabilities in predicting various physical outcomes through machine learning-assisted approaches. However, it remains challenging to generalize current methods, usually crafted for a specific problem, to other more complex or broader scenarios. To address this challenge, we developed graph neural network (GNN) models with enhanced generalizability derived from the distinct GNN architecture and neural operator techniques. As a proof of concept, we employ our GNN models to predict finite element (FE) simulation results for three-dimensional solid mechanics problems with varying boundary conditions. Results show that our GNN model achieves accurate and robust performance in predicting the stress and deformation profiles of structures compared with FE simulations. Furthermore, the neural operator embedded GNN approach enables learning and predicting various solid mechanics problems in a generalizable fashion, making it a promising approach for surrogate modeling.

Research has demonstrated that Black patients are more likely to be diagnosed with Schizophrenia Spectrum Disorder (SSD) compared to White patients. This integrative review focused on factors associated with symptoms, symptom expression, and the diagnosis of Black patients with SSD. Four data bases were searched from 2013 to 2023 and included studies written in English, studied patients with symptoms or diagnoses of SSDs; reported the symptoms or diagnoses of SSD in Black patients; examined any factors related to symptoms or diagnosis of SSD; and were published in peer-reviewed journals. Studies reported that differences in diagnosis may be due to how Black patients' experience symptoms of SSD, socio-geographic factors such as geography and migration, and provider demographic factors such as age, gender, and race. Results indicate the need for further discussion and awareness of how individual, socio-geographic, and provider factors may be associated with racial differences in symptoms, symptom expression, and the diagnosis of Back patients with SSD.

Patient Activation Measure-13 (PAM-13) is a valid and widely used questionnaire that assess an individual’s knowledge, confidence, and skills for self-management of their chronic illness. Although there is some evidence regarding its reliability, the test–retest reliability has not been investigated among patients with substance use disorders (SUDs) or schizophrenia spectrum disorders. We investigated the internal consistency and test–retest reliability of PAM-13 in these populations. Test–retest reliability was analysed using data from 29 patients with SUDs and 28 with schizophrenia spectrum disorders. Cronbach’s α and Intraclass Correlation Coefficient (ICC) scores were used to examine internal consistency and test–retest reliability, respectively. Of the 60 collected test–retest questionnaires, 57 were included in the analyses. No mean differences between time one (T1) and time two (T2) were observed in either patient group, except for item 12 in schizophrenia spectrum disorders patients (p < 0.05). Internal consistency for T1 and T2 was 0.75 and 0.84 in SUDs patients and 0.87 and 0.81 in schizophrenia spectrum disorders patients, respectively. The ICC was r = 0.86 in patients with SUDs and r = 0.93 in patients with schizophrenia spectrum disorders. To conclude, PAM-13 showed good internal consistency and test–retest reliability in SUDs and schizophrenia spectrum disorders patients.

Heterogeneity in Schizophrenia Spectrum Disorders: Studies on Functional, Affective, and Metabolic Outcomes This dissertation aimed to explore the heterogeneity in schizophrenia spectrum disorders (SSD) by examining long-term clinical and subjective outcomes in older adults and by investigating the role of depressive symptomatology in relation to metabolic risk. Part 1: Long-term Course of Remission and Quality of Life in Older Adults with SSD With an aging population, the number of older individuals with SSD is increasing. Their care needs often differ from those of younger patients. A better understanding of the long-term course of SSD in later life is therefore essential. Contrary to earlier beliefs that schizophrenia inevitably leads to a dementia-like state, more recent evidence suggests a more diverse and sometimes favorable course in older patients. Symptomatic remission, defined as a state of low symptom severity, has become a useful clinical concept. However, recovery must also include the patient’s perspective, especially in terms of subjective quality of life (SQoL). Using data from the SOUL study, we analyzed 77 patients aged 60+ over a 5-year period. We found that 26% of patients improved in terms of symptomatic remission, while 23% remained stable in remission. Key predictors of a favorable course were lower baseline positive symptoms and having a partner. In a separate analysis of 75 patients from the same cohort, we examined changes in SQoL using the Manchester Short Assessment of Quality of Life (MANSA). About 36% showed clinically relevant improvement, while 20% experienced a decline. Factor analysis revealed three SQoL domains: ‘personal circumstances,’ ‘daily life,’ and ‘social relations,’ with the first two showing notable improvement. Older age predicted more change, whereas a higher baseline SQoL score predicted stability. These findings suggest that improvement in both clinical and subjective outcomes is achievable even at later ages. Part 2: Depressive Symptom Profiles and Metabolic Syndrome in SSD People with SSD have higher mortality rates and shorter life expectancy, partly due to somatic conditions like cardiovascular disease, obesity, and diabetes. Recent studies suggest a link between depressive symptoms and metabolic syndrome (MetS), especially symptoms related to energy dysregulation. We compared depressive symptom profiles in 449 individuals with SSD, 816 with depressive disorders (DD), and 417 healthy controls (aged 18–50). Somatic/vegetative symptoms were more similar between SSD and DD groups than cognitive/mood symptoms. Among SSD patients, depressive symptom severity was associated with male gender, lower social functioning, use of benzodiazepines, and higher positive symptom burden. In a second study of 513 individuals with SSD, we examined the relationship between energy-related depressive symptoms (e.g., hypersomnia, hyperphagia, fatigue) and MetS. A third of the sample met criteria for MetS. Energy-related symptoms—but not melancholic symptoms—were significantly associated with both the presence and severity of MetS. Notably, hypersomnia showed the strongest link. Conclusion This dissertation underscores the heterogeneity in SSD, both in long-term outcomes and comorbid conditions. Older individuals with SSD can experience improvements in remission and quality of life, especially with social support. Additionally, specific depressive symptom patterns—particularly energy-related symptoms—may help identify patients at greater metabolic risk, supporting the need for integrated care approaches that address both mental and physical health.

