Coordinated cortical thickness alterations across six neurodevelopmental and psychiatric disorders

Abstract

Neuropsychiatric disorders are increasingly conceptualized as overlapping spectra sharing multi-level neurobiological alterations. However, whether transdiagnostic cortical alterations covary in a biologically meaningful way is currently unknown. Here, we studied co-alteration networks across six neurodevelopmental and psychiatric disorders, reflecting pathological structural covariance. In 12,024 patients and 18,969 controls from the ENIGMA consortium, we observed that co-alteration patterns followed normative connectome organization and were anchored to prefrontal and temporal disease epicenters. Manifold learning revealed frontal-to-temporal and sensory/limbic-to-occipitoparietal transdiagnostic gradients, differentiating shared illness effects on cortical thickness along these axes. The principal gradient aligned with a normative cortical thickness covariance gradient and established a transcriptomic link to cortico-cerebello-thalamic circuits. Moreover, transdiagnostic gradients segregated functional networks involved in basic sensory, attentional/perceptual, and domain-general cognitive processes, and distinguished between regional cytoarchitectonic profiles. Together, our findings indicate that shared illness effects occur in a synchronized fashion and along multiple levels of hierarchical cortical organization.

Fig. 1. Hubs and epicenters shaping transdiagnostic co-alteration patterns.

A Disorder-specific Cohen’s d maps indicating case-control differences in cortical thickness. B Normative connectivity matrices derived from resting-state functional magnetic resonance imaging (rs-fMRI) and diffusion-weighted tensor imaging (DTI) from the Human Connectome Project (HCP) and hubs (degree centrality). C Left: Computation of co-alteration hubs. Degree centrality was computed as the sum of above-threshold (80%) connections at each parcel using disorder maps from the Enhancing Neuroimaging Genetics through Meta-analyses (ENIGMA) consortium. Right: Visualization of the epicenter mapping approach using resting state functional connectivity (rsFC) or DTI. Seed-based connectivity profiles were systematically correlated with co-alteration hubs (using Pearson’s r and assessing significance via two-sided spin-tests, correcting for spatial auto-correlation, without further correction for multiple comparisons). D Transdiagnostic disease epicenters are depicted as correlations between co-alteration hubs and HCP normative seed-based connectivity profiles (rs-fMRI or diffusion tensor imaging (DTI)), thresholded at p_spin < 0.05 (this panel shows DTI examples). High correlations imply high likelihood of a structure constituting a disease epicenter. Top five functional and structural disease epicenters are framed in white/black. Source data are provided as a Source Data file. ADHD = Attention-deficit/hyperactivity disorder; ASD = Autism spectrum disorder, BD = Bipolar disorder, MDD = Major depressive disorder, OCD = Obsessive-compulsive disorder, SCZ = Schizophrenia spectrum disorders.

Fig. 2. Macroscale organization of transdiagnostic covariance in cortical thickness alterations.

A A cross-disorder structural covariance matrix was thresholded at 80% and decomposed using diffusion map embedding. Covariance along the principal (G1) and second (G2) gradients is depicted on the right. B Transdiagnostic gradients G1 and G2. C Correlation between a normative axis of cortical thickness (CT) covariance and transdiagnostic gradients. D Cross-condition gradients stratified according to von Economo-Koskinas cytoarchitectonic classes. E Developmental gene enrichment analysis based on 232 genes for which spatial expression patterns correlated with G1 (of which 146 showed a positive correlation, i.e., were overexpressed in prefrontal compared to temporal regions). F Meta-analysis for diverse cognitive functions obtained from NeuroSynth. We computed parcel-wise z-statistics, capturing node-function associations, and calculated the center of gravity of each function along 20 five-percentile bins of G1 and G2. Function terms are ordered by the weighted mean of their location along the gradients. Source data are provided as a Source Data file.

Fig. 3. Embedding of six disorders within transdiagnostic co-alteration networks.

A Computation of transdiagnostic and within-disorder co-alteration matrices. B Region-wise correspondence between disorder-specific and transdiagnostic co-alteration profiles. Disorder-specific inter-regional difference scores were inverted so that higher correlations with transdiagnostic patterns indicate higher coupling. C Similarity of illness effects between disorders, i.e., correlations of Cohen’s d maps, and how they cluster together in a two-cluster solution (D). Position of individual disorders within a transdiagnostic co-alteration space based on E the correlation between transdiagnostic hubs and Cohen’s d maps (x-axis) and the overlap between transdiagnostic and disorder-specific epicenters (y-axis); and F the correlation between the principal (G1) and secondary (G2) transdiagnostic gradients with Cohen’s d maps. Source data are provided as a Source Data file. ASD = Autism spectrum disorder, SCZ = Schizophrenia spectrum diagnoses, MDD = Major depressive disorder, ADHD = Attention-deficit/hyperactivity disorder, BD = Bipolar disorder, OCD = Obsessive-compulsive disorder.