Cortical surface area abnormalities in young adolescents at ultra-high risk who later develop psychosis

Introduction: Identifying biomarkers of transition in individuals ultra-high risk for psychosis (UHR) has the potential to improve future outcomes [1]. According to structural MRI studies brain maturation during adolescence leads to loss of cortical surface area (SA) [2]. However, UHR and schizophrenia adolescents seem to have larger SA than controls at cross-section [3,4], and its preservation over time may serve as biomarker of transition to psychosis in individuals at familial high risk [5]. Aim: To test whether onset of psychotic symptoms in adolescents at UHR relates to changes in SA. Methods: Case-control study including youth aged 10-17 years, recruited at Hospital Clinic of Barcelona child and adolescent mental health services. UHR individuals were identified using slightly-modified Structured Interview for Prodromal Syndromes criteria. Healthy controls (HC) were recruited from the same geographical area. Exclusion criteria comprised personal history of psychotic symptoms, IQ<70, autism spectrum disorder, presence of neurological disorder, or antecedents of head trauma with loss of consciousness. Local Ethical Review Board approved the study. All participants underwent a comprehensive sociodemographic and clinical evaluation at baseline and after 6, 12 and 18 months follow-up to identify which individuals experienced transition to psychosis (UHR-P) and which didn´t (UHR-NP). Due to non-linear normal SA development during adolescence we divided the sample into “early” (under 15) and “late” (over 15) adolescents. High-resolution magnetic resonance structural images were acquired by 3Tesla scanner. Images were preprocessed employing automated procedures implemented in FreeSurfer 5.3.0, cortical parcellation employed the Desikan-Killiany brain atlas. Analyses: First, mean global and lobar (frontal, parietal, temporal, occipital, insula and cingulate) SA measurements were computed. ANCOVA was performed to test differences between groups in SPSS 22.0, including gender, age, and total intracranial volume as covariates. Significance was set at p<.05, corrected using the false discovery rate (FDR). Results: 76 subjects were included (31 HC versus 33 UHRNP versus. 12 UHR-P, mean ages: 15.5± = 1.5 vs 15.4±1.8 versus 14.8±1.7 (F = .59, p = 0.553); gender (%female): 64.5% versus 63.6% versus 66.7%, (χ2 = .04, p = 0.982). No significant differences in global SA between groups were found (1702.3cm2 vs 1696.3cm2 versus 1696.9cm2, F = .048, p = 0.953). Between-group analyses didn’t show significant group differences in cortical SA by lobes. However, when dividing the sample into early (n = 32) and late adolescents (n = 44); younger adolescents showed a significant group effect in the left frontal lobe (F = 7.06, pFDR = 0.03) with larger SA in UHR-P versus HC (p = .004) and UHR-P versus UHRNP (p = .007); and in the right cingulate cortex (F = 6.38, pFDR = 0.03), with larger SA in UHR-P versus HC (p = 0.006). Late adolescents didn’t show any significant differences between groups. Discussion: “Early” adolescents at UHR showed significantly larger left frontal and right cingulate cortical SA when compared to HC. Furthermore, UHR-P individuals had larger SA than their UHR-NP counterparts, suggesting that cortical SA could serve as a possible marker for early detection of psychosis in UHR individuals during early adolescence. More so, relatively smaller SA in UHR-NP individuals could be a resilience factor for transition to psychosis. Longitudinal changes in SA have the potential to increase our understanding on pathophysiology of transition to clinical illness.