Abstract Lesion network mapping (LNM) describes a group of methods using normative functional connectivity data to map disparate brain lesions and stimulation sites onto common brain networks. Van den Heuvel and colleagues recently showed that these methods lack disease specificity, instead producing maps that converge toward intrinsic properties of the normative connectome dataset. Here, we investigate symptom LNM (sLNM), a recent advancement in the method which attempts to increase the robustness of results by incorporating symptom severity and incorporating replication across multiple datasets and prediction of clinical outcomes. Using clinical datasets of depression and Broca’s aphasia, we show that sLNM maps from unrelated disorders nonetheless converge despite using null models which break the specific lesion–symptom structure in the datasets. Using simulated datasets with a known ground-truth disease network, we show that sLNM results are systematically biased towards the normative connectome’s first principal component (PC1), which drives spurious convergence across unrelated datasets. We further show that the apparent clinical predictive capability of these maps are non-specific: network maps derived from unrelated disorders such as migraine and aphasia predict brain stimulation improvement in depression as well as — or better than — the cohort’s own sLNM map. However, controlling for PC1 reduces spurious convergence across unrelated datasets and improves clinical prediction specificity, supporting the notion that disease-specific signal exists within sLNM but is confounded by the globally present PC1 signal in the normative connectome. These findings offer a practical correction applicable to existing and future sLNM studies.

As the field expands from early research into the human connectome, there has been a fast expansion in the number of analytical approaches to study resting state functional MRI (rsfMRI) data. With increasing focus on individual differences, topographical brain maps of spatial organization have emerged in addition to traditional functional connectomes. Here, we developed a deep-learning model to embed maps of network topography and faithfully translate to individualized connectomes. Results confirmed the validity of the surface vision transformer based on reconstruction accuracy (0.73±0.09) and accurate topography-to-connectome translation (0.43±0.08). Importantly, translated connectomes retained identifiability and brain-cognition associations. These findings establish a direct mapping from spatial topography to connectomes that can be used to integrate scientific insights across rsfMRI sub-fields. This is an important step towards broadening our conceptualization of the connectome and supporting a broader integration of findings to inform a complete understanding of the human connectome.

This study aims to investigate functional and neurotransmitter signaling in the prefrontal-hippocampal pathway in relation to depression in a cohort of Thai transgender women. Twenty participants completed mental health surveys and imaging between January and March 2024. Depression severity was measured by Patient Health Questionnaire-9 (PHQ-9) scores. Higher PHQ-9 scores were associated with lower GABA levels in the hippocampus, and with lower fractional amplitude of low-frequency fluctuations (fALFF) in the dorsolateral prefrontal cortex. However, removal of the hippocampal GABA outlier resulted in a non-significant relationship with PHQ-9. Therefore, future studies with larger datasets should further investigate the association between GABA and depression in a transgender cohort. These findings revealed interactions between neurotransmitter signaling and functional brain activity of the hippocampal-prefrontal circuit in depression.

Objective: Though caffeine use during pregnancy is common, its longitudinal associations with child behavioral and physical health outcomes remain poorly understood. Here, we estimated associations between prenatal caffeine exposure, body mass index (BMI), and behavior as children enter adolescence. Method: Longitudinal data and caregiver-reported prenatal caffeine exposure were obtained from the ongoing Adolescent Brain and Cognitive Development (ABCD)(SM) Study, which recruited 11,875 children aged 9-11 years at baseline from 21 sites across the United States starting June 1, 2016. Prenatal caffeine exposure was analyzed as a 4-level categorical variable, and further group contrasts were used to characterize "any exposure" and "daily exposure" groups. Outcomes included psychopathology characteristics in children, sleep problems, and BMI. Potentially confounding covariates included familial (e.g., income, familial psychopathology), pregnancy (e.g., prenatal substance exposure), and child (e.g., caffeine use) variables. Results: Among 10,873 children (5686 boys [52.3 %]; mean [SD] age, 9.9 [0.6] years) with nonmissing prenatal caffeine exposure data, 6560 (60 %) were exposed to caffeine prenatally. Relative to no exposure, daily caffeine exposure was associated with higher child BMI (beta = 0.08; FDR-corrected p = 0.02), but was not associated with child behavior following correction for multiple testing. Those exposed to two or more cups of caffeine daily (n = 1028) had greater sleep problems than those with lower/no exposure (beta > 0.92; FDR-corrected p < 0.04). Conclusion: Daily prenatal caffeine exposure is associated with heightened childhood BMI, and when used multiple times a day greater sleep problems even after accounting for potential confounds. Whether this relationship is a consequence of prenatal caffeine exposure or its correlated factors remains unknown.

