Reproducible brain-wide association studies require thousands of individuals.

Magnetic resonance imaging (MRI) has transformed our understanding of the human brain through well-replicated mapping of abilities to specific structures (for example, lesion studies) and functions1-3 (for example, task functional MRI (fMRI)). Mental health research and care have yet to realize similar advances from MRI. A primary challenge has been replicating associations between inter-individual differences in brain structure or function and complex cognitive or mental health phenotypes (brain-wide association studies (BWAS)). Such BWAS have typically relied on sample sizes appropriate for classical brain mapping4 (the median neuroimaging study sample size is about 25), but potentially too small for capturing reproducible brain-behavioural phenotype associations5,6. Here we used three of the largest neuroimaging datasets currently available-with a total sample size of around 50,000 individuals-to quantify BWAS effect sizes and reproducibility as a function of sample size. BWAS associations were smaller than previously thought, resulting in statistically underpowered studies, inflated effect sizes and replication failures at typical sample sizes. As sample sizes grew into the thousands, replication rates began to improve and effect size inflation decreased. More robust BWAS effects were detected for functional MRI (versus structural), cognitive tests (versus mental health questionnaires) and multivariate methods (versus univariate). Smaller than expected brain-phenotype associations and variability across population subsamples can explain widespread BWAS replication failures. In contrast to non-BWAS approaches with larger effects (for example, lesions, interventions and within-person), BWAS reproducibility requires samples with thousands of individuals.

Hemispheric asymmetry in cortical thinning reflects intrinsic organization of the neurotransmitter systems and homotopic functional connectivity.

Hemispheric lateralization and its origins have been of great interest in neuroscience for over a century. The left-right asymmetry in cortical thickness may stem from differential maturation of the cerebral cortex in the two hemispheres. Here, we investigated the spatial pattern of hemispheric differences in cortical thinning during adolescence, and its relationship with the density of neurotransmitter receptors and homotopic functional connectivity. Using longitudinal data from IMAGEN study (N = 532), we found that many cortical regions in the frontal and temporal lobes thinned more in the right hemisphere than in the left. Conversely, several regions in the occipital and parietal lobes thinned less in the right (vs. left) hemisphere. We then revealed that regions thinning more in the right (vs. left) hemispheres had higher density of neurotransmitter receptors and transporters in the right (vs. left) side. Moreover, the hemispheric differences in cortical thinning were predicted by homotopic functional connectivity. Specifically, regions with stronger homotopic functional connectivity showed a more symmetrical rate of cortical thinning between the left and right hemispheres, compared with regions with weaker homotopic functional connectivity. Based on these findings, we suggest that the typical patterns of hemispheric differences in cortical thinning may reflect the intrinsic organization of the neurotransmitter systems and related patterns of homotopic functional connectivity.

Coupled changes between ruminating thoughts and resting-state brain networks during the transition into adulthood.

Perseverative negative thoughts, known as rumination, might arise from emotional challenges and preclude mental health when transitioning into adulthood. Due to its multifaceted nature, rumination can take several ruminative response styles, that diverge in manifestations, severity, and mental health outcomes. Still, prospective ruminative phenotypes remain elusive insofar. Longitudinal study designs are ideal for stratifying ruminative response styles, especially with resting-state functional MRI whose setup naturally elicits people's ruminative traits. Here, we considered self-rated questionnaires on rumination and psychopathology, along with resting-state functional MRI data in 595 individuals assessed at age 18 and 22 from the IMAGEN cohort. We conducted independent component analysis to characterize eight single static resting-state functional networks in each subject and session and furthermore conducted a dynamic analysis, tackling the time variations of functional networks during the entire scanning time. We then investigated their longitudinal mediation role between changes in three ruminative response styles (reflective pondering, brooding, and depressive rumination) and changes in internalizing and co-morbid externalizing symptoms. Four static and two dynamic networks longitudinally differentiated these ruminative styles and showed complemental sensitivity to internalizing and co-morbid externalizing symptoms. Among these networks, the right frontoparietal network covaried with all ruminative styles but did not play any mediation role towards psychopathology. The default mode, the salience, and the limbic networks prospectively stratified these ruminative styles, suggesting that maladaptive ruminative styles are associated with altered corticolimbic function. For static measures, only the salience network played a longitudinal causal role between brooding rumination and internalizing symptoms. Dynamic measures highlighted the default-mode mediation role between the other ruminative styles and co-morbid externalizing symptoms. In conclusion, we identified the ruminative styles' psychometric and neural outcome specificities, supporting their translation into applied research on young adult mental healthcare.