Deep learning has led to major advances in fields like natural language processing, computer vision, and other Euclidean data domains. Yet, many important fields have data defined on irregular domains, requiring graphs to be explicitly modeled. One such application is drug discovery. Recently, research has found that using graph neural network (GNN) models, given enough data, can perform better than using human-engineered fingerprints or descriptors in predicting molecular properties of potential antibiotics.We explore these state-of-the-art AI models on predicting desirable molecular properties for drugs that can inhibit SARS-CoV-2. We build upon the GNN models with ideas from recent breakthroughs in geometric deep learning, inspired by the topologies of the molecules. In this poster paper, we present an overview of the drug discovery framework, drug-target interaction framework, and GNNs. Preliminary results on two COVID-19 related datasets are encouraging, achieving a ROC-AUC of 0.72 for FDA-approved chemical library screened against SARS-CoV-2 in vitro.

Past events have revealed that widespread blackouts are mostly a result of cascading failures in the power grid. Understanding the underlining mechanisms of cascading failures can help in developing strategies to minimize the risk of such events. Moreover, a real-time detection of precursors to cascading failures will help operators take measures to prevent their propagation. Currently, the well-established probabilistic and physics-based models of cascading failures offer low computational efficiency, hindering them to be used only as offline tools. In this work, we develop a data-driven methodology for online estimation of the risk of cascading failures. We utilize a physics-based cascading failure model to generate a cascading failure dataset considering different operating conditions and failure scenarios, thus obtaining a sample space covering a large set of power grid states that are labeled as safe or unsafe. We use the synthetic data to train deep learning architectures, namely Feed-forward Neural Networks (FNN) and Graph Neural Networks (GNN). With the development of GNNs, improved performance is achieved with graph-structured data, and GNNs can generalize to graphs of diverse sizes. A comparison between FNN and GNN is made and the GNNs inductive capability is demonstrated via test grids. Furthermore, we apply transfer learning to improve the performance of a pre-trained GNN model on power grids not seen in the training process. The GNN model shows accuracy and balanced accuracy above 96% on selected test datasets not used in the training. Conversely, the FNN shows accuracy above 85% and balanced accuracy above 81% on test datasets unseen during training. Overall, the GNN model is successful in determining, if one or several simultaneous outages result in a critical grid state, under specific grid operating conditions.

