clusterBMA: Bayesian model averaging for clustering.

Various methods have been developed to combine inference across multiple sets of results for unsupervised clustering, within the ensemble clustering literature. The approach of reporting results from one 'best' model out of several candidate clustering models generally ignores the uncertainty that arises from model selection, and results in inferences that are sensitive to the particular model and parameters chosen. Bayesian model averaging (BMA) is a popular approach for combining results across multiple models that offers some attractive benefits in this setting, including probabilistic interpretation of the combined cluster structure and quantification of model-based uncertainty. In this work we introduce clusterBMA, a method that enables weighted model averaging across results from multiple unsupervised clustering algorithms. We use clustering internal validation criteria to develop an approximation of the posterior model probability, used for weighting the results from each model. From a combined posterior similarity matrix representing a weighted average of the clustering solutions across models, we apply symmetric simplex matrix factorisation to calculate final probabilistic cluster allocations. In addition to outperforming other ensemble clustering methods on simulated data, clusterBMA offers unique features including probabilistic allocation to averaged clusters, combining allocation probabilities from 'hard' and 'soft' clustering algorithms, and measuring model-based uncertainty in averaged cluster allocation. This method is implemented in an accompanying R package of the same name. We use simulated datasets to explore the ability of the proposed technique to identify robust integrated clusters with varying levels of separation between subgroups, and with varying numbers of clusters between models. Benchmarking accuracy against four other ensemble methods previously demonstrated to be highly effective in the literature, clusterBMA matches or exceeds the performance of competing approaches under various conditions of dimensionality and cluster separation. clusterBMA substantially outperformed other ensemble methods for high dimensional simulated data with low cluster separation, with 1.16 to 7.12 times better performance as measured by the Adjusted Rand Index. We also explore the performance of this approach through a case study that aims to identify probabilistic clusters of individuals based on electroencephalography (EEG) data. In applied settings for clustering individuals based on health data, the features of probabilistic allocation and measurement of model-based uncertainty in averaged clusters are useful for clinical relevance and statistical communication.

Longitudinal associations between resting-state, interregional theta-beta phase-amplitude coupling, psychological distress, and wellbeing in 12-15-year-old adolescents.

Cross-frequency coupling between the phase of slower oscillatory activity and the amplitude of faster oscillatory activity in the brain (phase-amplitude coupling; PAC), is a promising new biological marker for mental health. Prior research has demonstrated that PAC is associated with mental health. However, most research has focused on within-region theta-gamma PAC in adults. Our recent preliminary study found increased theta-beta PAC was associated with increased psychological distress in 12 year olds. It is important to investigate how PAC biomarkers relate to mental health and wellbeing in youth. Thus, in this study, we investigated longitudinal associations between interregional (posterior-anterior cortex) resting-state theta-beta PAC (Modulation Index [MI]), psychological distress and wellbeing in N = 99 adolescents (aged 12-15 years). In the right hemisphere, there was a significant relationship, whereby increased psychological distress was associated with decreased theta-beta PAC and psychological distress increased with increased age. In the left hemisphere, there was a significant relationship, whereby decreased wellbeing was associated with decreased theta-beta PAC and wellbeing scores decreased with increased age. This study presents novel findings demonstrating longitudinal relationships between interregional, resting-state theta-beta PAC and mental health and wellbeing in early adolescents. This EEG marker may facilitate improved early identification of emerging psychopathology.

Structural connectivity and its association with social connectedness in early adolescence.

Adolescence is a critical period of social and neural development. Brain regions which process social information develop throughout adolescence as young people learn to navigate social environments. Studies investigating brain structural connectivity (indexed by white matter (WM) integrity), and social connectedness in adolescents have been limited until recently, with literature stemming mostly from adult samples, broad age ranges within adolescence or based on social network characteristics as opposed to social connectedness. This cross-sectional study of 12-year-olds (N = 73) explored the relationship between social connectedness (SCS) and structural connectivity in early adolescence, to gauge how this snapshot of WM development is associated with social behaviour. Whole brain voxel-wise diffusion tensor imaging was undertaken to determine correlations between SCS and fractional anisotropy (FA), radial (RD) and axial (AD) diffusivity of clusters within WM tracts. Significant negative relationships between FA and SCS scores were found in clusters within 11 WM tracts, with significant positive correlations between SCS and both RD and AD across clusters within 13 and 8 clusters, respectively. Clusters within the genu of the corpus callosum (CCgn) showed strong correlations for all three metrics, and regression models that included gender, age, and psychological distress, revealed SCS to be the only significant predictor of CCgn FA, RD and AD values. Overall, these findings suggest that those with lower social connectedness had a WM profile suggestive of reduced axonal density and/or coherence. Longitudinal research is needed to track such WM profiles during adolescent development and determine the associations with mental health and well-being outcomes.