Leveraging the adolescent brain cognitive development study to improve behavioral prediction from neuroimaging in smaller replication samples.

Neuroimaging is a popular method to map brain structural and functional patterns to complex human traits. Recently published observations cast doubt upon these prospects, particularly for prediction of cognitive traits from structural and resting state functional magnetic resonance imaging (MRI). We leverage baseline data from thousands of children in the Adolescent Brain Cognitive DevelopmentSM Study to inform the replication sample size required with univariate and multivariate methods across different imaging modalities to detect reproducible brain-behavior associations. We demonstrate that by applying multivariate methods to high-dimensional brain imaging data, we can capture lower dimensional patterns of structural and functional brain architecture that correlate robustly with cognitive phenotypes and are reproducible with only 41 individuals in the replication sample for working memory-related functional MRI, and ~ 100 subjects for structural and resting state MRI. Even with 100 random re-samplings of 100 subjects in discovery, prediction can be adequately powered with 66 subjects in replication for multivariate prediction of cognition with working memory task functional MRI. These results point to an important role for neuroimaging in translational neurodevelopmental research and showcase how findings in large samples can inform reproducible brain-behavior associations in small sample sizes that are at the heart of many research programs and grants.

Unraveling the molecular relevance of brain phenotypes: A comparative analysis of null models and test statistics.

Correlating transcriptional profiles with imaging-derived phenotypes has the potential to reveal possible molecular architectures associated with cognitive functions, brain development and disorders. Competitive null models built by resampling genes and self-contained null models built by spinning brain regions, along with varying test statistics, have been used to determine the significance of transcriptional associations. However, there has been no systematic evaluation of their performance in imaging transcriptomics analyses. Here, we evaluated the performance of eight different test statistics (mean, mean absolute value, mean squared value, max mean, median, Kolmogorov-Smirnov (KS), Weighted KS and the number of significant correlations) in both competitive null models and self-contained null models. Simulated brain maps (n = 1,000) and gene sets (n = 500) were used to calculate the probability of significance (Psig) for each statistical test. Our results suggested that competitive null models may result in false positive results driven by co-expression within gene sets. Furthermore, we demonstrated that the self-contained null models may fail to account for distribution characteristics (e.g., bimodality) of correlations between all available genes and brain phenotypes, leading to false positives. These two confounding factors interacted differently with test statistics, resulting in varying outcomes. Specifically, the sign-sensitive test statistics (i.e., mean, median, KS, Weighted KS) were influenced by co-expression bias in the competitive null models, while median and sign-insensitive test statistics were sensitive to the bimodality bias in the self-contained null models. Additionally, KS-based statistics produced conservative results in the self-contained null models, which increased the risk of false negatives. Comprehensive supplementary analyses with various configurations, including realistic scenarios, supported the results. These findings suggest utilizing sign-insensitive test statistics such as mean absolute value, max mean in the competitive null models and the mean as the test statistic for the self-contained null models. Additionally, adopting the confounder-matched (e.g., coexpression-matched) null models as an alternative to standard null models can be a viable strategy. Overall, the present study offers insights into the selection of statistical tests for imaging transcriptomics studies, highlighting areas for further investigation and refinement in the evaluation of novel and commonly used tests.

Adolescent to young adult longitudinal development of subcortical volumes in two European sites with four waves.

