Brain-behaviour modes of covariation in healthy and clinically depressed young people

Association between hearing ability and cortical morphology in the elderly: multiparametric mapping, cognitive relevance, and neurobiological underpinnings

BACKGROUND

Hearing impairment is a common condition in the elderly. However, a comprehensive understanding of its neural correlates is still lacking.

METHODS

We recruited 284 elderly adults who underwent structural MRI, magnetic resonance spectroscopy, audiometry, and cognitive assessments. Individual hearing abilities indexed by pure tone average (PTA) were correlated with multiple structural MRI-derived cortical morphological indices. For regions showing significant correlations, mediation analyses were performed to examine their role in the relationship between hearing ability and cognitive function. Finally, the correlation maps between hearing ability and cortical morphology were linked with publicly available connectomic gradient, transcriptomic, and neurotransmitter maps.

FINDINGS

Poorer hearing was related to cortical thickness (CT) reductions in widespread regions and gyrification index (GI) reductions in the right Area 52 and Insular Granular Complex. The GI in the right Area 52 mediated the relationship between hearing ability and executive function. This mediating effect was further modulated by glutamate and N-acetylaspartate levels in the right auditory region. The PTA-CT correlation map followed microstructural connectomic hierarchy, were related to genes involved in certain biological processes (e.g., glutamate metabolic process), cell types (e.g., excitatory neurons and astrocytes), and developmental stages (i.e., childhood to young adulthood), and covaried with dopamine receptor 1, dopamine transporter, and fluorodopa. The PTA-GI correlation map was related to 5-hydroxytryptamine receptor 2a.

INTERPRETATION

Poorer hearing is associated with cortical thinning and folding reductions, which may be engaged in the relationship between hearing impairment and cognitive decline in the elderly and have different neurobiological substrates.

FUNDING

See the Acknowledgements section.

A distinctive subcortical functional connectivity pattern linking negative affect and treatment outcome in major depressive disorder

Major depressive disorder (MDD) is associated with functional disturbances in subcortical regions. In this naturalistic prospective study (NCT03294525), we aimed to investigate relationships among subcortical functional connectivity (FC), mood symptom profiles and treatment outcome in MDD using multivariate methods. Medication-free participants with MDD (n = 135) underwent a functional magnetic resonance imaging scan at baseline and completed posttreatment clinical assessment after 8 weeks of antidepressant monotherapy. We used partial least squares (PLS) correlation analysis to explore the association between subcortical FC and mood symptom profiles. FC score, reflecting the weighted representation of each individual in this association, was computed. Replication analysis was undertaken in an independent sample (n = 74). We also investigated the relationship between FC score and treatment outcome in the main sample. A distinctive subcortical connectivity pattern was found to be associated with negative affect. In general, higher FC between the caudate, putamen and thalamus was associated with greater negative affect. This association was partly replicated in the independent sample (similarity between the two samples: r = 0.66 for subcortical connectivity, r = 0.75 for mood symptom profile). Lower FC score predicted both remission and response to treatment after 8 weeks of antidepressant monotherapy. The emphasis here on the role of dorsal striatum and thalamus consolidates prior work of subcortical connectivity in MDD. The findings provide insight into the pathogenesis of MDD, linking subcortical FC with negative affect. However, while the FC score significantly predicted treatment outcome, the low odds ratio suggests that finding predictive biomarkers for depression remains an aspiration.

Dimensional and Categorical Solutions to Parsing Depression Heterogeneity in a Large Single-Site Sample

BACKGROUND

Recent studies have reported significant advances in modeling the biological basis of heterogeneity in major depressive disorder, but investigators have also identified important technical challenges, including scanner-related artifacts, a propensity for multivariate models to overfit, and a need for larger samples with more extensive clinical phenotyping. The goals of the current study were to evaluate dimensional and categorical solutions to parsing heterogeneity in depression that are stable and generalizable in a large, single-site sample.

METHODS

We used regularized canonical correlation analysis to identify data-driven brain-behavior dimensions that explain individual differences in depression symptom domains in a large, single-site dataset comprising clinical assessments and resting-state functional magnetic resonance imaging data for 328 patients with major depressive disorder and 461 healthy control participants. We examined the stability of clinical loadings and model performance in held-out data. Finally, hierarchical clustering on these dimensions was used to identify categorical depression subtypes.

