Multiple clustering for identifying subject clusters and brain sub-networks using functional connectivity matrices without vectorization

Exploring neural heterogeneity in inattention and hyperactivity

Inattention and hyperactivity are cardinal symptoms of Attention Deficit Hyperactivity Disorder (ADHD). These characteristics have also been observed across a range of other neurodevelopmental conditions, such as autism and dyspraxia, suggesting that they might best be studied across diagnostic categories. Here, we evaluated the associations between inattention and hyperactivity behaviours and features of the structural brain network (connectome) in a large transdiagnostic sample of children (Centre for Attention, Learning, and Memory; n = 383). In our sample, we found that a single latent factor explains 77.6% of variance in scores across multiple questionnaires measuring inattention and hyperactivity. Partial Least-Squares (PLS) regression revealed that variability in this latent factor could not be explained by a linear component representing nodewise properties of connectomes. We then investigated the type and extent of neural heterogeneity in a subset of our sample with clinically-elevated levels of inattention and hyperactivity. Multidimensional scaling combined with k-means clustering revealed two neural subtypes in children with elevated levels of inattention and hyperactivity (n = 232), differentiated primarily by nodal communicability-a measure which demarcates the extent to which neural signals propagate through specific brain regions. These different clusters had similar behavioural profiles, which included high levels of inattention and hyperactivity. However, one of the clusters scored higher on multiple cognitive assessment measures of executive function. We conclude that inattention and hyperactivity are so common in children with neurodevelopmental difficulties because they emerge through multiple different trajectories of brain development. In our own data, we can identify two of these possible trajectories, which are reflected by measures of structural brain network topology and cognition.

An explainable autoencoder with multi-paradigm fMRI fusion for identifying differences in dynamic functional connectivity during brain development

Multi-paradigm deep learning models show great potential for dynamic functional connectivity (dFC) analysis by integrating complementary information. However, many of them cannot use information from different paradigms effectively and have poor explainability, that is, the ability to identify significant features that contribute to decision making. In this paper, we propose a multi-paradigm fusion-based explainable deep sparse autoencoder (MF-EDSAE) to address these issues. Considering explainability, the MF-EDSAE is constructed based on a deep sparse autoencoder (DSAE). For integrating information effectively, the MF-EDASE contains the nonlinear fusion layer and multi-paradigm hypergraph regularization. We apply the model to the Philadelphia Neurodevelopmental Cohort and demonstrate it achieves better performance in detecting dynamic FC (dFC) that differ significantly during brain development than the single-paradigm DSAE. The experimental results show that children have more dispersive dFC patterns than adults. The function of the brain transits from undifferentiated systems to specialized networks during brain development. Meanwhile, adults have stronger connectivities between task-related functional networks for a given task than children. As the brain develops, the patterns of the global dFC change more quickly when stimulated by a task.

Dopamine-Mediated Major Depressive Disorder in the Neural Circuit of Ventral Tegmental Area-Nucleus Accumbens-Medial Prefrontal Cortex: From Biological Evidence to Computational Models

Major depressive disorder (MDD) is a serious psychiatric disorder, with an increasing incidence in recent years. The abnormal dopaminergic pathways of the midbrain cortical and limbic system are the key pathological regions of MDD, particularly the ventral tegmental area- nucleus accumbens- medial prefrontal cortex (VTA-NAc-mPFC) neural circuit. MDD usually occurs with the dysfunction of dopaminergic neurons in VTA, which decreases the dopamine concentration and metabolic rate in NAc/mPFC brain regions. However, it has not been fully explained how abnormal dopamine concentration levels affect this neural circuit dynamically through the modulations of ion channels and synaptic activities. We used Hodgkin-Huxley and dynamical receptor binding model to establish this network, which can quantitatively explain neural activity patterns observed in MDD with different dopamine concentrations by changing the kinetics of some ion channels. The simulation replicated some important pathological patterns of MDD at the level of neurons and circuits with low dopamine concentration, such as the decreased action potential frequency in pyramidal neurons of mPFC with significantly reduced burst firing frequency. The calculation results also revealed that NaP and KS channels of mPFC pyramidal neurons played key roles in the functional regulation of this neural circuit. In addition, we analyzed the synaptic currents and local field potentials to explain the mechanism of MDD from the perspective of dysfunction of excitation-inhibition balance, especially the disinhibition effect in the network. The significance of this article is that we built the first computational model to illuminate the effect of dopamine concentrations for the NAc-mPFC-VTA circuit between MDD and normal groups, which can be used to quantitatively explain the results of existing physiological experiments, predict the results for unperformed experiments and screen possible drug targets.

Clustering of Multiple Psychiatric Disorders Using Functional Connectivity in the Data-Driven Brain Subnetwork

Recently, the dimensional approach has attracted much attention, bringing a paradigm shift to a continuum of understanding of different psychiatric disorders. In line with this new paradigm, we examined whether there was common functional connectivity related to various psychiatric disorders in an unsupervised manner without explicitly using diagnostic label information. To this end, we uniquely applied a newly developed network-based multiple clustering method to resting-state functional connectivity data, which allowed us to identify pairs of relevant brain subnetworks and subject cluster solutions accordingly. Thus, we identified four subject clusters, which were characterized as major depressive disorder (MDD), young healthy control (young HC), schizophrenia (SCZ)/bipolar disorder (BD), and autism spectrum disorder (ASD), respectively, with the relevant brain subnetwork represented by the cerebellum-thalamus-pallidum-temporal circuit. The clustering results were validated using independent datasets. This study is the first cross-disorder analysis in the framework of unsupervised learning of functional connectivity based on a data-driven brain subnetwork.