Adolescent subcortical structural brain development might underlie psychopathological symptoms, which often emerge in adolescence. At the same time, sex differences exist in psychopathology, which might be mirrored in underlying sex differences in structural development. However, previous studies showed inconsistencies in subcortical trajectories and potential sex differences. Therefore, we aimed to investigate the subcortical structural trajectories and their sex differences across adolescence using for the first time a single cohort design, the same quality control procedure, software, and a general additive mixed modeling approach. We investigated two large European sites from ages 14 to 24 with 503 participants and 1408 total scans from France and Germany as part of the IMAGEN project including four waves of data acquisition. We found significantly larger volumes in males versus females in both sites and across all seven subcortical regions. Sex differences in age-related trajectories were observed across all regions in both sites. Our findings provide further evidence of sex differences in longitudinal adolescent brain development of subcortical regions and thus might eventually support the relationship of underlying brain development and different adolescent psychopathology in boys and girls.

Genetics impact risk of Alzheimer's disease through mechanisms modulating structural brain morphology in late life.

BACKGROUND: Alzheimer's disease (AD)-related neuropathological changes can occur decades before clinical symptoms. We aimed to investigate whether neurodevelopment and/or neurodegeneration affects the risk of AD, through reducing structural brain reserve and/or increasing brain atrophy, respectively. METHODS: We used bidirectional two-sample Mendelian randomisation to estimate the effects between genetic liability to AD and global and regional cortical thickness, estimated total intracranial volume, volume of subcortical structures and total white matter in 37 680 participants aged 8-81 years across 5 independent cohorts (Adolescent Brain Cognitive Development, Generation R, IMAGEN, Avon Longitudinal Study of Parents and Children and UK Biobank). We also examined the effects of global and regional cortical thickness and subcortical volumes from the Enhancing NeuroImaging Genetics through Meta-Analysis (ENIGMA) Consortium on AD risk in up to 37 741 participants. RESULTS: Our findings show that AD risk alleles have an age-dependent effect on a range of cortical and subcortical brain measures that starts in mid-life, in non-clinical populations. Evidence for such effects across childhood and young adulthood is weak. Some of the identified structures are not typically implicated in AD, such as those in the striatum (eg, thalamus), with consistent effects from childhood to late adulthood. There was little evidence to suggest brain morphology alters AD risk. CONCLUSIONS: Genetic liability to AD is likely to affect risk of AD primarily through mechanisms affecting indicators of brain morphology in later life, rather than structural brain reserve. Future studies with repeated measures are required for a better understanding and certainty of the mechanisms at play.

PACE: a Novel Eating Behavior Phenotype to Assess Risk for Obesity in Middle Childhood.

BACKGROUND: Behavioral phenotypes that predict future weight gain are needed to identify children susceptible to obesity. OBJECTIVES: This prospective study developed an eating behavior risk score to predict change in adiposity over 1 y in children. METHODS: Data from 6 baseline visits (Time 1, T1) and a 1-y follow-up visit (Time 2, T2) were collected from 76, 7- to 8-y-old healthy children recruited from Central Pennsylvania. At T1, children had body mass index (BMI) percentiles <90 and were classified with either high (n = 33; maternal BMI ≥30 kg/m2) or low (n = 43; maternal BMI ≤25 kg/m2) familial risk for obesity. Appetitive traits and eating behaviors were assessed at T1. Adiposity was measured at T1 and T2 using dual-energy x-ray absorptiometry, with a main outcome of fat mass index (FMI; total body fat mass divided by height in meters squared). Hierarchical linear regressions determined which eating measures improved prediction of T2 FMI after adjustment for covariates in the baseline model (T1 FMI, sex, income, familial risk, and Tanner stage). RESULTS: Four eating measures-Portion susceptibility, Appetitive traits, loss of control eating, and eating rate-were combined into a standardized summary score called PACE. PACE improved the baseline model to predict 80% variance in T2 FMI. PACE was positively associated with the increase in FMI in children from T1 to T2, independent of familial risk (r = 0.58, P < 0.001). Although PACE was higher in girls than boys (P < 0.05), it did not differ by familial risk, income, or education. CONCLUSIONS: PACE represents a cumulative eating behavior risk score that predicts adiposity gain over 1 y in middle childhood. If PACE similarly predicts adiposity gain in a cohort with greater racial and socioeconomic diversity, it will inform the development of interventions to prevent obesity. This trial was registered at clinicaltrials.gov as NCT03341247.

Trans-ancestry meta-analysis of genome wide association studies of inhibitory control.

