Sex-specific cortical brain differences in children at familial high risk for schizophrenia or bipolar disorder

Background
Schizophrenia (SZ) and bipolar disorder (BP) are severe mental disorders that typically emerge during late
adolescence and early adulthood. Both disorders are associated with structural brain abnormalities, including
reduced cortical thickness and volume. Identifying structural brain differences already present before the onset
of symptoms is essential to understanding the emergence of these disorders. A family history of SZ or BP is
the strongest risk factor for severe mental illness, with 15-40% of children at familial high risk (FHR) developing
a psychotic disorder or BP. Studying enriched cohorts, such as children at FHR, therefore provides the best
opportunity to examine neurodevelopmental processes preceding illness. Studies of children and adolescents
at FHR for SZ (FHR-SZ) show smaller brain and gray matter volumes, thinner cortices, and reduced surface
area compared to children at FHR for BP (FHR-BP) and controls, with some differences linked to later
psychosis onset. Children and adolescents at FHR-BP exhibit subtler alterations compared to controls,
including larger ventricular size. However, most prior FHR studies in children and adolescents did not examine
sex differences despite known sex differences in brain maturational trajectories in this age range.
Methods
We analyzed neuroimaging data from the Danish High Risk and Resilience Study (VIA), a multi-site populationbased cohort study, to examine structural brain differences associated with FHR for severe mental illness. The
sample included 278 children: 101 at FHR-SZ (mean[SD] age: 12.1[0.3] years; 49.5% female), 64 at FHR-BP
(mean[SD] age: 12.1[0.3] years; 50% female), and 113 age- and sex-matched population-based controls (PBC;
mean[SD]: 12.1[0.3] years; 49.6% female). We extracted structural brain measures from T1-weighted images
using FreeSurfer version 7.1.1. Since developmental trajectories of brain structure are well known to vary by
sex, we specifically tested for group (FHR-SZ, FHR-BP, PBC)-by-sex differences in brain volume, cortical
volume, cortical thickness, and surface area, controlling for age, site, and a proxy for in-scanner movement.
Results
We found significant group-by-sex interactions for brain volume (eta2=0.038, F=5.270, p=0.006, q=0.021,
BF10=8.362), cortical volume (eta2=0.031, F=4.364, p=0.014, q=0.021, BF10=3.523), and surface area
(eta2=0.030, F=4.220, p=0.016, q=0.021, BF10=2.982), with no main effects of group. We did not observe
significant group-by-sex (eta2=0.002, F=0.20, p=0.801, BF10=0.088) or group (eta2=0.019, F=2.775, p=0.064,
BF10=0.507) effects for mean cortical thickness. Since these results suggested sex-specific group effects in
brain structure, we ran post hoc pairwise group analyses in females and males separately. These analyses
revealed that FHR-SZ males exhibited smaller brain volume, cortical volume, and surface area than PBC
males (Cohen’s d range=-0.624 to -0.489; p range=0.002 to 0.015), while FHR-BP females displayed larger
brain and cortical volumes compared to PBC females (Cohen’s d range=0.525 to 0.537; p range=0.017 to
0.020). No significant differences were observed for cortical thickness (ps>0.210).
Discussion
Our results indicate that already in middle childhood, children at FHR-SZ and FHR-BP show sex-specific global
volumetric brain differences. This suggests that the volumetric differences may be endophenotypic markers of
biological vulnerability prior to the potential onset of SZ and BP. Our results further highlight the importance of
considering sex in developmental neuropsychiatric research. We are currently conducting follow-up
longitudinal analyses to investigate how these structural brain differences change during adolescence, and to
test their association with psychotic-like experiences and emerging psychopathology. This work will clarify the
clinical and predictive utility of the early sex-specific structural brain markers of FHR-SZ and FHR-BP identified
here.