Dataset of brain functional connectome and its maturation in adolescents.

We provided the dataset of brain connectome matrices, their similarities measures to self and others longitudinally, and Kessler's psychological distress scales (K10) including the response to each question. The dataset can be used to replicate the results of the manuscript titled "A longitudinal study of functional connectome uniqueness and its association with psychological distress in adolescence". The functional connectome (whole-brain and 13 networks) matrices were calculated from the resting-state functional MRIs (rs-fMRIs). We collected rs-fMRI and Kessler's psychological distress scale (K10) in 77 adolescents longitudinally up to 9 times from 12 years of age every four months. After removal of data with excessive motion, 262 functional connectome matrices were provided with this paper. The 300 regions of interest (ROIs) were defined using the Greene lab brain atlas. The functional connectome matrices were calculated as correlations between time series from any pair of ROIs extracted from pre-processed fMRIs. This dataset could be potentially used to1.Understand developmental changes in the functional brain connectivity,2.As a normal control database of functional connectome matrices,3.Develop and validate connectome and network-related analysing methods.

EEG-based clusters differentiate psychological distress, sleep quality and cognitive function in adolescents.

INTRODUCTION: To better understand the relationships between neurophysiology, cognitive function and psychopathology risk in adolescence there is value in identifying data-driven subgroups based on measurements of brain activity and function, and then comparing cognition and mental health between such subgroups. METHODS: We developed a flexible and scaleable multi-stage analysis pipeline to identify data-driven clusters of 12-year-olds (M = 12.64, SD = 0.32) based on frequency characteristics calculated from resting state, eyes-closed electroencephalography (EEG) recordings. For this preliminary cross-sectional study, EEG data was collected from 59 individuals in the Longitudinal Adolescent Brain Study (LABS) being undertaken in Queensland, Australia. Applying multiple unsupervised clustering algorithms to these EEG features, we identified well-separated subgroups of individuals. To study patterns of difference in cognitive function and mental health symptoms between clusters, we applied Bayesian regression models to probabilistically identify differences in these measures between clusters. RESULTS: We identified 5 core clusters associated with distinct subtypes of resting state EEG frequency content. Bayesian models demonstrated substantial differences in psychological distress, sleep quality and cognitive function between clusters. By examining associations between neurophysiology and health measures across clusters, we have identified preliminary risk and protective profiles linked to EEG characteristics. CONCLUSION: This method provides the potential to identify neurophysiological subgroups of adolescents in the general population based on resting state EEG, and associated patterns of health and cognition that are not observed at the whole group level. This approach offers potential utility in clinical risk prediction for mental and cognitive health outcomes throughout adolescent development.

A longitudinal study of functional connectome uniqueness and its association with psychological distress in adolescence.

Each human brain has a unique functional synchronisation pattern (functional connectome) analogous to a fingerprint that underpins brain functions and related behaviours. Here we examine functional connectome (whole-brain and 13 networks) maturation by measuring its uniqueness in adolescents who underwent brain scans longitudinally from 12 years of age every four months. The uniqueness of a functional connectome is defined as its ratio of self-similarity (from the same subject at a different time point) to the maximal similarity-to-others (from a given subject and any others at a different time point). We found that the unique whole brain connectome exists in 12 years old adolescents, with 92% individuals having a whole brain uniqueness value greater than one. The cingulo-opercular network (CON; a long-acting 'brain control network' configuring information processing) demonstrated marginal uniqueness in early adolescence with 56% of individuals showing uniqueness greater than one (i.e., more similar to her/his own CON four months later than those from any other subjects) and this increased longitudinally. Notably, the low uniqueness of the CON correlates (ß = -18.6, FDR-Q < < 0.001) with K10 levels at the subsequent time point. This association suggests that the individualisation of CON network is related to psychological distress levels. Our findings highlight the potential of longitudinal neuroimaging to capture mental health problems in young people who are undergoing profound neuroplasticity and environment sensitivity period.