Deficits in effective executive function, including inhibitory control are associated with risk for a number of psychiatric disorders and significantly impact everyday functioning. These complex traits have been proposed to serve as endophenotypes, however, their genetic architecture is not yet well understood. To identify the common genetic variation associated with inhibitory control in the general population we performed the first trans-ancestry genome wide association study (GWAS) combining data across 8 sites and four ancestries (N = 14,877) using cognitive traits derived from the stop-signal task, namely - go reaction time (GoRT), go reaction time variability (GoRT SD) and stop signal reaction time (SSRT). Although we did not identify genome wide significant associations for any of the three traits, GoRT SD and SSRT demonstrated significant and similar SNP heritability of 8.2%, indicative of an influence of genetic factors. Power analyses demonstrated that the number of common causal variants contributing to the heritability of these phenotypes is relatively high and larger sample sizes are necessary to robustly identify associations. In Europeans, the polygenic risk for ADHD was significantly associated with GoRT SD and the polygenic risk for schizophrenia was associated with GoRT, while in East Asians polygenic risk for schizophrenia was associated with SSRT. These results support the potential of executive function measures as endophenotypes of neuropsychiatric disorders. Together these findings provide the first evidence indicating the influence of common genetic variation in the genetic architecture of inhibitory control quantified using objective behavioural traits derived from the stop-signal task.

Anxiety onset in adolescents: a machine-learning prediction.

Recent longitudinal studies in youth have reported MRI correlates of prospective anxiety symptoms during adolescence, a vulnerable period for the onset of anxiety disorders. However, their predictive value has not been established. Individual prediction through machine-learning algorithms might help bridge the gap to clinical relevance. A voting classifier with Random Forest, Support Vector Machine and Logistic Regression algorithms was used to evaluate the predictive pertinence of gray matter volumes of interest and psychometric scores in the detection of prospective clinical anxiety. Participants with clinical anxiety at age 18-23 (N = 156) were investigated at age 14 along with healthy controls (N = 424). Shapley values were extracted for in-depth interpretation of feature importance. Prospective prediction of pooled anxiety disorders relied mostly on psychometric features and achieved moderate performance (area under the receiver operating curve = 0.68), while generalized anxiety disorder (GAD) prediction achieved similar performance. MRI regional volumes did not improve the prediction performance of prospective pooled anxiety disorders with respect to psychometric features alone, but they improved the prediction performance of GAD, with the caudate and pallidum volumes being among the most contributing features. To conclude, in non-anxious 14 year old adolescents, future clinical anxiety onset 4-8 years later could be individually predicted. Psychometric features such as neuroticism, hopelessness and emotional symptoms were the main contributors to pooled anxiety disorders prediction. Neuroanatomical data, such as caudate and pallidum volume, proved valuable for GAD and should be included in prospective clinical anxiety prediction in adolescents.

Arc controls alcohol cue relapse by a central amygdala mechanism.

Alcohol use disorder (AUD) is a chronic and fatal disease. The main impediment of the AUD therapy is a high probability of relapse to alcohol abuse even after prolonged abstinence. The molecular mechanisms of cue-induced relapse are not well established, despite the fact that they may offer new targets for the treatment of AUD. Using a comprehensive animal model of AUD, virally-mediated and amygdala-targeted genetic manipulations by CRISPR/Cas9 technology and ex vivo electrophysiology, we identify a mechanism that selectively controls cue-induced alcohol relapse and AUD symptom severity. This mechanism is based on activity-regulated cytoskeleton-associated protein (Arc)/ARG3.1-dependent plasticity of the amygdala synapses. In humans, we identified single nucleotide polymorphisms in the ARC gene and their methylation predicting not only amygdala size, but also frequency of alcohol use, even at the onset of regular consumption. Targeting Arc during alcohol cue exposure may thus be a selective new mechanism for relapse prevention.

Differential associations of adolescent versus young adult cannabis initiation with longitudinal brain change and behavior.