RESULTS

The optimal regularized canonical correlation analysis model yielded 3 robust and generalizable brain-behavior dimensions that explained individual differences in depressed mood and anxiety, anhedonia, and insomnia. Hierarchical clustering identified 4 depression subtypes, each with distinct clinical symptom profiles, abnormal resting-state functional connectivity patterns, and antidepressant responsiveness to repetitive transcranial magnetic stimulation.

CONCLUSIONS

Our results define dimensional and categorical solutions to parsing neurobiological heterogeneity in major depressive disorder that are stable, generalizable, and capable of predicting treatment outcomes, each with distinct advantages in different contexts. They also provide additional evidence that regularized canonical correlation analysis and hierarchical clustering are effective tools for investigating associations between functional connectivity and clinical symptoms.

Symptom clustering of major depression in a national telehealth sample

BACKGROUND

Major depressive disorder (MDD) is a heterogeneous disorder whose possible symptom combinations have not been well delineated. The aim of this study was to explore the heterogeneity of symptoms experienced by those with MDD to characterize phenotypic presentations.

METHODS

Cross-sectional data (N = 10,158) from a large telemental health platform were used to identify subtypes of MDD. Symptom data, gathered from both clinically-validated surveys and intake questions, were analyzed via polychoric correlations, principal component analysis, and cluster analysis.

RESULTS

Principal components analysis (PCA) of baseline symptom data revealed 5 components, including anxious distress, core emotional, agitation/irritability, insomnia, and anergic/apathy components. PCA-based cluster analysis resulted in four MDD phenotypes, the largest of which was characterized by a prominent elevation on the anergic/apathy component, but also core emotional. The four clusters differed on demographic and clinical characteristics.

LIMITATIONS

The primary limitation of this study is that the phenotypes uncovered are limited by the questions asked. These phenotypes will need to be cross validated with other samples, potentially expanded to include biological/genetic variables, and followed longitudinally.

CONCLUSIONS

The heterogeneity in MDD, as illustrated by the phenotypes in this sample, may explain the heterogeneity of treatment response in large-scale treatment trials. These phenotypes can be used to study varying rates of recovery following treatment and to develop clinical decision support tools and artificial intelligence algorithms. Strengths of this study include its size, breadth of included symptoms, and novel use of a telehealth platform.

Multivariate analytical approaches for investigating brain-behavior relationships

Background

Many studies of brain-behavior relationships rely on univariate approaches where each variable of interest is tested independently, which does not allow for the simultaneous investigation of multiple correlated variables. Alternatively, multivariate approaches allow for examining relationships between psychopathology and neural substrates simultaneously. There are multiple multivariate methods to choose from that each have assumptions which can affect the results; however, many studies employ one method without a clear justification for its selection. Additionally, there are few studies illustrating how differences between methods manifest in examining brain-behavior relationships. The purpose of this study was to exemplify how the choice of multivariate approach can change brain-behavior interpretations.

Method

We used data from 9,027 9- to 10-year-old children from the Adolescent Brain Cognitive DevelopmentSM Study (ABCD Study®) to examine brain-behavior relationships with three commonly used multivariate approaches: canonical correlation analysis (CCA), partial least squares correlation (PLSC), and partial least squares regression (PLSR). We examined the associations between psychopathology dimensions including general psychopathology, attention-deficit/hyperactivity symptoms, conduct problems, and internalizing symptoms with regional brain volumes.

Results

The results of CCA, PLSC, and PLSR showed both consistencies and differences in the relationship between psychopathology symptoms and brain structure. The leading significant component yielded by each method demonstrated similar patterns of associations between regional brain volumes and psychopathology symptoms. However, the additional significant components yielded by each method demonstrated differential brain-behavior patterns that were not consistent across methods.

Conclusion

Here we show that CCA, PLSC, and PLSR yield slightly different interpretations regarding the relationship between child psychopathology and brain volume. In demonstrating the divergence between these approaches, we exemplify the importance of carefully considering the method's underlying assumptions when choosing a multivariate approach to delineate brain-behavior relationships.

