Rashidah Unaib Al-Zayid
Department of Radiology, Family Medicine Academy, Riyadh, Saudi Arabia

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Transdiagnostic brain age gap in schizophrenia and bipolar disorder: a leakage-controlled OpenNeuro analysis Aline Hafidzah; Rashidah Unaib Al-Zayid; Cindy Susanti
Scientia Psychiatrica Vol. 7 No. 3 (2026): Scientia Psychiatrica
Publisher : HM Publisher in collaboration with CMHC Research Center

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37275/scipsy.v7i3.218

Abstract

Background. Brain age gap (BAG), predicted minus chronological brain age, is often elevated in psychotic and mood disorders. Its magnitude depends on the reference sample, age range, scanner, image quality, and age-bias correction. Objective. To estimate transdiagnostic BAG differences in schizophrenia and bipolar disorder with a control-only model, nested validation, and adjustment for scanner and ghost artifact. Methods. This secondary cross-sectional study used the University of California, Los Angeles (UCLA) Consortium for Neuropsychiatric Phenomics dataset (OpenNeuro ds000030): 125 controls, 49 participants with bipolar disorder, and 50 with schizophrenia. Ridge regression used 68 Desikan–Killiany cortical-thickness and 37 intracranial-volume-normalized FreeSurfer measures. Models were trained only in controls within repeated 10-fold validation, with hyperparameters and age-bias correction estimated in training data. Group effects came from heteroskedasticity-consistent linear models adjusted for age, gender, scanner, and ghost artifact. Results. In controls, raw out-of-fold mean absolute error was 6.48 years (95% bootstrap confidence interval [CI] 5.83–7.11), r = 0.562, and R² = 0.277. Adjusted BAG differed from controls by 0.55 years in bipolar disorder (95% CI −0.50 to 1.59; p = 0.304; false-discovery-rate q = 0.304) and by 0.63 years in schizophrenia (95% CI −0.51 to 1.77; p = 0.276; q = 0.304). The schizophrenia–bipolar contrast was 0.08 years (95% CI −1.16 to 1.33; p = 0.895). Sensitivity analyses gave similar estimates. Conclusion. Both patient groups showed small positive estimates with wide overlap, and the confidence intervals spanned zero. The findings support a reproducible open-data workflow, not individual diagnosis or a causal claim of accelerated aging.