Information processing in the brain spans from localised sensorimotor processes to higher-level cognition that integrates across multiple regions. Interactions between and within these subsystems enable multiscale information processing. Despite this multiscale characteristic, functional brain connectivity is often either estimated based on 10-30 distributed modes or parcellations with 100-1000 localised parcels, both missing across-scale functional interactions. We present Multiscale Probabilistic Functional Modes (mPFMs), a new mapping which comprises modes over various scales of granularity, thus enabling direct estimation of functional connectivity within- and across-scales. Crucially, mPFMs emerged from data-driven multilevel Bayesian modelling of large functional MRI (fMRI) populations. We demonstrate that mPFMs capture both distributed brain modes and their co-existing subcomponents. In addition to validating mPFMs using simulations and real data, we show that mPFMs can predict ~900 personalised traits from UK Biobank more accurately than current standard techniques. Therefore, mPFMs can offer a paradigm shift in functional connectivity modelling and yield enhanced fMRI biomarkers for traits and diseases.

BACKGROUND: Patients with depression vary from one another in their clinical and neuroimaging presentation, but the relationship between clinical and neuroimaging sources of variation is poorly understood. Determining sources of heterogeneity in depression is important to gain insights into its diverse and complex neural etiology. In this study, we aimed to test whether depression heterogeneity is characterized by subgroups that differ both clinically and neurobiologically and/or whether multiple neuroimaging profiles give rise to the same clinical presentation. METHODS: This study utilized population-based data from the UK Biobank over multiple imaging sites. Clinically dissociated groups were selected to isolate clinical characteristics of depression (symptoms of anhedonia, depressed mood, and somatic disturbance; severity indices of lifetime chronicity and acute impairment; and late onset). Residual neuroimaging heterogeneity within each group was assessed using neuroimaging-driven clustering. RESULTS: The clinically dissociated subgroups had significantly larger neuroimaging normative deviations than a comparison heterogeneous group and had distinct neuroimaging profiles from each other. Imaging-driven clustering within each clinically dissociated group identified 2 stable subtypes within the acute impairment group that differed significantly in cognitive ability despite identical clinical profiles. CONCLUSIONS: The study identified distinct neuroimaging profiles related to particular clinical depression features that may explain inconsistencies in the literature and subclusters within the acute impairment group with cognitive differences that were only differentiable by neuroimaging. Our results provide evidence that multiple neuroimaging profiles may give rise to the same clinical presentation, emphasizing the presence of complex interactions between clinical and neuroimaging sources of heterogeneity.

A key challenge in predicting a person’s internal state of mind is that there are a wide range of contributing factors that each have a subtle, yet significant, influence on mental health. To address this challenge, we trained machine learning algorithms on multiple sources of variation that could meaningfully contribute to psychological distress. Data mining techniques were used to identify key risk factors for predicting current symptoms and longitudinal outcomes from the Adolescent Brain Cognitive Developmental dataset (n = 11,552). Our results consistently revealed that social conflicts were the strongest indicators of psychopathology. Family fighting was the best predictor of current symptoms, whereas social victimization among peers was the top predictor of changes across time and future symptom severity. Sex-differences also emerged as a critical factor for predicting psychopathology, as females exhibited greater symptoms on average, and their symptoms became more severe over time. The long-term mental health of females was best predicted by social exclusion and reputational damage, whereas aggression and anti-social tendencies were the strongest predictors among males. While these findings provide novel insight into the developmental origins of psychopathology, our best performing models could only explain up to 40% of the variation between individuals. Future research is needed to obtain a more complete understanding of all the factors that meaningfully contribute to mental health.

Older adults often show improved emotional regulation with age, a phenomenon known as the aging paradox. This age-related increase in emotional regulation capacity is attributed to enhanced prefrontal cortex control over amygdala reactivity. However, because racial discrimination and economic disadvantage cause chronic stress, typical age-related neural associations may be altered in marginalized groups. Using task-functional MRI data from 8,711 UK Biobank participants aged 50-78, we investigated whether age-related associations in emotion-related brain function, specifically amygdala activation and vmPFC-amygdala connectivity, varied across racial and socioeconomic status (SES) groups. We found that older age was associated with decreased amygdala activation, which is consistent with improved emotional regulation. Yet, lower socioeconomic status was associated with increased amygdala activation, suggesting heightened stress-related reactivity. No significant age-related effects on vmPFC-amygdala connectivity were observed at the population level. Black participants showed a stronger age-related decline in functional connectivity compared to other racial groups. These findings call for more inclusive and diverse neuroimaging studies to better understand brain health across marginalized groups.