Comparing the stability and reproducibility of brain-behaviour relationships found using Canonical Correlation Analysis and Partial Least Squares within the ABCD Sample

Introduction

Canonical Correlation Analysis (CCA) and Partial Least Squares Correlation (PLS) detect associations between two data matrices based on computing a linear combination between the two matrices (called latent variables; LVs). These LVs maximize correlation (CCA) and covariance (PLS). These different maximization criteria may render one approach more stable and reproducible than the other when working with brain and behavioural data at the population-level. This study compared the LVs which emerged from CCA and PLS analyses of brain-behaviour relationships from the Adolescent Brain Cognitive Development (ABCD) dataset and examined their stability and reproducibility.

Methods

Structural T1-weighted imaging and behavioural data were accessed from the baseline Adolescent Brain Cognitive Development dataset (N > 9000, ages = 9-11 years). The brain matrix consisted of cortical thickness estimates in different cortical regions. The behavioural matrix consisted of 11 subscale scores from the parent-reported Child Behavioral Checklist (CBCL) or 7 cognitive performance measures from the NIH Toolbox. CCA and PLS models were separately applied to the brain-CBCL analysis and brain-cognition analysis. A permutation test was used to assess whether identified LVs were statistically significant. A series of resampling statistical methods were used to assess stability and reproducibility of the LVs.

Results

When examining the relationship between cortical thickness and CBCL scores, the first LV was found to be significant across both CCA and PLS models (singular value: CCA = .13, PLS = .39, p < .001). LV1 from the CCA model found that covariation of CBCL scores was linked to covariation of cortical thickness. LV1 from the PLS model identified decreased cortical thickness linked to lower CBCL scores. There was limited evidence of stability or reproducibility of LV1 for both CCA and PLS. When examining the relationship between cortical thickness and cognitive performance, there were 6 significant LVs for both CCA and PLS (p < .01). The first LV showed similar relationships between CCA and PLS and was found to be stable and reproducible (singular value: CCA = .21, PLS = .43, p < .001).

Conclusion

CCA and PLS identify different brain-behaviour relationships with limited stability and reproducibility when examining the relationship between cortical thickness and parent-reported behavioural measures. However, both methods identified relatively similar brain-behaviour relationships that were stable and reproducible when examining the relationship between cortical thickness and cognitive performance. The results of the current study suggest that stability and reproducibility of brain-behaviour relationships identified by CCA and PLS are influenced by characteristics of the analyzed sample and the included behavioural measurements when applied to a large pediatric dataset.

A systematic review of the neural correlates of well-being reveals no consistent associations

Findings from behavioral and genetic studies indicate a potential role for the involvement of brain structures and brain functioning in well-being. We performed a systematic review on the association between brain structures or brain functioning and well-being, including 56 studies. The 11 electroencephalography (EEG) studies suggest a larger alpha asymmetry (more left than right brain activation) to be related to higher well-being. The 18 Magnetic Resonance Imaging (MRI) studies, 26 resting-state functional MRI studies and two functional near-infrared spectroscopy (fNIRS) studies identified a wide range of brain regions involved in well-being, but replication across studies was scarce, both in direction and strength of the associations. The inconsistency could result from small sample sizes of most studies and a possible wide-spread network of brain regions with small effects involved in well-being. Future directions include well-powered brain-wide association studies and innovative methods to more reliably measure brain activity in daily life.

