Estimating age based on neuroimaging‐derived data has become a popular approach to developing markers for brain integrity and health. While a variety of machine‐learning algorithms can provide accurate predictions of age based on brain characteristics, there is significant variation in model accuracy reported across studies. We predicted age in two population‐based datasets, and assessed the effects of age range, sample size and age‐bias correction on the model performance metrics Pearson's correlation coefficient (r), the coefficient of determination (R²), Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). The results showed that these metrics vary considerably depending on cohort age range; r and R² values are lower when measured in samples with a narrower age range. RMSE and MAE are also lower in samples with a narrower age range due to smaller errors/brain age delta values when predictions are closer to the mean age of the group. Across subsets with different age ranges, performance metrics improve with increasing sample size. Performance metrics further vary depending on prediction variance as well as mean age difference between training and test sets, and age‐bias corrected metrics indicate high accuracy—also for models showing poor initial performance. In conclusion, performance metrics used for evaluating age prediction models depend on cohort and study‐specific data characteristics, and cannot be directly compared across different studies. Since age‐bias corrected metrics generally indicate high accuracy, even for poorly performing models, inspection of uncorrected model results provides important information about underlying model attributes such as prediction variance.

The deviation between chronological age and age predicted using brain MRI is a putative marker of overall brain health. Age prediction based on structural MRI data shows high accuracy in common brain disorders. However, brain aging is complex and heterogenous, both in terms of individual differences and the underlying biological processes. Here, we implemented a multimodal model to estimate brain age using different combinations of cortical area, thickness and sub‐cortical volumes, cortical and subcortical T1/T2‐weighted ratios, and cerebral blood flow (CBF) based on arterial spin labeling. For each of the 11 models we assessed the age prediction accuracy in healthy controls (HC, n = 750) and compared the obtained brain age gaps (BAGs) between age‐matched subsets of HC and patients with Alzheimer's disease (AD, n = 54), mild (MCI, n = 90) and subjective (SCI, n = 56) cognitive impairment, schizophrenia spectrum (SZ, n = 159) and bipolar disorder (BD, n = 135). We found highest age prediction accuracy in HC when integrating all modalities. Furthermore, two‐group case–control classifications revealed highest accuracy for AD using global T1‐weighted BAG, while MCI, SCI, BD and SZ showed strongest effects in CBF‐based BAGs. Combining multiple MRI modalities improves brain age prediction and reveals distinct deviations in patients with psychiatric and neurological disorders. The multimodal BAG was most accurate in predicting age in HC, while group differences between patients and HC were often larger for BAGs based on single modalities. These findings indicate that multidimensional neuroimaging of patients may provide a brain‐based mapping of overlapping and distinct pathophysiology in common disorders.

Metabolites in plasma form biosignatures of a range of common complex human diseases. Discovering variants with pleiotropic effects across metabolites can reveal underlying biological mechanisms. We therefore performed uni- and multivariate genome-wide association studies (GWAS) on 249 circulating metabolic markers across 328,006 UK Biobank and Estonian Biobank participants. We investigated rare variation through whole exome sequencing gene burden tests, analysed the role of body mass index through Mendelian randomization, and performed genome-wide interaction analyses with sex. We discovered 15,585 loci summed over the univariate GWAS, with high pleiotropy across markers, linked to a wide range of disorders. Findings from common and rare variant gene tests converged on lipid homeostasis pathways. 31 loci interacted with sex, mapped to genes involved in cholesterol processing. The findings offer insights into the genetic architecture of circulating metabolites, revealing pleiotropic loci, highlighting the role of rare variation, and uncovering sex-specific molecular mechanisms of lipid metabolism.

Neuroimaging research has shown brain morphological abnormalities associated with violence and psychosis, but individual differences are substantial and results not consistent across studies. Normative modeling of brain MRI-features facilitates a systematic mapping of individual brain characteristics of complex phenotypes also in small samples but has not yet been applied to forensic psychiatry populations. We explored brain heterogeneity in persons with a history of severe violence with a comorbid schizophrenia spectrum disorder (SSD-V; n = 38), non-violent persons with schizophrenia spectrum disorders (SSD-NV; n = 138), persons with a history of severe violence without comorbid schizophrenia spectrum disorder (nonSSD-V; n = 20), and healthy non-violent participants (HC; n = 196) from lifetime normative trajectories of cortical thickness, surface area, and subcortical volumes. Normative models based on Freesurfer derived regions of interest from 58,836 individuals were used to investigate individual deviances, group differences, and associations to psychopathy traits. We found overall heterogeneous patterns of individual deviations from the norm, which were most prominent for regions within the collateral transverse sulcus, lingual gyrus, and cerebellum among SSD-V, a pattern that differed from SSD-NV (parieto-occipital and suborbital sulci), and nonSSD-V (paracentral and middle frontal regions). We found no significant associations to psychopathy traits. By applying normative modeling, we demonstrate heterogeneous patterns of brain morphometry deviations associated with violence and psychosis. While the results warrant replication, studies addressing individual brain deviations may contribute to improved understanding of the neurobiological underpinnings of comorbid violence and psychosis, which ultimately may have clinical impact on treatment and forensic psychiatric evaluations.

