Data-Driven, Generalizable Prediction of Adolescent Sleep Disturbances in the Multisite ABCD Study.

STUDY OBJECTIVES: Sleep disturbances are common in adolescence and associated with a host of negative outcomes. Here we assess associations between multifaceted sleep disturbances and a broad set of psychological, cognitive, and demographic variables using a data-driven approach, canonical correlation analysis (CCA). METHODS: Baseline data from 9,093 participants from the Adolescent Brain Cognitive Development℠ (ABCD) Study were examined using CCA, a multivariate statistical approach that identifies many-to-many associations between two sets of variables by finding combinations for each set of variables that maximize their correlation. We combined CCA with leave-one-site-out cross-validation across ABCD sites to examine the robustness of results and generalizability to new participants. The statistical significance of canonical correlations was determined by non-parametric permutation tests that accounted for twin, family, and site structure. To assess the stability of the associations identified at baseline, CCA was repeated using two-year follow-up data from 4,247 ABCD Study participants. RESULTS: Two significant sets of associations were identified: 1) difficulty initiating and maintaining sleep and excessive daytime somnolence were strongly linked to nearly all domains of psychopathology (r-squared = 0.36, p < .0001); 2) sleep breathing disorders were linked to BMI and African American/Black race (r-squared = 0.08, p < .0001). These associations generalized to unseen participants at all 22 ABCD sites and were replicated using two-year follow-up data. CONCLUSIONS: These findings underscore interwoven links between sleep disturbances in early adolescence and psychological, social, and demographic factors.

Socioeconomic resources in youth are linked to divergent patterns of network integration and segregation across the brain's transmodal axis.

Socioeconomic resources (SER) calibrate the developing brain to the current context, which can confer or attenuate risk for psychopathology across the lifespan. Recent multivariate work indicates that SER levels powerfully influence intrinsic functional connectivity patterns across the entire brain. Nevertheless, the neurobiological meaning of these widespread alterations remains poorly understood, despite its translational promise for early risk identification, targeted intervention, and policy reform. In the present study, we leverage the resources of graph theory to precisely characterize multivariate and univariate associations between household SER and the functional integration and segregation (i.e., participation coefficient, within-module degree) of brain regions across major cognitive, affective, and sensorimotor systems during the resting state in 5,821 youth (ages 9-10 years) from the Adolescent Brain Cognitive Development (ABCD) Study. First, we establish that decomposing the brain into profiles of integration and segregation captures more than half of the multivariate association between SER and functional connectivity with greater parsimony (100-fold reduction in number of features) and interpretability. Second, we show that the topological effects of SER are not uniform across the brain; rather, higher SER levels are related to greater integration of somatomotor and subcortical systems, but greater segregation of default mode, orbitofrontal, and cerebellar systems. Finally, we demonstrate that the effects of SER are spatially patterned along the unimodal-transmodal gradient of brain organization. These findings provide critical interpretive context for the established and widespread effects of SER on brain organization, indicating that SER levels differentially configure the intrinsic functional architecture of developing unimodal and transmodal systems. This study highlights both sensorimotor and higher-order networks that may serve as neural markers of environmental stress and opportunity, and which may guide efforts to scaffold healthy neurobehavioral development among disadvantaged communities of youth.

Socioeconomic resources are associated with distributed alterations of the brain's intrinsic functional architecture in youth.

Little is known about how exposure to limited socioeconomic resources (SER) in childhood gets "under the skin" to shape brain development, especially using rigorous whole-brain multivariate methods in large, adequately powered samples. The present study examined resting state functional connectivity patterns from 5821 youth in the Adolescent Brain Cognitive Development (ABCD) study, employing multivariate methods across three levels: whole-brain, network-wise, and connection-wise. Across all three levels, SER was associated with widespread alterations across the connectome. However, critically, we found that parental education was the primary driver of neural associations with SER. These parental education associations with the developing connectome exhibited notable concentrations in somatosensory and subcortical regions, and they were partially accounted for by home enrichment activities, child's cognitive abilities, and child's grades, indicating interwoven links between parental education, child stimulation, and child cognitive performance. These results add a new data-driven, multivariate perspective on links between household SER and the child's developing functional connectome.

Differentiated nomological networks of internalizing, externalizing, and the general factor of psychopathology ('p factor') in emerging adolescence in the ABCD study.