Deriving psychiatric symptom-based biomarkers from multivariate relationships between psychophysiological and biochemical measures

Identification of biomarkers for psychiatric disorders remains very challenging due to substantial symptom heterogeneity and diagnostic comorbidity, limiting the ability to map symptoms to underlying neurobiology. Dimensional symptom clusters, such as anhedonia, hyperarousal, etc., are complex and arise due to interactions of a multitude of complex biological relationships. The primary aim of the current investigation was to use multi-set canonical correlation analysis (mCCA) to derive biomarkers (biochemical, physiological) linked to dimensional symptoms across the anxiety and depressive spectrum. Active-duty service members (N = 2,592) completed standardized depression, anxiety and posttraumatic stress questionnaires and several psychophysiological and biochemical assays. Using this approach, we identified two phenotype associations between distinct physiological and biological phenotypes. One was characterized by symptoms of dysphoric arousal (anhedonia, anxiety, hypervigilance) which was associated with low blood pressure and startle reactivity. This finding is in line with previous studies suggesting blunted physiological reactivity is associated with subpopulations endorsing anxiety with comorbid depressive features. A second phenotype of anxious fatigue (high anxiety and reexperiencing/avoidance symptoms coupled with fatigue) was associated with elevated blood levels of norepinephrine and the inflammatory marker C-reactive protein in conjunction with high blood pressure. This second phenotype may describe populations in which inflammation and high sympathetic outflow might contribute to anxious fatigue. Overall, these findings support the growing consensus that distinct neuropsychiatric symptom patterns are associated with differential physiological and blood-based biological profiles and highlight the potential of mCCA to reveal important psychiatric symptom biomarkers from several psychophysiological and biochemical measures.

Canonical Correlation Analysis and Partial Least Squares for Identifying Brain-Behavior Associations: A Tutorial and a Comparative Study

Canonical correlation analysis (CCA) and partial least squares (PLS) are powerful multivariate methods for capturing associations across 2 modalities of data (e.g., brain and behavior). However, when the sample size is similar to or smaller than the number of variables in the data, standard CCA and PLS models may overfit, i.e., find spurious associations that generalize poorly to new data. Dimensionality reduction and regularized extensions of CCA and PLS have been proposed to address this problem, yet most studies using these approaches have some limitations. This work gives a theoretical and practical introduction into the most common CCA/PLS models and their regularized variants. We examine the limitations of standard CCA and PLS when the sample size is similar to or smaller than the number of variables. We discuss how dimensionality reduction and regularization techniques address this problem and explain their main advantages and disadvantages. We highlight crucial aspects of the CCA/PLS analysis framework, including optimizing the hyperparameters of the model and testing the identified associations for statistical significance. We apply the described CCA/PLS models to simulated data and real data from the Human Connectome Project and Alzheimer's Disease Neuroimaging Initiative (both of n > 500). We use both low- and high-dimensionality versions of these data (i.e., ratios between sample size and variables in the range of ∼1-10 and ∼0.1-0.01, respectively) to demonstrate the impact of data dimensionality on the models. Finally, we summarize the key lessons of the tutorial.

The promise of a model-based psychiatry: building computational models of mental ill health

Computational models have great potential to revolutionise psychiatry research and clinical practice. These models are now used across multiple subfields, including computational psychiatry and precision psychiatry. Their goals vary from understanding mechanisms underlying disorders to deriving reliable classification and personalised predictions. Rapid growth of new tools and data sources (eg, digital data, gamification, and social media) requires an understanding of the constraints and advantages of different modelling approaches in psychiatry. In this Series paper, we take a critical look at the range of computational models that are used in psychiatry and evaluate their advantages and disadvantages for different purposes and data sources. We describe mechanism-driven and mechanism-agnostic computational models and discuss how interpretability of models is crucial for clinical translation. Based on these evaluations, we provide recommendations on how to build computational models that are clinically useful.

Regional gene expression signatures are associated with sex-specific functional connectivity changes in depression

The neural substrates of depression may differ in men and women, but the underlying mechanisms are incompletely understood. Here, we show that depression is associated with sex-specific patterns of abnormal functional connectivity in the default mode network and in five regions of interest with sexually dimorphic transcriptional effects. Regional differences in gene expression in two independent datasets explained the neuroanatomical distribution of abnormal connectivity. These gene sets varied by sex and were strongly enriched for genes implicated in depression, synapse function, immune signaling, and neurodevelopment. In an independent sample, we confirmed the prediction that individual differences in default mode network connectivity are explained by inferred brain expression levels for six depression-related genes, including PCDH8, a brain-specific protocadherin integral membrane protein implicated in activity-related synaptic reorganization. Together, our results delineate both shared and sex-specific changes in the organization of depression-related functional networks, with implications for biomarker development and fMRI-guided therapeutic neuromodulation.

