Neuroanatomical correlates of HPA axis biomarkers among children at elevated risk for schizophrenia

Abstract

Background: The hypothalamic-pituitary-adrenal (HPA) axis disturbances observed in psychosis might feasibly contribute to some of the structural brain abnormalities associated with the disorder (i.e., via the potentially toxic effects of elevated glucocorticoid levels on the brain). However, the association between brain structure and cortisol has rarely been investigated. The current study examined relationships between salivary cortisol and regional brain volumes among at-risk children aged 11-14 years, who present multiple antecedents of schizophrenia (ASz=30) or a family history of illness (FHx=22), and typically-developing children (TD=32). Methods: Volumes of specific brain regions known to express glucocorticoid receptors were extracted from structural brain images using FreeSurfer. Diurnal cortisol levels and the cortisol awakening response (CAR) were determined from multiple salivary cortisol samples obtained throughout the day. Results: In the FHx group only, higher diurnal cortisol was correlated with smaller right amygdala and white matter hypointensity volumes (r=-0.51 and -0.62, respectively, p<0.05). In post-hoc analyses, a higher CAR was associated with smaller bilateral cerebellum, bilateral hippocampus, right ventral diencephalon, and total subcortical grey matter volumes among FHx children with a first-degree relative with schizophrenia. In FHx children with an affected second-degree relative, increased diurnal cortisol levels were associated with smaller left hippocampal and white matter hypointensity volumes. Conclusions: Our findings are consistent with the few previous studies examining cortisol and brain structure among adults with established psychosis. These associations may not necessarily reflect a direct effect of cortisol on the developing brain; indeed, abnormal neurodevelopment affecting these brain regions may contribute to HPA-axis dysregulation. Funding: This work was supported by a Sir Henry Wellcome Postdoctoral Fellowship (107395/Z/15/Z) and a Waterloo Foundation Child Development Fund grant (164/1719)