Graph Neural Networks (GNNs) have become indispensable tools for learning from graph-structured data, catering to various applications such as social network analysis and fraud detection for financial services. At the heart of these networks are the edges, which are crucial in guiding GNN models' predictions. In many scenarios, these edges represent sensitive information, such as personal associations or financial dealings, which require privacy assurance. However, their contributions to GNN model predictions may, in turn, be exploited by the adversary to compromise their privacy. Motivated by these conflicting requirements, this paper investigates edge privacy in contexts where adversaries possess only black-box access to the target GNN model, restricted further by access controls, preventing direct insights into arbitrary node outputs. Moreover, we are the first to extensively examine situations where the target graph continuously evolves—a common trait of many real-world graphs. In this setting, we present a range of attacks that leverage the message-passing mechanism of GNNs. We evaluated the effectiveness of our attacks using nine real-world datasets, encompassing both static and dynamic graphs, across four different GNN architectures. The results demonstrate that our attack outperforms existing methods across various GNN architectures, consistently achieving an F1 score of at least 0.8 in static scenarios. Furthermore, our attack retains robustness in dynamic graph scenarios, maintaining F1 scores up to 0.8, unlike previous methods that only achieve F1 scores around 0.2.

Graph neural networks (GNNs) have made rapid developments in the recent years. Due to their great ability in modeling graph-structured data, GNNs are vastly used in various applications, including high-stakes scenarios such as financial analysis, traffic predictions, and drug discovery. Despite their great potential in benefiting humans in the real world, recent study shows that GNNs can leak private information, are vulnerable to adversarial attacks, can inherit and magnify societal bias from training data and lack interpretability, which have risk of causing unintentional harm to the users and society. For example, existing works demonstrate that attackers can fool the GNNs to give the outcome they desire with unnoticeable perturbation on training graph. GNNs trained on social networks may embed the discrimination in their decision process, strengthening the undesirable societal bias. Consequently, trust-worthy GNNs in various aspects are emerging to prevent the harm from GNN models and increase the users’ trust in GNNs. In this paper, we give a comprehensive survey of GNNs in the computational aspects of privacy, robustness, fairness, and explainability. For each aspect, we give the taxonomy of the related methods and formulate the general frameworks for the multiple categories of trustworthy GNNs. We also discuss the future research directions of each aspect and connections between these aspects to help achieve trustworthiness.

Most distribution system operators may not accurately record or completely maintain the phase connections for the numerous LV customers. Different consumer phase identification (CPI) approaches based on voltages, powers or other measurements are proposed in the literature. Due to the technical challenges in collecting voltage measurements, power measurement based approaches are preferable. Hence, this paper proposes a novel power based CPI methodology applying Graph Neural Networks (GNNs). The CPI methodology generates synthetic transformer power profiles assuming random combinations of phases for the measured load profiles, which are used altogether to train the GNN model. The GNN model is then tested using measured transformer and load power profiles. The performance of the methodology is evaluated in a test low voltage grid of 55 loads under various conditions of Photovoltaic penetration, Photovoltaics under maintenance, measurement errors, unmetered consumption, uncertain grid asset parameters and inaccurate phase connections. Further tests on a real low voltage grid with 111 loads prove the scalability of the methodology. The attained results show that the GNN model can achieve accuracy above 90% in most cases, outperforming various state-of-the-art methods. • The Graph Neural Network identifies the phase labels with accuracy higher than 90%. • Under noisy or unmetered demand, the Graph Neural Network achieves accuracy over 90%. • Under various Photovoltaic cases, the Graph Neural Network yields accuracy over 90%. • The Graph Neural Network yields accuracy of 92% in case of multi-phase connections. • The grid losses and uncertain assets’ parameters have minor impact on model accuracy.

Graph Neural Networks (GNNs) are the current state-of-the-art models in learning node representations for many predictive tasks on graphs. Typically, GNNs reuses the same set of model parameters across all nodes in the graph to improve the training efficiency and exploit the translationally-invariant properties in many datasets. However, the parameter sharing scheme prevents GNNs from distinguishing two nodes that are isomorphic and that the translation invariance property may not exhibit in real-world graphs. In this paper, we present Graph Representation Learning with Adaptive Mixtures (GRAM), a novel approach for learning node representations in a graph by introducing multiple independent GNN models and a trainable mixture distribution for each node. GRAM is a general framework that can be readily embedded into existing models for enhancing representation learning. It allows for adaptive node clustering with mixtures and joint optimization of all GNN models simultaneously. To achieve this, we develop an Expectation-Maximization algorithm with stochastic gradient descent in a scalable manner. Specifically, in the E-Step, we approximate the posterior probability distribution of the latent cluster membership based on the k-th independent GNN models in the last iteration. In the M-step, we update all GNN models and the mixture parameters based on the refined latent cluster membership for downstream tasks, e.g., node classification. To fully exploit the graph structural information, we further implement a regularization term on unlabeled nodes, which encourages adjacent nodes to have the same mixture distribution. By jointly optimizing the model in an end-to-end manner, GRAM learns to capture different abstract patterns on graphs for different node clusters, which yields discriminative representation for each node. We evaluate GRAM on five benchmark datasets with extensive experiments. GRAM is demonstrated to consistently boost state-of-the-art GNN variants in node classification tasks.

