Cortical thickness differences stratified by duration since first episode of psychosis in the GENUS dataset
Nora Y. Sun (1) , Sahil Sood (1), Victor Zeng (2,3), Joyce Zhou (2,3), Kareem Hamada (2,3,4), Matcheri Keshavan (2,3), Gabriella Blokland (5), Elisabetta C. del Re (2,3,6)
1. Harvard University, Cambridge, MA
2. Harvard Medical School, Boston, MA
3. Beth Israel Deaconess Medical Center, Boston, MA
4. McLean Hospital, Belmont, MA
5. Maastricht University, Maastricht, The Netherland
6. Veterans Affairs Boston Healthcare System, Brockton, MA
Background: Past studies have yielded mixed results on the impact of duration of illness (DUI) versus aging on cortical thickness (CT) in schizophrenia (SZ). Here, we employ a large dataset (GENUS) that includes participants affected by SZ, individuals at familial high risk for psychosis (FHR) and controls (HC). We focus on brain regions involved in language circuits hypothesizing strong effect of illness chronicity.
Methods: T1-MPRAGE scans (n = 3,007) were quality checked and processed using FreeSurfer 7.1.0 with an in-house pipeline to extract CT for seven regions involved in language processing: Banks of superior temporal sulcus, middle temporal, fusiform, inferior parietal, supramarginal, pars opercularis, pars triangularis and transverse temporal gyrus. SZ were categorized according to DUI < or > 5 years DUI. MANCOVA was employed to compare CT across groups. Multiple comparisons were corrected with False Discovery Rate.
Results: HC have the greatest mean CT in all regions of interest (ROI), followed by FHR, low (=<5 yrs) DUI, and finally high (>5 yrs) DUI. When comparing HC and FHR to probands, all 8 ROI show significantly smaller thickness in probands. Comparison of low versus high DUI probands indicate smaller CT of the middletemporal, banks of the superior temporal sulcus, and supramarginal regions in probands with longer DUI. While FHR have lower CT than HC, the differences in thickness are not significant.
Conclusion: Our investigation identified smaller CT in brain regions involved in language, beyond aging effects and according to DUI, indicating that language regions may be involved in disease progression. Further analyses examining potential correlations between CT differences and cognitive and clinical symptoms is warranted.