Migraine disability, pain catastrophizing, and headache severity are associated with evoked pain and targeted by mind-body therapy

ABSTRACT

Meta-analysis suggests that migraine patients are no more sensitive to experimentally evoked pain than healthy control subjects. At the same time, studies have linked some migraine symptoms to quantitative sensory testing (QST) profiles. Unfortunately, previous studies associating migraine symptoms and QST have important methodological shortcomings, stemming from small sample sizes, and frequent use of univariate statistics for multivariate research questions. In the current study, we seek to address these limitations by using a large sample of episodic migraine patients (n = 103) and a multivariate analysis that associates pain ratings from many thermal intensities simultaneously with 12 clinical measures ranging from headache frequency to sleep abnormalities. We identified a single dimension of association between thermal QST and migraine symptoms that relates to pain ratings for all stimulus intensities and a subset of migraine symptoms relating to disability (Headache Impact Test 6 and Brief Pain Inventory interference), catastrophizing (Pain Catastrophizing Scale), and pain severity (average headache pain, Brief Pain Inventory severity, and Short-Form McGill Pain Questionnaire 2). Headache frequency, allodynia, affect, and sleep disturbances were unrelated to this dimension. Consistent with previous research, we did not observe any difference in QST ratings between migraine patients and healthy control subjects. Additionally, we found that the linear combination of symptoms related to QST was modified by the mind-body therapy enhanced mindfulness-based stress reduction (MBSR+). These results suggest that QST has a selective relationship with pain symptoms even in the absence of between-subjects differences between chronic pain patients and healthy control subjects.

Brain-phenotype models fail for individuals who defy sample stereotypes

Individual differences in brain functional organization track a range of traits, symptoms and behaviours1-12. So far, work modelling linear brain-phenotype relationships has assumed that a single such relationship generalizes across all individuals, but models do not work equally well in all participants13,14. A better understanding of in whom models fail and why is crucial to revealing robust, useful and unbiased brain-phenotype relationships. To this end, here we related brain activity to phenotype using predictive models-trained and tested on independent data to ensure generalizability15-and examined model failure. We applied this data-driven approach to a range of neurocognitive measures in a new, clinically and demographically heterogeneous dataset, with the results replicated in two independent, publicly available datasets16,17. Across all three datasets, we find that models reflect not unitary cognitive constructs, but rather neurocognitive scores intertwined with sociodemographic and clinical covariates; that is, models reflect stereotypical profiles, and fail when applied to individuals who defy them. Model failure is reliable, phenotype specific and generalizable across datasets. Together, these results highlight the pitfalls of a one-size-fits-all modelling approach and the effect of biased phenotypic measures18-20 on the interpretation and utility of resulting brain-phenotype models. We present a framework to address these issues so that such models may reveal the neural circuits that underlie specific phenotypes and ultimately identify individualized neural targets for clinical intervention.

Three Dimensions of Association Link Migraine Symptoms and Functional Connectivity

Migraine is a heterogeneous disorder with variable symptoms and responsiveness to therapy. Because of previous analytic shortcomings, variance in migraine symptoms has been inconsistently related to brain function. In the current analysis, we used data from two sites (n = 143, male and female humans), and performed canonical correlation analysis, relating resting-state functional connectivity (RSFC) with a broad range of migraine symptoms, ranging from headache characteristics to sleep abnormalities. This identified three dimensions of covariance between symptoms and RSFC. The first dimension related to headache intensity, headache frequency, pain catastrophizing, affect, sleep disturbances, and somatic abnormalities, and was associated with frontoparietal and dorsal attention network connectivity, both of which are major cognitive networks. Additionally, RSFC scores from this dimension, both the baseline value and the change from baseline to postintervention, were associated with responsiveness to mind-body therapy. The second dimension was related to an inverse association between pain and anxiety, and to default mode network connectivity. The final dimension was related to pain catastrophizing, and salience, sensorimotor, and default mode network connectivity. In addition to performing canonical correlation analysis, we evaluated the current clustering of migraine patients into episodic and chronic subtypes, and found no evidence to support this clustering. However, when using RSFC scores from the three significant dimensions, we identified a novel clustering of migraine patients into four biotypes with unique functional connectivity patterns. These findings provide new insight into individual variability in migraine, and could serve as the foundation for novel therapies that take advantage of migraine heterogeneity.SIGNIFICANCE STATEMENT Using a large multisite dataset of migraine patients, we identified three dimensions of multivariate association between symptoms and functional connectivity. This analysis revealed neural networks that relate to all measured symptoms, but also to specific symptom ensembles, such as patient propensity to catastrophize painful events. Using these three dimensions, we found four biotypes of migraine informed by clinical and neural variation together. Such findings pave the way for precision medicine therapy for migraine.