Leveraging ~10 years of prospective longitudinal data on 704 participants, we examined the effects of adolescent versus young adult cannabis initiation on MRI-assessed cortical thickness development and behavior. Data were obtained from the IMAGEN study conducted across eight European sites. We identified IMAGEN participants who reported being cannabis-naïve at baseline and had data available at baseline, 5-year, and 9-year follow-up visits. Cannabis use was assessed with the European School Survey Project on Alcohol and Drugs. T1-weighted MR images were processed through the CIVET pipeline. Cannabis initiation occurring during adolescence (14-19 years) and young adulthood (19-22 years) was associated with differing patterns of longitudinal cortical thickness change. Associations between adolescent cannabis initiation and cortical thickness change were observed primarily in dorso- and ventrolateral portions of the prefrontal cortex. In contrast, cannabis initiation occurring between 19 and 22 years of age was associated with thickness change in temporal and cortical midline areas. Follow-up analysis revealed that longitudinal brain change related to adolescent initiation persisted into young adulthood and partially mediated the association between adolescent cannabis use and past-month cocaine, ecstasy, and cannabis use at age 22. Extent of cannabis initiation during young adulthood (from 19 to 22 years) had an indirect effect on psychotic symptoms at age 22 through thickness change in temporal areas. Results suggest that developmental timing of cannabis exposure may have a marked effect on neuroanatomical correlates of cannabis use as well as associated behavioral sequelae. Critically, this work provides a foundation for neurodevelopmentally informed models of cannabis exposure in humans.

Covariation of preadult environmental exposures, adult brain imaging phenotypes, and adult personality traits.

Exposure to preadult environmental exposures may have long-lasting effects on mental health by affecting the maturation of the brain and personality, two traits that interact throughout the developmental process. However, environment-brain-personality covariation patterns and their mediation relationships remain unclear. In 4297 healthy participants (aged 18-30 years), we combined sparse multiple canonical correlation analysis with independent component analysis to identify the three-way covariation patterns of 59 preadult environmental exposures, 760 adult brain imaging phenotypes, and five personality traits, and found two robust environment-brain-personality covariation models with sex specificity. One model linked greater stress and less support to weaker functional connectivity and activity in the default mode network, stronger activity in subcortical nuclei, greater thickness and volume in the occipital, parietal and temporal cortices, and lower agreeableness, consciousness and extraversion as well as higher neuroticism. The other model linked higher urbanicity and better socioeconomic status to stronger functional connectivity and activity in the sensorimotor network, smaller volume and surface area and weaker functional connectivity and activity in the medial prefrontal cortex, lower white matter integrity, and higher openness to experience. We also conducted mediation analyses to explore the potential bidirectional mediation relationships between adult brain imaging phenotypes and personality traits with the influence of preadult environmental exposures and found both environment-brain-personality and environment-personality-brain pathways. We finally performed moderated mediation analyses to test the potential interactions between macro- and microenvironmental exposures and found that one category of exposure moderated the mediation pathways of another category of exposure. These results improve our understanding of the effects of preadult environmental exposures on the adult brain and personality traits and may facilitate the design of targeted interventions to improve mental health by reducing the impact of adverse environmental exposures.

Task fMRI paradigms may capture more behaviorally relevant information than resting-state functional connectivity.

Characterizing the optimal fMRI paradigms for detecting behaviorally relevant functional connectivity (FC) patterns is a critical step to furthering our knowledge of the neural basis of behavior. Previous studies suggested that FC patterns derived from task fMRI paradigms, which we refer to as task-based FC, are better correlated with individual differences in behavior than resting-state FC, but the consistency and generalizability of this advantage across task conditions was not fully explored. Using data from resting-state fMRI and three fMRI tasks from the Adolescent Brain Cognitive Development Study ® (ABCD), we tested whether the observed improvement in behavioral prediction power of task-based FC can be attributed to changes in brain activity induced by the task design. We decomposed the task fMRI time course of each task into the task model fit (the fitted time course of the task condition regressors from the single-subject general linear model) and the task model residuals, calculated their respective FC, and compared the behavioral prediction performance of these FC estimates to resting-state FC and the original task-based FC. The FC of the task model fit was better than the FC of the task model residual and resting-state FC at predicting a measure of general cognitive ability or two measures of performance on the fMRI tasks. The superior behavioral prediction performance of the FC of the task model fit was content-specific insofar as it was only observed for fMRI tasks that probed similar cognitive constructs to the predicted behavior of interest. To our surprise, the task model parameters, the beta estimates of the task condition regressors, were equally if not more predictive of behavioral differences than all FC measures. These results showed that the observed improvement of behavioral prediction afforded by task-based FC was largely driven by the FC patterns associated with the task design. Together with previous studies, our findings highlighted the importance of task design in eliciting behaviorally meaningful brain activation and FC patterns.