BACKGROUND: Structural models of psychopathology consistently identify internalizing (INT) and externalizing (EXT) specific factors as well as a superordinate factor that captures their shared variance, the p factor. Questions remain, however, about the meaning of these data-driven dimensions and the interpretability and distinguishability of the larger nomological networks in which they are embedded. METHODS: The sample consisted of 10 645 youth aged 9-10 years participating in the multisite Adolescent Brain and Cognitive Development (ABCD) Study. p, INT, and EXT were modeled using the parent-rated Child Behavior Checklist (CBCL). Patterns of associations were examined with variables drawn from diverse domains including demographics, psychopathology, temperament, family history of substance use and psychopathology, school and family environment, and cognitive ability, using instruments based on youth-, parent-, and teacher-report, and behavioral task performance. RESULTS: p exhibited a broad pattern of statistically significant associations with risk variables across all domains assessed, including temperament, neurocognition, and social adversity. The specific factors exhibited more domain-specific patterns of associations, with INT exhibiting greater fear/distress and EXT exhibiting greater impulsivity. CONCLUSIONS: In this largest study of hierarchical models of psychopathology to date, we found that p, INT, and EXT exhibit well-differentiated nomological networks that are interpretable in terms of neurocognition, impulsivity, fear/distress, and social adversity. These networks were, in contrast, obscured when relying on the a priori Internalizing and Externalizing dimensions of the CBCL scales. Our findings add to the evidence for the validity of p, INT, and EXT as theoretically and empirically meaningful broad psychopathology liabilities.

Widespread attenuating changes in brain connectivity associated with the general factor of psychopathology in 9- and 10-year olds.

Convergent research identifies a general factor ("P factor") that confers transdiagnostic risk for psychopathology. Large-scale networks are key organizational units of the human brain. However, studies of altered network connectivity patterns associated with the P factor are limited, especially in early adolescence when most mental disorders are first emerging. We studied 11,875 9- and 10-year olds from the Adolescent Brain and Cognitive Development (ABCD) study, of whom 6593 had high-quality resting-state scans. Network contingency analysis was used to identify altered interconnections associated with the P factor among 16 large-scale networks. These connectivity changes were then further characterized with quadrant analysis that quantified the directionality of P factor effects in relation to neurotypical patterns of positive versus negative connectivity across connections. The results showed that the P factor was associated with altered connectivity across 28 network cells (i.e., sets of connections linking pairs of networks); pPERMUTATION values < 0.05 FDR-corrected for multiple comparisons. Higher P factor scores were associated with hypoconnectivity within default network and hyperconnectivity between default network and multiple control networks. Among connections within these 28 significant cells, the P factor was predominantly associated with "attenuating" effects (67%; pPERMUTATION < 0.0002), i.e., reduced connectivity at neurotypically positive connections and increased connectivity at neurotypically negative connections. These results demonstrate that the general factor of psychopathology produces attenuating changes across multiple networks including default network, involved in spontaneous responses, and control networks involved in cognitive control. Moreover, they clarify mechanisms of transdiagnostic risk for psychopathology and invite further research into developmental causes of distributed attenuated connectivity.

Brain-wide functional connectivity patterns support general cognitive ability and mediate effects of socioeconomic status in youth.

General cognitive ability (GCA) is an individual difference dimension linked to important academic, occupational, and health-related outcomes and its development is strongly linked to differences in socioeconomic status (SES). Complex abilities of the human brain are realized through interconnections among distributed brain regions, but brain-wide connectivity patterns associated with GCA in youth, and the influence of SES on these connectivity patterns, are poorly understood. The present study examined functional connectomes from 5937 9- and 10-year-olds in the Adolescent Brain Cognitive Development (ABCD) multi-site study. Using multivariate predictive modeling methods, we identified whole-brain functional connectivity patterns linked to GCA. In leave-one-site-out cross-validation, we found these connectivity patterns exhibited strong and statistically reliable generalization at 19 out of 19 held-out sites accounting for 18.0% of the variance in GCA scores (cross-validated partial η2). GCA-related connections were remarkably dispersed across brain networks: across 120 sets of connections linking pairs of large-scale networks, significantly elevated GCA-related connectivity was found in 110 of them, and differences in levels of GCA-related connectivity across brain networks were notably modest. Consistent with prior work, socioeconomic status was a strong predictor of GCA in this sample, and we found that distributed GCA-related brain connectivity patterns significantly statistically mediated this relationship (mean proportion mediated: 15.6%, p < 2 × 10-16). These results demonstrate that socioeconomic status and GCA are related to broad and diffuse differences in functional connectivity architecture during early adolescence, potentially suggesting a mechanism through which socioeconomic status influences cognitive development.