Linking interindividual variability in brain structure to behaviour

What are the brain structural correlates of interindividual differences in behaviour? More than a decade ago, advances in structural MRI opened promising new avenues to address this question. The initial wave of research then progressively led to substantial conceptual and methodological shifts, and a replication crisis unveiled the limitations of traditional approaches, which involved searching for associations between local measurements of neuroanatomy and behavioural variables in small samples of healthy individuals. Given these methodological issues and growing scepticism regarding the idea of one-to-one mapping of psychological constructs to brain regions, new perspectives emerged. These not only embrace the multivariate nature of brain structure-behaviour relationships and promote generalizability but also embrace the representation of the relationships between brain structure and behavioural data by latent dimensions of interindividual variability. Here, we examine the past and present of the study of brain structure-behaviour associations in healthy populations and address current challenges and open questions for future investigations.

A hierarchical Bayesian model to find brain-behaviour associations in incomplete data sets

Canonical Correlation Analysis (CCA) and its regularised versions have been widely used in the neuroimaging community to uncover multivariate associations between two data modalities (e.g., brain imaging and behaviour). However, these methods have inherent limitations: (1) statistical inferences about the associations are often not robust; (2) the associations within each data modality are not modelled; (3) missing values need to be imputed or removed. Group Factor Analysis (GFA) is a hierarchical model that addresses the first two limitations by providing Bayesian inference and modelling modality-specific associations. Here, we propose an extension of GFA that handles missing data, and highlight that GFA can be used as a predictive model. We applied GFA to synthetic and real data consisting of brain connectivity and non-imaging measures from the Human Connectome Project (HCP). In synthetic data, GFA uncovered the underlying shared and specific factors and predicted correctly the non-observed data modalities in complete and incomplete data sets. In the HCP data, we identified four relevant shared factors, capturing associations between mood, alcohol and drug use, cognition, demographics and psychopathological measures and the default mode, frontoparietal control, dorsal and ventral networks and insula, as well as two factors describing associations within brain connectivity. In addition, GFA predicted a set of non-imaging measures from brain connectivity. These findings were consistent in complete and incomplete data sets, and replicated previous findings in the literature. GFA is a promising tool that can be used to uncover associations between and within multiple data modalities in benchmark datasets (such as, HCP), and easily extended to more complex models to solve more challenging tasks.

Individual Differences in the Affective Response to Pandemic-related Stressors in COVID-19 Healthcare Workers

Background

We investigated the evolving prevalence of mood and anxiety symptoms among healthcare workers from May, 2020 to January, 2021; risk factors for adverse outcomes; and characteristic modes of affective responses to pandemic-related stressors.

Methods

2,307 healthcare workers (78.9% female, modal age 25-34) participated in an online survey assessing depression (Patient Health Questionnaire [PHQ-9]) and anxiety symptoms (Generalized Anxiety Disorder scale [GAD-7]), demographic variables, and self-reported impact of pandemic-related stressors. 334 subjects were reassessed ∼6 months later.

Results

The prevalence of clinically significant depression and anxiety was 45.3% and 43.3%, respectively, and a majority (59.9%-62.9%) of those individuals had persistent significant symptoms at 6-month follow-up. Younger age, female gender, and specific occupations (support staff > nurses > physicians) were associated with increased depressive and anxiety symptoms. The most important risk factors were social isolation and fear of contracting COVID-19. The prevalence of clinically significant mood and anxiety symptoms increased by 39.8% from May, 2020 to January, 2021. PHQ-9 and GAD-7 scores were highly correlated and associated with nearly identical risk factors, suggesting that they are not capturing independent constructs in this sample. Principal components analysis identified seven orthogonal symptom domains with unique risk factors.