Sex differences in fronto-limbic white matter tracts in youth with mood disorders.

AIMS: Patients with depression and bipolar disorder have previously been shown to have impaired white matter (WM) integrity compared with healthy controls. This study aimed to investigate potential sex differences that may provide further insight into the pathophysiology of these highly debilitating mood disorders. METHODS: Participants aged 17 to 30 years (168 with depression [60% females], 107 with bipolar disorder [74% females], and 61 controls [64% females]) completed clinical assessment, self-report measures, and a neuropsychological assessment battery. Participants also underwent magnetic resonance imaging from which diffusion tensor imaging data were collected among five fronto-limbic WM tracts: cingulum bundle (cingulate gyrus and hippocampus subsections), fornix, stria terminalis, and the uncinate fasciculus. Mean fractional anisotropy (FA) scores were compared between groups using analyses of variance with sex and diagnosis as fixed factors. RESULTS: Among the nine WM tracts analyzed, one revealed a significant interaction between sex and diagnosis, controlling for age. Male patients with bipolar disorder had significantly lower FA scores in the fornix compared with the other groups. Furthermore, partial correlations revealed a significant positive association between FA scores for the fornix and psychomotor speed. CONCLUSIONS: Our findings suggest that males with bipolar disorder may be at increased risk of disruptions in WM integrity, especially in the fornix, which is thought to be responsible for a range of cognitive functions. More broadly, our findings suggest that sex differences may exist in WM integrity and thereby alter our understanding of the pathophysiology of mood disorders.

Phase-Amplitude Coupling, Mental Health and Cognition: Implications for Adolescence.

Identifying biomarkers of developing mental disorder is crucial to improving early identification and treatment-a key strategy for reducing the burden of mental disorders. Cross-frequency coupling between two different frequencies of neural oscillations is one such promising measure, believed to reflect synchronization between local and global networks in the brain. Specifically, in adults phase-amplitude coupling (PAC) has been shown to be involved in a range of cognitive processes, including working and long-term memory, attention, language, and fluid intelligence. Evidence suggests that increased PAC mediates both temporary and lasting improvements in working memory elicited by transcranial direct-current stimulation and reductions in depressive symptoms after transcranial magnetic stimulation. Moreover, research has shown that abnormal patterns of PAC are associated with depression and schizophrenia in adults. PAC is believed to be closely related to cortico-cortico white matter (WM) microstructure, which is well established in the literature as a structural mechanism underlying mental health. Some cognitive findings have been replicated in adolescents and abnormal patterns of PAC have also been linked to ADHD in young people. However, currently most research has focused on cross-sectional adult samples. Whereas initial hypotheses suggested that PAC was a state-based measure due to an early focus on cognitive, task-based research, current evidence suggests that PAC has both state-based and stable components. Future longitudinal research focusing on PAC throughout adolescent development could further our understanding of the relationship between mental health and cognition and facilitate the development of new methods for the identification and treatment of youth mental health.

White Matter Integrity According to the Stage of Mental Disorder in Youth.

The present study investigated differences in white matter (WM) integrity between 96 young people with affective and/or psychotic symptoms classified at an early stage of mental disorder (i.e. 'attenuated syndrome'; stage 1b), 85 young people classified at a more advanced stage of mental disorder (i.e. 'discrete disorder'; stage 2), and 81 demographically matched healthy controls using diffusion tensor imaging. The relationship between WM integrity (indexed by fractional anisotropy; FA) across the tracts and neuropsychological functioning was also investigated. A significant reduction in FA was identified in those with more advanced disorder in the body of the corpus callosum. Clinical stage groups were associated with significant neuropsychological impairment, which was significantly greater in those with discrete disorders. Compared to those in the earlier stage of disorder, participants at the later clinical stage showed decreased FA in the body of the corpus callosum that was associated with worse performance in attentional set formation maintenance, shifting and flexibility. These results provide further support for clinical staging of mental disorder and highlight the potential for utilising neuroanatomical biomarkers to support the classification of stages of mental disorder in the future.