Machine learning approaches linking brain function to behavior in the ABCD STOP task.

The stop-signal task (SST) is one of the most common fMRI tasks of response inhibition, and its performance measure, the stop-signal reaction-time (SSRT), is broadly used as a measure of cognitive control processes. The neurobiology underlying individual or clinical differences in response inhibition remain unclear, consistent with the general pattern of quite modest brain-behavior associations that have been recently reported in well-powered large-sample studies. Here, we investigated the potential of multivariate, machine learning (ML) methods to improve the estimation of individual differences in SSRT with multimodal structural and functional region of interest-level neuroimaging data from 9- to 11-year-olds children in the ABCD Study. Six ML algorithms were assessed across modalities and fMRI tasks. We verified that SST activation performed best in predicting SSRT among multiple modalities including morphological MRI (cortical surface area/thickness), diffusion tensor imaging, and fMRI task activations, and then showed that SST activation explained 12% of the variance in SSRT using cross-validation and out-of-sample lockbox data sets (n = 7298). Brain regions that were more active during the task and that showed more interindividual variation in activation were better at capturing individual differences in performance on the task, but this was only true for activations when successfully inhibiting. Cortical regions outperformed subcortical areas in explaining individual differences but the two hemispheres performed equally well. These results demonstrate that the detection of reproducible links between brain function and performance can be improved with multivariate approaches and give insight into a number of brain systems contributing to individual differences in this fundamental cognitive control process.

Brain volumes in alcohol use disorder: Do females and males differ? A whole-brain magnetic resonance imaging mega-analysis.

Emerging evidence suggests distinct neurobiological correlates of alcohol use disorder (AUD) between sexes, which however remain largely unexplored. This work from ENIGMA Addiction Working Group aimed to characterize the sex differences in gray matter (GM) and white matter (WM) correlates of AUD using a whole-brain, voxel-based, multi-tissue mega-analytic approach, thereby extending our recent surface-based region of interest findings on a nearly matching sample using a complementary methodological approach. T1-weighted magnetic resonance imaging (MRI) data from 653 people with AUD and 326 controls was analyzed using voxel-based morphometry. The effects of group, sex, group-by-sex, and substance use severity in AUD on brain volumes were assessed using General Linear Models. Individuals with AUD relative to controls had lower GM volume in striatal, thalamic, cerebellar, and widespread cortical clusters. Group-by-sex effects were found in cerebellar GM and WM volumes, which were more affected by AUD in females than males. Smaller group-by-sex effects were also found in frontotemporal WM tracts, which were more affected in AUD females, and in temporo-occipital and midcingulate GM volumes, which were more affected in AUD males. AUD females but not males showed a negative association between monthly drinks and precentral GM volume. Our results suggest that AUD is associated with both shared and distinct widespread effects on GM and WM volumes in females and males. This evidence advances our previous region of interest knowledge, supporting the usefulness of adopting an exploratory perspective and the need to include sex as a relevant moderator variable in AUD.

Independent contribution of polygenic risk for schizophrenia and cannabis use in predicting psychotic-like experiences in young adulthood: testing gene × environment moderation and mediation.