Despite depression being a leading cause of global disability, neuroimaging studies have struggled to identify replicable neural correlates of depression or explain limited variance. This challenge may, in part, stem from the intertwined state (current symptoms; variable) and trait (general propensity; stable) experiences of depression. Here, we sought to disentangle state from trait experiences of depression by leveraging a longitudinal cohort and stratifying individuals into four groups: those in remission (‘trait depression group’), those with large longitudinal severity changes in depression symptomatology (‘state depression group’), and their respective matched control groups (total analytic n=1,030). We hypothesized that spatial network organization would be linked to trait depression due to its temporal stability, whereas functional connectivity between networks would be more sensitive to state-dependent depression symptoms due to its capacity to fluctuate. We identified 15 large-scale probabilistic functional networks from resting-state fMRI data and performed group comparisons on the amplitude, connectivity, and spatial overlap between these networks, using matched control participants as reference. Our findings revealed higher amplitude in visual networks for the trait depression group at the time of remission, in contrast to controls. This observation may suggest altered visual processing in individuals predisposed to developing depression over time. No significant group differences were observed in any other network measures for the trait-control comparison, nor in any measures for the state-control comparison. These results underscore the overlooked contribution of visual networks to the psychopathology of depression and provide evidence for distinct neural correlates between state and trait experiences of depression.

Background The use of machine learning to classify diagnostic cases versus controls defined based on diagnostic ontologies such as the International Classification of Diseases, Tenth Revision (ICD-10) from neuroimaging features is now commonplace across a wide range of diagnostic fields. However, transdiagnostic comparisons of such classifications are lacking. Such transdiagnostic comparisons are important to establish the specificity of classification models, set benchmarks, and assess the value of diagnostic ontologies.Results We investigated case-control classification accuracy in 17 different ICD-10 diagnostic groups from Chapter V (mental and behavioral disorders) and Chapter VI (diseases of the nervous system) using data from the UK Biobank. Classification models were trained using either neuroimaging (structural or functional brain magnetic resonance imaging feature sets) or sociodemographic features. Random forest classification models were adopted using rigorous shuffle-splits to estimate stability as well as accuracy of case-control classifications. Diagnostic classification accuracies were benchmarked against age classification (oldest vs. youngest) from the same feature sets and against additional classifier types (k-nearest neighbors and linear support vector machine). In contrast to age classification accuracy, which was high for all feature sets, few ICD-10 diagnostic groups were classified significantly above chance (namely, demyelinating diseases based on structural neuroimaging features and depression based on sociodemographic and functional neuroimaging features).Conclusion These findings highlight challenges with the current disease classification system, leading us to recommend caution with the use of ICD-10 diagnostic groups as target labels in brain-based disease prediction studies.

Electrophysiological recordings during ketamine anesthesia have revealed a slow alternating pattern of high- and low- frequency activity (a "gamma-burst" pattern) that develops with the onset of general anesthesia. We examine the role of NMDA receptor antagonism in generating the gamma-burst pattern and the link between gamma-bursts and dissociative anesthesia. We compare the effects of ketamine with those of the highly selective NMDA receptor antagonist CGS 19755 on multi-site intracranial electrophysiology and behavior in rhesus macaques. Remarkably, we find that animals given a moderate dose of CGS 19755 are able to perform a difficult memory task, while at the same time showing electrophysiological activity similar to ketamine anesthesia, with one key difference: a lack of delta-band LFP modulation. This difference demonstrates that ketamine's ability to drive strong delta-band oscillations relies on additional mechanisms beyond NMDA receptor antagonism alone, and points to a key role for the activity underlying delta-band oscillations in causing anesthesia.