The restructuring and optimization of the cerebral cortex from early childhood and through adolescence is an essential feature of human brain development, underlying immense cognitive improvements. Beyond established morphometric cortical assessments, the T1w/T2w ratio quantifies partly separate biological processes, and might inform models of typical neurocognitive development and developmental psychopathology. In the present study, we computed vertex-wise T1w/T2w ratio across the cortical surface in 621 youths (3-21 years) sampled from the Pediatric Imaging, Neurocognition, and Genetics (PING) study and tested for associations with individual differences in age, sex, and both general and specific cognitive abilities. The results showed a near global linear age-related increase in T1w/T2w ratio across the brain surface, with a general posterior to anterior increasing gradient in association strength. Moreover, results indicated that boys in late adolescence had regionally higher T1w/T2w ratio as compared to girls. Across individuals, T1w/T2w ratio was negatively associated with general and several specific cognitive abilities mainly within anterior cortical regions. Our study indicates age-related differences in T1w/T2w ratio throughout childhood, adolescence and young adulthood, in line with the known protracted myelination of the cortex. Moreover, the study supports T1w/T2w ratio as a promising surrogate measure of individual differences in intracortical brain structure in neurodevelopment.

Background: The corpus callosum (CC) is a brain structure with a high heritability and potential role in psychiatric disorders. However, the genetic architecture of the CC and the genetic link with psychiatric disorders remain largely unclear. We investigated the genetic architectures of the volume of the CC and its subregions and the genetic overlap with psychiatric disorders. Methods: We applied multivariate genome-wide association study (GWAS) to genetic and T1-weighted magnetic resonance imaging (MRI) data of 40,894 individuals from the UK Biobank, aiming to boost genetic discovery and to assess the pleiotropic effects across volumes of the five subregions of the CC (posterior, mid-posterior, central, mid-anterior and anterior) obtained by FreeSurfer 7.1. Multivariate GWAS was run combining all subregions, co-varying for relevant variables. Gene-set enrichment analyses were performed using MAGMA. Linkage disequilibrium score regression (LDSC) was used to determine Single nucleotide polymorphism (SNP)-based heritability of total CC volume and volumes of its subregions as well as their genetic correlations with relevant psychiatric traits. Results: We identified 70 independent loci with distributed effects across the five subregions of the CC (p < 5 × 10⁻⁸). Additionally, we identified 33 significant loci in the anterior subregion, 23 in the mid-anterior, 29 in the central, 7 in the mid-posterior and 56 in the posterior subregion. Gene-set analysis revealed 156 significant genes contributing to volume of the CC subregions (p < 2.6 × 10⁻⁶). LDSC estimated the heritability of CC to (h²_SNP = 0.38, SE = 0.03) and subregions ranging from 0.22 (SE = 0.02) to 0.37 (SE = 0.03). We found significant genetic correlations of total CC volume with bipolar disorder (BD, r_g = −0.09, SE = 0.03; p = 5.9 × 10⁻³) and drinks consumed per week (r_g = −0.09, SE = 0.02; p = 4.8 × 10⁻⁴), and volume of the mid-anterior subregion with BD (r_g = −0.12, SE = 0.02; p = 2.5 × 10⁻⁴), major depressive disorder (MDD) (r_g = −0.12, SE = 0.04; p = 3.6 × 10⁻³), drinks consumed per week (r_g = −0.13, SE = 0.04; p = 1.8 × 10⁻³) and cannabis use (r_g = −0.09, SE = 0.03; p = 8.4 × 10⁻³). Conclusions: Our results demonstrate that the CC has a polygenic architecture implicating multiple genes and show that CC subregion volumes are heritable. We found that distinct genetic factors are involved in the development of anterior and posterior subregions, consistent with their divergent functional specialisation. Significant genetic correlation between volumes of the CC and BD, drinks per week, MDD and cannabis consumption subregion volumes with psychiatric traits is noteworthy and deserving of further investigation.