Traditional Graph Neural Network (GNN) learning patterns can only achieve optimal performance under the assumption of independent and identically distributed data. However, in real-world scenarios, numerous unpredictable factors give rise to the Out-of-Distribution (OOD) problem. Moreover, this distribution discrepancy may lead to unreliable predictions by GNNs models within unknown OOD domains. Therefore, this paper proposes a machine learning method to enhance the stability of GNN cross-site classification from the perspective of stable learning. The aim is to extract genuine causal features while eliminating spurious causal features. By introducing a feature sample weighting decorrelation technique in the random Fourier transform space and combining it with a baseline GNN model, a Stable-GNN model and a constrained sampling weight gradient update algorithm are designed. Its theoretical proof indicates that this algorithm can ensure the decrease in the loss, thus remedying the shortcomings of the original algorithm in weight updating. While guaranteeing predictive performance on the training distribution data, it also reduces prediction bias on data from unseen test distributions. Compared with the GNN model and the Stable-GNN model (S-GNN), the experimental results of the S-GNN model demonstrate that it not only surpasses the current state-of-the-art GNN models, but also offers a flexible framework for strengthening existing GNNs.

Graph representational learning can detect topological patterns by leveraging both the network structure as well as nodal features. The basis of our exploration involves the application of graph neural network architectures and machine learning to resting-state functional Magnetic Resonance Imaging (rs-fMRI) data for the purpose of detecting schizophrenia. Our study uses single-site data to avoid the shortcomings in generalizability of neuroimaging data obtained from multiple sites. The performance of our graph neural network models is on par with that of our machine learning models, each of which is trained using 69 graph-theoretical measures computed from functional correlations between various regions of interest (ROI) in a brain graph. Our deep graph convolutional neural network (DGCNN) demonstrates a promising average accuracy score of 0.82 and a sensitivity score of 0.84. This study provides insights into the role of advanced graph theoretical methods and machine learning on fMRI data to detect schizophrenia by harnessing changes in brain functional connectivity. The results of this study demonstrate the capabilities of using both traditional ML techniques as well as graph neural network-based methods to detect schizophrenia using features extracted from fMRI data. The study also proposes two methods to obtain potential biomarkers for the disease, many of which are corroborated by research in this area and can further help in the understanding of schizophrenia as a mental disorder.

The high energy demand of the human brain obligates robust mitochondrial energy metabolism, while mitochondrial dysfunctions have been linked to neuropsychiatric disorders, including schizophrenia spectrum disorders (SSD). However, in vivo assessments that can directly inform brain mitochondrial functioning and its etiopathophysiological path to SSD remain difficult to obtain. We hypothesized that system and brain mitochondrial dysfunctions in SSD may be indexed by elevated cell-free mitochondrial DNA (cf-mtDNA) levels in the blood and in neuronal extracellular vesicles (nEVs). We also explored if these mtDNA marker elevations were associated with brain metabolites as measured by magnetic resonance spectroscopy (MRS). We examined blood cf-mtDNA in 58 SSD patients and 33 healthy controls, followed by assessing nEV mtDNA and metabolite levels using MRS in a subgroup of patients and controls. We found that people with SSD had significantly elevated cf-mtDNA levels in both the blood (p = 0.0002) and neuronal EVs (p = 0.003) compared to controls. These mtDNA abnormalities can be linked back to brain lactate+ levels such that higher blood and nEV mtDNA levels were significantly associated with higher lactate+ levels measured at the anterior cingulate cortex (r = 0.53, 0.53; p = 0.008, 0.03, respectively) in SSD patients. Furthermore, higher developmental stress and trauma were significantly associated with higher cf-mtDNA levels in both the blood and neuronal EVs in SSD patients (r = 0.29, 0.49; p = 0.01, 0.03, respectively). In conclusion, if replicated and fully developed, blood and neuronal EV-based cell-free mtDNA may provide a clinically accessible biomarker to more directly evaluate the mitochondrial hypothesis and the abnormal bioenergetics pathways in schizophrenia.