The General Factor of Psychopathology in the Adolescent Brain Cognitive Development (ABCD) Study: A Comparison of Alternative Modeling Approaches.

Many models of psychopathology include a single general factor of psychopathology (GFP) or "p factor" to account for covariation across symptoms. The Adolescent Brain Cognitive Development (ABCD) Study provides a rich opportunity to study the development of the GFP. However, a variety of approaches for modeling the GFP have emerged, raising questions about how modeling choices impact estimated GFP scores. We used the ABCD baseline assessment (ages 9-10 years-old; N=11,875) of the parent-rated Child Behavior Checklist (CBCL) to examine the implications of modeling the GFP using items versus scales; using a priori CBCL scales versus data-driven dimensions; and using bifactor, higher-order, or single-factor models. Children's rank-ordering on the GFP was stable across models, with GFP scores similarly related to criterion variables. Results suggest that while theoretical debates about modeling the GFP continue, the practical implications of these choices for rank-ordering children and assessing external associations will often be modest.

Boost in Test-Retest Reliability in Resting State fMRI with Predictive Modeling.

Recent studies found low test-retest reliability in functional magnetic resonance imaging (fMRI), raising serious concerns among researchers, but these studies mostly focused on the reliability of individual fMRI features (e.g., individual connections in resting state connectivity maps). Meanwhile, neuroimaging researchers increasingly employ multivariate predictive models that aggregate information across a large number of features to predict outcomes of interest, but the test-retest reliability of predicted outcomes of these models has not previously been systematically studied. Here we apply 10 predictive modeling methods to resting state connectivity maps from the Human Connectome Project dataset to predict 61 outcome variables. Compared with mean reliability of individual resting state connections, we find mean reliability of the predicted outcomes of predictive models is substantially higher for all 10 modeling methods assessed. Moreover, improvement was consistently observed across all scanning and processing choices (i.e., scan lengths, censoring thresholds, volume- vs. surface-based processing). For the most reliable methods, the reliability of predicted outcomes was mostly, though not exclusively, in the "good" range (above 0.60). Finally, we identified three mechanisms that help to explain why predicted outcomes of predictive models have higher reliability than individual imaging features. We conclude that researchers can potentially achieve higher test-retest reliability by making greater use of predictive models.

Structure in the stream of consciousness: Evidence from a verbalized thought protocol and automated text analytic methods.

A key question about the spontaneous stream of thought (SST), often called the stream of consciousness, concerns its serial structure: How are thoughts in an extended sequence related to each other? In this study, we used a verbalized thought protocol to investigate "clump-and-jump" structure in SST-clusters of related thoughts about a topic followed by a jump to a new topic, in a repeating pattern. Several lines of evidence convergently supported the presence of clump-and-jump structure: high interrater agreement in identifying jumps, corroboration of rater-assigned jumps by automated text analytic methods, identification of clumps and jumps by a data-driven algorithm, and the inferred presence of clumps and jumps in unverbalized SST. We also found evidence that jumps involve a discontinuous shift in which a new clump is only modestly related to the previous one. These results illuminate serial structure in SST and invite research into the processes that generate the clump-and-jump pattern.

Toward a "treadmill test" for cognition: Improved prediction of general cognitive ability from the task activated brain.

General cognitive ability (GCA) refers to a trait-like ability that contributes to performance across diverse cognitive tasks. Identifying brain-based markers of GCA has been a longstanding goal of cognitive and clinical neuroscience. Recently, predictive modeling methods have emerged that build whole-brain, distributed neural signatures for phenotypes of interest. In this study, we employ a predictive modeling approach to predict GCA based on fMRI task activation patterns during the N-back working memory task as well as six other tasks in the Human Connectome Project dataset (n = 967), encompassing 15 task contrasts in total. We found tasks are a highly effective basis for prediction of GCA: The 2-back versus 0-back contrast achieved a 0.50 correlation with GCA scores in 10-fold cross-validation, and 13 out of 15 task contrasts afforded statistically significant prediction of GCA. Additionally, we found that task contrasts that produce greater frontoparietal activation and default mode network deactivation-a brain activation pattern associated with executive processing and higher cognitive demand-are more effective in the prediction of GCA. These results suggest a picture analogous to treadmill testing for cardiac function: Placing the brain in a more cognitively demanding task state significantly improves brain-based prediction of GCA.