Decomposing cortical thickness heterogeneity by evaluating PRS of psychiatric and medical disorders

528 

Abstract Submission 

Hadis Jameei¹, William Reay², Rebecca Cooper¹, Sina Mansour L¹, Murray Cairns², Andrew Zalesky¹, Maria Di Biase¹

¹University of Melbourne, Melbourne, VIC, ²University of Newcastle, Newcastle, NSW

Hadis Jameei  
University of Melbourne
Melbourne, VIC

William Reay  
University of Newcastle
Newcastle, NSW

Rebecca Cooper  
University of Melbourne
Melbourne, VIC

Sina Mansour L  
University of Melbourne
Melbourne, VIC

Murray Cairns  
University of Newcastle
Newcastle, NSW

Andrew Zalesky  
University of Melbourne
Melbourne, VIC

Maria Di Biase  
University of Melbourne
Melbourne, VIC

A common feature of neuropsychiatric disorders is reduced cortical thickness, yet the extent and spatial distribution of these reductions are heterogeneous across patients, even within diagnostic groups. To understand this variability from an etiological standpoint, recent studies have examined common genetic risk variants of distinct psychiatric disorders, revealing significant, albeit weak, associations with cortical thickness. We speculate that variability is better explained by genetic risk profiles for diverse psychiatric and non-psychiatric conditions, which are overrepresented in psychiatric populations. Here, we evaluate polygenic risk correlations between psychiatric disorders and commonly comorbid chronic diseases and test their effects on deviations in brain cortical thickness variation.

This study comprises healthy adults with available genotype and magnetic resonance imaging data from the UK Biobank (N=7,873, age=56±8, 45% male). Polygenic risk scores (PRSs) were computed for 21 binary traits, including 5 psychiatric traits and 16 chronic medical conditions overrepresented amongst individuals with a psychiatric diagnosis (Figure 1). Normative models constructed with Generalized Additive Models for Location, Scale and Shape (GAMLSS) calculated person-specific cortical thickness deviations from the median for a given age/sex. Linear regressions tested pair-wise correlations between disorder PRSs, and associations between each PRS with regional cortical thickness deviations. Spatial correspondence (Spin) tests evaluated overlap between traits in terms of their PRS associations with regional deviation profiles. A permutation test was devised to account for the degree of trait-pair PRS correlations in assessing spatial correspondence.

Polygenic risk for the 16 chronic medical conditions exhibited smaller pair-wise correlations across individuals compared to psychiatric disorders, suggesting greater diagnostic specificity and/or more distinct etiologies. Out of 90 psychiatric-psychiatric and psychiatric-chronic trait pairs, 13 pairs departed from expected patterns of genetic and spatial convergence (Figure 2). Of 12 psychiatric-chronic trait pairs that were genetically correlated, only 2 (17%) revealed significant spatial correspondence in their regional distributions of cortical thickness deviations (r range=0.36-0.39; p<0.05). In contrast, 3 (4%) trait pairs exhibited stronger associations in regional cortical deviation profiles relative to that expected by their genetic correlations (r range=0.41-0.44; p<0.05).

Psychiatric disorders with shared genetic components can exhibit varied cortical deviation profiles, consistent with pleiotropic-like effects.6 Conversely, chronic diseases with distinct genetic bases to psychiatric disorders can map to similar deviation profiles of cortical thickness. Our results suggest that moving beyond analyses of single genetic traits leads to improved characterization of complex relationships between genetic architecture and phenotypic variability in cortical thickness. These advancements can help toward developing predictive models of neuro-phenotypic outcomes in psychiatry.

Psychiatric (eg. Depression, Anxiety, Schizophrenia) ¹

Genetic Modeling and Analysis Methods ²

Phenotype-Genotype

Psychiatric Disorders

STRUCTURAL MRI

Other - single nucleotide polymorphisms, chronic medical condition