Conclusions

Clinically significant mood and anxiety symptoms are highly prevalent and persistent among healthcare workers, and are associated with numerous risk factors, the strongest of which are related to pandemic stressors and potentially modifiable. Interventions aimed at reducing social isolation and mitigating the impact of fear of infection warrant further study.

Dissecting diagnostic heterogeneity in depression by integrating neuroimaging and genetics

Depression is a heterogeneous and etiologically complex psychiatric syndrome, not a unitary disease entity, encompassing a broad spectrum of psychopathology arising from distinct pathophysiological mechanisms. Motivated by a need to advance our understanding of these mechanisms and develop new treatment strategies, there is a renewed interest in investigating the neurobiological basis of heterogeneity in depression and rethinking our approach to diagnosis for research purposes. Large-scale genome-wide association studies have now identified multiple genetic risk variants implicating excitatory neurotransmission and synapse function and underscoring a highly polygenic inheritance pattern that may be another important contributor to heterogeneity in depression. Here, we review various sources of phenotypic heterogeneity and approaches to defining and studying depression subtypes, including symptom-based subtypes and biology-based approaches to decomposing the depression syndrome. We review "dimensional," "categorical," and "hybrid" approaches to parsing phenotypic heterogeneity in depression and defining subtypes using functional neuroimaging. Next, we review recent progress in neuroimaging genetics (correlating neuroimaging patterns of brain function with genetic data) and its potential utility for generating testable hypotheses concerning molecular and circuit-level mechanisms. We discuss how genetic variants and transcriptomic profiles may confer risk for depression by modulating brain structure and function. We conclude by highlighting several promising areas for future research into the neurobiological underpinnings of heterogeneity, including efforts to understand sexually dimorphic mechanisms, the longitudinal dynamics of depressive episodes, and strategies for developing personalized treatments and facilitating clinical decision-making.

A technical review of canonical correlation analysis for neuroscience applications

Collecting comprehensive data sets of the same subject has become a standard in neuroscience research and uncovering multivariate relationships among collected data sets have gained significant attentions in recent years. Canonical correlation analysis (CCA) is one of the powerful multivariate tools to jointly investigate relationships among multiple data sets, which can uncover disease or environmental effects in various modalities simultaneously and characterize changes during development, aging, and disease progressions comprehensively. In the past 10 years, despite an increasing number of studies have utilized CCA in multivariate analysis, simple conventional CCA dominates these applications. Multiple CCA-variant techniques have been proposed to improve the model performance; however, the complicated multivariate formulations and not well-known capabilities have delayed their wide applications. Therefore, in this study, a comprehensive review of CCA and its variant techniques is provided. Detailed technical formulation with analytical and numerical solutions, current applications in neuroscience research, and advantages and limitations of each CCA-related technique are discussed. Finally, a general guideline in how to select the most appropriate CCA-related technique based on the properties of available data sets and particularly targeted neuroscience questions is provided.

Approaches to Defining Common and Dissociable Neurobiological Deficits Associated With Psychopathology in Youth

Psychiatric disorders show high rates of comorbidity and nonspecificity of presenting clinical symptoms, while demonstrating substantial heterogeneity within diagnostic categories. Notably, many of these psychiatric disorders first manifest in youth. We review progress and next steps in efforts to parse heterogeneity in psychiatric symptoms in youths by identifying abnormalities within neural circuits. To address this fundamental challenge in psychiatry, a number of methods have been proposed. We provide an overview of these methods, broadly organized into dimensional versus categorical approaches and single-view versus multiview approaches. Dimensional approaches including factor analysis and canonical correlation analysis aim to capture dimensional associations between psychopathology and brain measures across a continuous spectrum from health to disease. In contrast, categorical approaches, such as clustering and community detection, aim to identify subtypes of individuals within a class of symptoms or brain features. We highlight several studies that apply these methods to samples of youths and discuss issues to consider when using these approaches. Finally, we end by highlighting avenues for future research.