Neuroimaging research has uncovered a multitude of neural abnormalities associated with psychopathology, but few prediction-based studies have been conducted during adolescence, and even fewer used neurobiological features that were extracted across multiple neuroimaging modalities. This gap in the literature is critical, as deriving accurate brain-based models of psychopathology is an essential step towards understanding key neural mechanisms and identifying high-risk individuals. As such, we trained adaptive tree-boosting algorithms on multimodal neuroimaging features from the Lifespan Human Connectome Developmental (HCP-D) sample that contained 956 participants between the ages of 8 to 22 years old. Our feature space consisted of 1037 anatomical, 1090 functional, and 192 diffusion MRI features, which were used to derive models that separately predicted internalizing symptoms, externalizing symptoms, and the general psychopathology factor. We found that multimodal models were the most accurate, but all brain-based models of psychopathology yielded out-of-sample predictions that were weakly correlated with actual symptoms (r2 < 0.15). White matter microstructural properties, including orientation dispersion indices and intracellular volume fractions, were the most predictive of general psychopathology, followed by cortical thickness and functional connectivity. Spatially, the most predictive features of general psychopathology were primarily localized within the default mode and dorsal attention networks. These results were mostly consistent across all dimensions of psychopathology, except orientation dispersion indices and the default mode network were not as heavily weighted in the prediction of internalizing and externalizing symptoms. Taken with prior literature, it appears that neurobiological features are an important part of the equation for predicting psychopathology but relying exclusively on neural markers is clearly not sufficient, especially among adolescent samples with subclinical symptoms. Consequently, risk factor models of psychopathology may benefit from incorporating additional sources of information that have also been shown to explain individual differences, such as psychosocial factors, environmental stressors, and genetic vulnerabilities.

Background:Both cognitive and non-cognitive (e.g., traits like curiosity) factors are critical for social and emotional functioning and independently predict educational attainment. These factors are heritable and genetically correlated with a range of health-relevant traits and behaviors in adulthood (e.g., risk-taking, psychopathology). However, whether these associations are present during adolescence, and to what extent these relationships diverge, could have implications for adolescent health and well-being. Methods:Using data from 5,517 youth of European ancestry from the ongoing Adolescent Brain Cognitive DevelopmentSM Study, we examined associations between polygenic scores (PGS) for cognitive and non-cognitive factors and outcomes related to cognition, socioeconomic status, risk tolerance and decision-making, substance initiation, psychopathology, and brain structure. Results:Cognitive and non-cognitive PGSs were both positively associated with cognitive performance and family income, and negatively associated with ADHD and severity of psychotic-like experiences. The cognitive PGS was also associated with greater risk-taking, delayed discounting, and anorexia, as well as lower likelihood of nicotine initiation. The cognitive PGS was further associated with cognition scores and anorexia in within-sibling analyses, suggesting these results do not solely reflect the effects of assortative mating or passive gene-environment correlations. The cognitive PGS showed significantly stronger associations with cortical volumes than the non-cognitive PGS and was associated with right hemisphere caudal anterior cingulate and pars-orbitalis in within-sibling analyses, while the non-cognitive PGS showed stronger associations with white matter fractional anisotropy and a significant within-sibling association for right superior corticostriate-frontal cortex. Conclusions:Our findings suggest that PGSs for cognitive and non-cognitive factors show similar associations with cognition and socioeconomic status as well as other psychosocial outcomes, but distinct associations with regional neural phenotypes in this adolescent sample.

Research into the brain basis of psychopathology is challenging due to the heterogeneity of psychiatric disorders, extensive comorbidities, underdiagnosis or overdiagnosis, multifaceted interactions with genetics and life experiences, and the highly multivariate nature of neural correlates. Therefore, increasingly larger datasets that measure more variables in larger cohorts are needed to gain insights. In this review, we present current “best practice” approaches for using existing databases, collecting and sharing new repositories for big data analyses, and future directions for big data in neuroimaging and psychiatry with an emphasis on contributing to collaborative efforts and the challenges of multi-study data analysis.

Research into neuroimaging biomarkers for Late Life Depression (LLD) has identified neural correlates of LLD including increased white matter hyperintensities and reduced hippocampal volume. However, studies into neuroimaging biomarkers for LLD largely fail to converge. This lack of replicability is potentially due to challenges linked to construct variability, etiological heterogeneity, and experimental rigor. We discuss suggestions to help address these challenges, including improved construct standardization, increased sample sizes, multimodal approaches to parse heterogeneity, and the use of individualized analytical models.