BACKGROUND: It has not yet been determined if the commonly reported cannabis-psychosis association is limited to individuals with pre-existing genetic risk for psychotic disorders. METHODS: We examined whether the relationship between polygenic risk score for schizophrenia (PRS-Sz) and psychotic-like experiences (PLEs), as measured by the Community Assessment of Psychic Experiences-42 (CAPE-42) questionnaire, is mediated or moderated by lifetime cannabis use at 16 years of age in 1740 of the individuals of the European IMAGEN cohort. Secondary analysis examined the relationships between lifetime cannabis use, PRS-Sz and the various sub-scales of the CAPE-42. Sensitivity analyses including covariates, including a PRS for cannabis use, were conducted and results were replicated using data from 1223 individuals in the Dutch Utrecht cannabis cohort. RESULTS: PRS-Sz significantly predicted cannabis use (p = 0.027) and PLE (p = 0.004) in the IMAGEN cohort. In the full model, considering PRS-Sz and covariates, cannabis use was also significantly associated with PLE in IMAGEN (p = 0.007). Results remained consistent in the Utrecht cohort and through sensitivity analyses. Nevertheless, there was no evidence of a mediation or moderation effects. CONCLUSIONS: These results suggest that cannabis use remains a risk factor for PLEs, over and above genetic vulnerability for schizophrenia. This research does not support the notion that the cannabis-psychosis link is limited to individuals who are genetically predisposed to psychosis and suggests a need for research focusing on cannabis-related processes in psychosis that cannot be explained by genetic vulnerability.

Trajectories of cortical structures associated with stress across adolescence: a bivariate latent change score approach.

BACKGROUND: Stress exposure in childhood and adolescence has been linked to reductions in cortical structures and cognitive functioning. However, to date, most of these studies have been cross-sectional, limiting the ability to make long-term inferences, given that most cortical structures continue to develop through adolescence. METHODS: Here, we used a subset of the IMAGEN population cohort sample (N = 502; assessment ages: 14, 19, and 22 years; mean age: 21.945 years; SD = 0.610) to understand longitudinally the long-term interrelations between stress, cortical development, and cognitive functioning. To these ends, we first used a latent change score model to examine four bivariate relations - assessing individual differences in change in the relations between adolescent stress exposure and volume, surface area, and cortical thickness of cortical structures, as well as cognitive outcomes. Second, we probed for indirect neurocognitive effects linking stress to cortical brain structures and cognitive functions using rich longitudinal mediation modeling. RESULTS: Latent change score modeling showed that greater baseline adolescence stress at age 14 predicted a small reduction in the right anterior cingulate volume (Std. ß = -.327, p = .042, 95% CI [-0.643, -0.012]) and right anterior cingulate surface area (Std. ß = -.274, p = .038, 95% CI [-0.533, -0.015]) across ages 14-22. These effects were very modest in nature and became nonsignificant after correcting for multiple comparisons. Our longitudinal analyses found no evidence of indirect effects in the two neurocognitive pathways linking adolescent stress to brain and cognitive outcomes. CONCLUSION: Findings shed light on the impact of stress on brain reductions, particularly in the prefrontal cortex that have consistently been implicated in the previous cross-sectional studies. However, the magnitude of effects observed in our study is smaller than that has been reported in past cross-sectional work. This suggests that the potential impact of stress during adolescence on brain structures may likely be more modest than previously noted.

Performance scaling for structural MRI surface parcellations: a machine learning analysis in the ABCD Study.

The use of predefined parcellations on surface-based representations of the brain as a method for data reduction is common across neuroimaging studies. In particular, prediction-based studies typically employ parcellation-driven summaries of brain measures as input to predictive algorithms, but the choice of parcellation and its influence on performance is often ignored. Here we employed preprocessed structural magnetic resonance imaging (sMRI) data from the Adolescent Brain Cognitive Development Study® to examine the relationship between 220 parcellations and out-of-sample predictive performance across 45 phenotypic measures in a large sample of 9- to 10-year-old children (N = 9,432). Choice of machine learning (ML) pipeline and use of alternative multiple parcellation-based strategies were also assessed. Relative parcellation performance was dependent on the spatial resolution of the parcellation, with larger number of parcels (up to ~4,000) outperforming coarser parcellations, according to a power-law scaling of between 1/4 and 1/3. Performance was further influenced by the type of parcellation, ML pipeline, and general strategy, with existing literature-based parcellations, a support vector-based pipeline, and ensembling across multiple parcellations, respectively, as the highest performing. These findings highlight the choice of parcellation as an important influence on downstream predictive performance, showing in some cases that switching to a higher resolution parcellation can yield a relatively large boost to performance.