Causes and Consequences of Diagnostic Heterogeneity in Depression: Paths to Discovering Novel Biological Depression Subtypes

Depression is a highly heterogeneous syndrome that bears only modest correlations with its biological substrates, motivating a renewed interest in rethinking our approach to diagnosing depression for research purposes and new efforts to discover subtypes of depression anchored in biology. Here, we review the major causes of diagnostic heterogeneity in depression, with consideration of both clinical symptoms and behaviors (symptomatology and trajectory of depressive episodes) and biology (genetics and sexually dimorphic factors). Next, we discuss the promise of using data-driven strategies to discover novel subtypes of depression based on functional neuroimaging measures, including dimensional, categorical, and hybrid approaches to parsing diagnostic heterogeneity and understanding its biological basis. The merits of using resting-state functional magnetic resonance imaging functional connectivity techniques for subtyping are considered along with a set of technical challenges and potential solutions. We conclude by identifying promising future directions for defining neurobiologically informed depression subtypes and leveraging them in the future for predicting treatment outcomes and informing clinical decision making.

Permutation inference for canonical correlation analysis

Canonical correlation analysis (CCA) has become a key tool for population neuroimaging, allowing investigation of associations between many imaging and non-imaging measurements. As age, sex and other variables are often a source of variability not of direct interest, previous work has used CCA on residuals from a model that removes these effects, then proceeded directly to permutation inference. We show that a simple permutation test, as typically used to identify significant modes of shared variation on such data adjusted for nuisance variables, produces inflated error rates. The reason is that residualisation introduces dependencies among the observations that violate the exchangeability assumption. Even in the absence of nuisance variables, however, a simple permutation test for CCA also leads to excess error rates for all canonical correlations other than the first. The reason is that a simple permutation scheme does not ignore the variability already explained by previous canonical variables. Here we propose solutions for both problems: in the case of nuisance variables, we show that transforming the residuals to a lower dimensional basis where exchangeability holds results in a valid permutation test; for more general cases, with or without nuisance variables, we propose estimating the canonical correlations in a stepwise manner, removing at each iteration the variance already explained, while dealing with different number of variables in both sides. We also discuss how to address the multiplicity of tests, proposing an admissible test that is not conservative, and provide a complete algorithm for permutation inference for CCA.

Multiple Holdouts With Stability: Improving the Generalizability of Machine Learning Analyses of Brain-Behavior Relationships

BACKGROUND

In 2009, the National Institute of Mental Health launched the Research Domain Criteria, an attempt to move beyond diagnostic categories and ground psychiatry within neurobiological constructs that combine different levels of measures (e.g., brain imaging and behavior). Statistical methods that can integrate such multimodal data, however, are often vulnerable to overfitting, poor generalization, and difficulties in interpreting the results.

METHODS

We propose an innovative machine learning framework combining multiple holdouts and a stability criterion with regularized multivariate techniques, such as sparse partial least squares and kernel canonical correlation analysis, for identifying hidden dimensions of cross-modality relationships. To illustrate the approach, we investigated structural brain-behavior associations in an extensively phenotyped developmental sample of 345 participants (312 healthy and 33 with clinical depression). The brain data consisted of whole-brain voxel-based gray matter volumes, and the behavioral data included item-level self-report questionnaires and IQ and demographic measures.

RESULTS

Both sparse partial least squares and kernel canonical correlation analysis captured two hidden dimensions of brain-behavior relationships: one related to age and drinking and the other one related to depression. The applied machine learning framework indicates that these results are stable and generalize well to new data. Indeed, the identified brain-behavior associations are in agreement with previous findings in the literature concerning age, alcohol use, and depression-related changes in brain volume.

CONCLUSIONS

Multivariate techniques (such as sparse partial least squares and kernel canonical correlation analysis) embedded in our novel framework are promising tools to link behavior and/or symptoms to neurobiology and thus have great potential to contribute to a biologically grounded definition of psychiatric disorders.