Alcohol use and smoking are common substance-use behaviors with well-established negative health effects, including decreased brain health. We examined whether alcohol use and smoking were associated with the same neuroimaging-derived brain measures. We further explored whether the effects of alcohol use and smoking on the brain were additive or interactive. We leveraged a cohort of 36,309 participants with neuroimaging data from the UK Biobank. We used linear regression to determine the association between 354 neuroimaging-derived brain measures and alcohol use defined as drinks per week, pack years of smoking, and drinks per week × pack years smoking interaction. To assess whether the brain associations with alcohol are broadly similar or different from the associations with smoking, we calculated the correlation between z-scores of association for drinks per week and pack years smoking. Results indicated overall moderate positive correlation in the associations across measures representing brain structure, magnetic susceptibility, and white matter tract microstructure, indicating greater similarity than difference in the brain measures associated with alcohol use and smoking. The only evidence of an interaction between drinks per week and pack years smoking was seen in measures representing magnetic susceptibility in subcortical structures. The effects of alcohol use and smoking on brain health appeared to be additive rather than multiplicative for all other brain measures studied. 97% (224/230) of associations with alcohol and 100% (167/167) of the associations with smoking that surpassed a p value threshold are in a direction that can be interpreted to reflect reduced brain health. Our results underscore the similarity of the adverse associations between use of these substances and neuroimaging derived brain measures.

Background:1.The use of machine learning to classify diagnostic cases versus controls defined based on diagnostic ontologies such as the ICD-10 from neuroimaging features is now commonplace across a wide range of diagnostic fields. However, transdiagnostic comparisons of such classifications are lacking. Such transdiagnostic comparisons are important to establish the specificity of classification models, set benchmarks, and assess the value of diagnostic ontologies. Results:2.We investigated case-control classification accuracy in 17 different ICD-10 diagnostic groups from Chapter V (mental and behavioral disorders) and Chapter VI (diseases of the nervous system) using data from the UK Biobank. Classification models were trained using either neuroimaging (structural or functional brain MRI feature sets) or socio-demographic features. Random forest classification models were adopted using rigorous shuffle splits to estimate stability as well as accuracy of case-control classifications. Diagnostic classification accuracies were benchmarked against age classification (oldest versus youngest) from the same feature sets and against additional classifier types (K-nearest neighbors and linear support vector machine). In contrast to age classification accuracy, which was high for all feature sets, few ICD-10 diagnostic groups were classified significantly above chance (namely, demyelinating diseases based on structural neuroimaging features, and depression based on socio-demographic and functional neuroimaging features). Conclusion:3.These findings highlight challenges with the current disease classification system, leading us to recommend caution with the use of ICD-10 diagnostic groups as target labels in brain-based disease prediction studies.

Prenatal cannabis exposure (PCE) is associated with mental health problems in early adolescence, but the possible neurobiological mechanisms remain unknown. In a large longitudinal sample of adolescents (ages 9-12 years, n = 9,322-10,186), we find that PCE is associated with localized differences in gray and white matter of the frontal and parietal cortices, their associated white matter tracts, and striatal resting-state connectivity, even after accounting for potential pregnancy, familial, and child confounds. Variability in forceps minor and pars triangularis diffusion metrics partially longitudinally mediate associations of PCE with attention problems and attention deficit hyperactivity disorder symptoms. PCE-related differences in brain development may confer vulnerability to worse mental health in early adolescence.

INTRODUCTION: The hippocampus atrophies with age and is implicated in neurodegenerative disorders including Alzheimer's disease (AD). We examined the interplay between age and apolipoprotein E (APOE) genotype on total hippocampal volume. METHODS: Using neuroimaging data from 37,463 UK Biobank participants, we applied linear regression to quantify the association of age and APOE with hippocampal volume and identified the age when volumes of epsilon 2/epsilon 3, epsilon 3/epsilon 4, and epsilon 4/epsilon 4 carriers significantly deviated from epsilon 3/epsilon 3 using generalized additive modeling. RESULTS: Total hippocampal volume declined with age, with significant differences by APOE genotype emerging after age 60. epsilon 3/epsilon 4 and epsilon 4/epsilon 4 carriers displayed reduced volumes from ages 69 and 61, respectively, while epsilon 2/epsilon 3 showed delayed decline starting at the age of 76. DISCUSSION: The association of APOE and hippocampal volume is age-dependent, with differences in volumes of epsilon 4/epsilon 4 carriers detected as early as age 61. This work underscores the importance of APOE genotype in determining when to begin screening for AD.